git/add-patch.c

1756 строки
50 KiB
C
Исходник Обычный вид История

built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
#include "cache.h"
#include "add-interactive.h"
#include "strbuf.h"
#include "run-command.h"
#include "strvec.h"
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
#include "pathspec.h"
#include "color.h"
built-in add -p: adjust hunk headers as needed When skipping a hunk that adds a different number of lines than it removes, we need to adjust the subsequent hunk headers of non-skipped hunks: in pathological cases, the context is not enough to determine precisely where the patch should be applied. This problem was identified in 23fea4c240 (t3701: add failing test for pathological context lines, 2018-03-01) and fixed in the Perl version in fecc6f3a68 (add -p: adjust offsets of subsequent hunks when one is skipped, 2018-03-01). And this patch fixes it in the C version of `git add -p`. In contrast to the Perl version, we try to keep the extra text on the hunk header (which typically contains the signature of the function whose code is changed in the hunk) intact. Note: while the C version does not support staging mode changes at this stage, we already prepare for this by simply skipping the hunk header if both old and new offset is 0 (this cannot happen for regular hunks, and we will use this as an indicator that we are looking at a special hunk). Likewise, we already prepare for hunk splitting by handling the absence of extra text in the hunk header gracefully: only the first split hunk will have that text, the others will not (indicated by an empty extra text start/end range). Preparing for hunk splitting already at this stage avoids an indentation change of the entire hunk header-printing block later, and is almost as easy to review as without that handling. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:51 +03:00
#include "diff.h"
#include "compat/terminal.h"
#include "prompt.h"
built-in add -p: adjust hunk headers as needed When skipping a hunk that adds a different number of lines than it removes, we need to adjust the subsequent hunk headers of non-skipped hunks: in pathological cases, the context is not enough to determine precisely where the patch should be applied. This problem was identified in 23fea4c240 (t3701: add failing test for pathological context lines, 2018-03-01) and fixed in the Perl version in fecc6f3a68 (add -p: adjust offsets of subsequent hunks when one is skipped, 2018-03-01). And this patch fixes it in the C version of `git add -p`. In contrast to the Perl version, we try to keep the extra text on the hunk header (which typically contains the signature of the function whose code is changed in the hunk) intact. Note: while the C version does not support staging mode changes at this stage, we already prepare for this by simply skipping the hunk header if both old and new offset is 0 (this cannot happen for regular hunks, and we will use this as an indicator that we are looking at a special hunk). Likewise, we already prepare for hunk splitting by handling the absence of extra text in the hunk header gracefully: only the first split hunk will have that text, the others will not (indicated by an empty extra text start/end range). Preparing for hunk splitting already at this stage avoids an indentation change of the entire hunk header-printing block later, and is almost as easy to review as without that handling. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:51 +03:00
enum prompt_mode_type {
PROMPT_MODE_CHANGE = 0, PROMPT_DELETION, PROMPT_ADDITION, PROMPT_HUNK,
PROMPT_MODE_MAX, /* must be last */
};
struct patch_mode {
/*
* The magic constant 4 is chosen such that all patch modes
* provide enough space for three command-line arguments followed by a
* trailing `NULL`.
*/
const char *diff_cmd[4], *apply_args[4], *apply_check_args[4];
unsigned is_reverse:1, index_only:1, apply_for_checkout:1;
const char *prompt_mode[PROMPT_MODE_MAX];
const char *edit_hunk_hint, *help_patch_text;
};
static struct patch_mode patch_mode_add = {
.diff_cmd = { "diff-files", NULL },
.apply_args = { "--cached", NULL },
.apply_check_args = { "--cached", NULL },
.prompt_mode = {
N_("Stage mode change [y,n,q,a,d%s,?]? "),
N_("Stage deletion [y,n,q,a,d%s,?]? "),
N_("Stage addition [y,n,q,a,d%s,?]? "),
N_("Stage this hunk [y,n,q,a,d%s,?]? ")
},
.edit_hunk_hint = N_("If the patch applies cleanly, the edited hunk "
"will immediately be marked for staging."),
.help_patch_text =
N_("y - stage this hunk\n"
"n - do not stage this hunk\n"
"q - quit; do not stage this hunk or any of the remaining "
"ones\n"
"a - stage this hunk and all later hunks in the file\n"
"d - do not stage this hunk or any of the later hunks in "
"the file\n")
};
static struct patch_mode patch_mode_stash = {
.diff_cmd = { "diff-index", "HEAD", NULL },
.apply_args = { "--cached", NULL },
.apply_check_args = { "--cached", NULL },
.prompt_mode = {
N_("Stash mode change [y,n,q,a,d%s,?]? "),
N_("Stash deletion [y,n,q,a,d%s,?]? "),
N_("Stash addition [y,n,q,a,d%s,?]? "),
N_("Stash this hunk [y,n,q,a,d%s,?]? "),
},
.edit_hunk_hint = N_("If the patch applies cleanly, the edited hunk "
"will immediately be marked for stashing."),
.help_patch_text =
N_("y - stash this hunk\n"
"n - do not stash this hunk\n"
"q - quit; do not stash this hunk or any of the remaining "
"ones\n"
"a - stash this hunk and all later hunks in the file\n"
"d - do not stash this hunk or any of the later hunks in "
"the file\n"),
};
static struct patch_mode patch_mode_reset_head = {
.diff_cmd = { "diff-index", "--cached", NULL },
.apply_args = { "-R", "--cached", NULL },
.apply_check_args = { "-R", "--cached", NULL },
.is_reverse = 1,
.index_only = 1,
.prompt_mode = {
N_("Unstage mode change [y,n,q,a,d%s,?]? "),
N_("Unstage deletion [y,n,q,a,d%s,?]? "),
N_("Unstage addition [y,n,q,a,d%s,?]? "),
N_("Unstage this hunk [y,n,q,a,d%s,?]? "),
},
.edit_hunk_hint = N_("If the patch applies cleanly, the edited hunk "
"will immediately be marked for unstaging."),
.help_patch_text =
N_("y - unstage this hunk\n"
"n - do not unstage this hunk\n"
"q - quit; do not unstage this hunk or any of the remaining "
"ones\n"
"a - unstage this hunk and all later hunks in the file\n"
"d - do not unstage this hunk or any of the later hunks in "
"the file\n"),
};
static struct patch_mode patch_mode_reset_nothead = {
.diff_cmd = { "diff-index", "-R", "--cached", NULL },
.apply_args = { "--cached", NULL },
.apply_check_args = { "--cached", NULL },
.index_only = 1,
.prompt_mode = {
N_("Apply mode change to index [y,n,q,a,d%s,?]? "),
N_("Apply deletion to index [y,n,q,a,d%s,?]? "),
N_("Apply addition to index [y,n,q,a,d%s,?]? "),
N_("Apply this hunk to index [y,n,q,a,d%s,?]? "),
},
.edit_hunk_hint = N_("If the patch applies cleanly, the edited hunk "
"will immediately be marked for applying."),
.help_patch_text =
N_("y - apply this hunk to index\n"
"n - do not apply this hunk to index\n"
"q - quit; do not apply this hunk or any of the remaining "
"ones\n"
"a - apply this hunk and all later hunks in the file\n"
"d - do not apply this hunk or any of the later hunks in "
"the file\n"),
};
static struct patch_mode patch_mode_checkout_index = {
.diff_cmd = { "diff-files", NULL },
.apply_args = { "-R", NULL },
.apply_check_args = { "-R", NULL },
.is_reverse = 1,
.prompt_mode = {
N_("Discard mode change from worktree [y,n,q,a,d%s,?]? "),
N_("Discard deletion from worktree [y,n,q,a,d%s,?]? "),
N_("Discard addition from worktree [y,n,q,a,d%s,?]? "),
N_("Discard this hunk from worktree [y,n,q,a,d%s,?]? "),
},
.edit_hunk_hint = N_("If the patch applies cleanly, the edited hunk "
"will immediately be marked for discarding."),
.help_patch_text =
N_("y - discard this hunk from worktree\n"
"n - do not discard this hunk from worktree\n"
"q - quit; do not discard this hunk or any of the remaining "
"ones\n"
"a - discard this hunk and all later hunks in the file\n"
"d - do not discard this hunk or any of the later hunks in "
"the file\n"),
};
static struct patch_mode patch_mode_checkout_head = {
.diff_cmd = { "diff-index", NULL },
.apply_for_checkout = 1,
.apply_check_args = { "-R", NULL },
.is_reverse = 1,
.prompt_mode = {
N_("Discard mode change from index and worktree [y,n,q,a,d%s,?]? "),
N_("Discard deletion from index and worktree [y,n,q,a,d%s,?]? "),
N_("Discard addition from index and worktree [y,n,q,a,d%s,?]? "),
N_("Discard this hunk from index and worktree [y,n,q,a,d%s,?]? "),
},
.edit_hunk_hint = N_("If the patch applies cleanly, the edited hunk "
"will immediately be marked for discarding."),
.help_patch_text =
N_("y - discard this hunk from index and worktree\n"
"n - do not discard this hunk from index and worktree\n"
"q - quit; do not discard this hunk or any of the remaining "
"ones\n"
"a - discard this hunk and all later hunks in the file\n"
"d - do not discard this hunk or any of the later hunks in "
"the file\n"),
};
static struct patch_mode patch_mode_checkout_nothead = {
.diff_cmd = { "diff-index", "-R", NULL },
.apply_for_checkout = 1,
.apply_check_args = { NULL },
.prompt_mode = {
N_("Apply mode change to index and worktree [y,n,q,a,d%s,?]? "),
N_("Apply deletion to index and worktree [y,n,q,a,d%s,?]? "),
N_("Apply addition to index and worktree [y,n,q,a,d%s,?]? "),
N_("Apply this hunk to index and worktree [y,n,q,a,d%s,?]? "),
},
.edit_hunk_hint = N_("If the patch applies cleanly, the edited hunk "
"will immediately be marked for applying."),
.help_patch_text =
N_("y - apply this hunk to index and worktree\n"
"n - do not apply this hunk to index and worktree\n"
"q - quit; do not apply this hunk or any of the remaining "
"ones\n"
"a - apply this hunk and all later hunks in the file\n"
"d - do not apply this hunk or any of the later hunks in "
"the file\n"),
};
static struct patch_mode patch_mode_worktree_head = {
.diff_cmd = { "diff-index", NULL },
.apply_args = { "-R", NULL },
.apply_check_args = { "-R", NULL },
.is_reverse = 1,
.prompt_mode = {
N_("Discard mode change from index and worktree [y,n,q,a,d%s,?]? "),
N_("Discard deletion from index and worktree [y,n,q,a,d%s,?]? "),
N_("Discard addition from index and worktree [y,n,q,a,d%s,?]? "),
N_("Discard this hunk from index and worktree [y,n,q,a,d%s,?]? "),
},
.edit_hunk_hint = N_("If the patch applies cleanly, the edited hunk "
"will immediately be marked for discarding."),
.help_patch_text =
N_("y - discard this hunk from worktree\n"
"n - do not discard this hunk from worktree\n"
"q - quit; do not discard this hunk or any of the remaining "
"ones\n"
"a - discard this hunk and all later hunks in the file\n"
"d - do not discard this hunk or any of the later hunks in "
"the file\n"),
};
static struct patch_mode patch_mode_worktree_nothead = {
.diff_cmd = { "diff-index", "-R", NULL },
.apply_args = { NULL },
.apply_check_args = { NULL },
.prompt_mode = {
N_("Apply mode change to index and worktree [y,n,q,a,d%s,?]? "),
N_("Apply deletion to index and worktree [y,n,q,a,d%s,?]? "),
N_("Apply addition to index and worktree [y,n,q,a,d%s,?]? "),
N_("Apply this hunk to index and worktree [y,n,q,a,d%s,?]? "),
},
.edit_hunk_hint = N_("If the patch applies cleanly, the edited hunk "
"will immediately be marked for applying."),
.help_patch_text =
N_("y - apply this hunk to worktree\n"
"n - do not apply this hunk to worktree\n"
"q - quit; do not apply this hunk or any of the remaining "
"ones\n"
"a - apply this hunk and all later hunks in the file\n"
"d - do not apply this hunk or any of the later hunks in "
"the file\n"),
};
built-in add -p: adjust hunk headers as needed When skipping a hunk that adds a different number of lines than it removes, we need to adjust the subsequent hunk headers of non-skipped hunks: in pathological cases, the context is not enough to determine precisely where the patch should be applied. This problem was identified in 23fea4c240 (t3701: add failing test for pathological context lines, 2018-03-01) and fixed in the Perl version in fecc6f3a68 (add -p: adjust offsets of subsequent hunks when one is skipped, 2018-03-01). And this patch fixes it in the C version of `git add -p`. In contrast to the Perl version, we try to keep the extra text on the hunk header (which typically contains the signature of the function whose code is changed in the hunk) intact. Note: while the C version does not support staging mode changes at this stage, we already prepare for this by simply skipping the hunk header if both old and new offset is 0 (this cannot happen for regular hunks, and we will use this as an indicator that we are looking at a special hunk). Likewise, we already prepare for hunk splitting by handling the absence of extra text in the hunk header gracefully: only the first split hunk will have that text, the others will not (indicated by an empty extra text start/end range). Preparing for hunk splitting already at this stage avoids an indentation change of the entire hunk header-printing block later, and is almost as easy to review as without that handling. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:51 +03:00
struct hunk_header {
unsigned long old_offset, old_count, new_offset, new_count;
/*
* Start/end offsets to the extra text after the second `@@` in the
* hunk header, e.g. the function signature. This is expected to
* include the newline.
*/
size_t extra_start, extra_end, colored_extra_start, colored_extra_end;
};
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
struct hunk {
size_t start, end, colored_start, colored_end, splittable_into;
ssize_t delta;
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
enum { UNDECIDED_HUNK = 0, SKIP_HUNK, USE_HUNK } use;
built-in add -p: adjust hunk headers as needed When skipping a hunk that adds a different number of lines than it removes, we need to adjust the subsequent hunk headers of non-skipped hunks: in pathological cases, the context is not enough to determine precisely where the patch should be applied. This problem was identified in 23fea4c240 (t3701: add failing test for pathological context lines, 2018-03-01) and fixed in the Perl version in fecc6f3a68 (add -p: adjust offsets of subsequent hunks when one is skipped, 2018-03-01). And this patch fixes it in the C version of `git add -p`. In contrast to the Perl version, we try to keep the extra text on the hunk header (which typically contains the signature of the function whose code is changed in the hunk) intact. Note: while the C version does not support staging mode changes at this stage, we already prepare for this by simply skipping the hunk header if both old and new offset is 0 (this cannot happen for regular hunks, and we will use this as an indicator that we are looking at a special hunk). Likewise, we already prepare for hunk splitting by handling the absence of extra text in the hunk header gracefully: only the first split hunk will have that text, the others will not (indicated by an empty extra text start/end range). Preparing for hunk splitting already at this stage avoids an indentation change of the entire hunk header-printing block later, and is almost as easy to review as without that handling. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:51 +03:00
struct hunk_header header;
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
};
struct add_p_state {
built-in add -p: adjust hunk headers as needed When skipping a hunk that adds a different number of lines than it removes, we need to adjust the subsequent hunk headers of non-skipped hunks: in pathological cases, the context is not enough to determine precisely where the patch should be applied. This problem was identified in 23fea4c240 (t3701: add failing test for pathological context lines, 2018-03-01) and fixed in the Perl version in fecc6f3a68 (add -p: adjust offsets of subsequent hunks when one is skipped, 2018-03-01). And this patch fixes it in the C version of `git add -p`. In contrast to the Perl version, we try to keep the extra text on the hunk header (which typically contains the signature of the function whose code is changed in the hunk) intact. Note: while the C version does not support staging mode changes at this stage, we already prepare for this by simply skipping the hunk header if both old and new offset is 0 (this cannot happen for regular hunks, and we will use this as an indicator that we are looking at a special hunk). Likewise, we already prepare for hunk splitting by handling the absence of extra text in the hunk header gracefully: only the first split hunk will have that text, the others will not (indicated by an empty extra text start/end range). Preparing for hunk splitting already at this stage avoids an indentation change of the entire hunk header-printing block later, and is almost as easy to review as without that handling. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:51 +03:00
struct add_i_state s;
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
struct strbuf answer, buf;
/* parsed diff */
struct strbuf plain, colored;
struct file_diff {
struct hunk head;
struct hunk *hunk;
size_t hunk_nr, hunk_alloc;
unsigned deleted:1, added:1, mode_change:1,binary:1;
} *file_diff;
size_t file_diff_nr;
/* patch mode */
struct patch_mode *mode;
const char *revision;
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
};
static void add_p_state_clear(struct add_p_state *s)
{
size_t i;
strbuf_release(&s->answer);
strbuf_release(&s->buf);
strbuf_release(&s->plain);
strbuf_release(&s->colored);
for (i = 0; i < s->file_diff_nr; i++)
free(s->file_diff[i].hunk);
free(s->file_diff);
clear_add_i_state(&s->s);
}
static void err(struct add_p_state *s, const char *fmt, ...)
{
va_list args;
va_start(args, fmt);
fputs(s->s.error_color, stderr);
vfprintf(stderr, fmt, args);
fputs(s->s.reset_color, stderr);
fputc('\n', stderr);
va_end(args);
}
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
static void setup_child_process(struct add_p_state *s,
struct child_process *cp, ...)
{
va_list ap;
const char *arg;
va_start(ap, cp);
while ((arg = va_arg(ap, const char *)))
strvec_push(&cp->args, arg);
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
va_end(ap);
cp->git_cmd = 1;
strvec_pushf(&cp->env_array,
INDEX_ENVIRONMENT "=%s", s->s.r->index_file);
built-in add -p: adjust hunk headers as needed When skipping a hunk that adds a different number of lines than it removes, we need to adjust the subsequent hunk headers of non-skipped hunks: in pathological cases, the context is not enough to determine precisely where the patch should be applied. This problem was identified in 23fea4c240 (t3701: add failing test for pathological context lines, 2018-03-01) and fixed in the Perl version in fecc6f3a68 (add -p: adjust offsets of subsequent hunks when one is skipped, 2018-03-01). And this patch fixes it in the C version of `git add -p`. In contrast to the Perl version, we try to keep the extra text on the hunk header (which typically contains the signature of the function whose code is changed in the hunk) intact. Note: while the C version does not support staging mode changes at this stage, we already prepare for this by simply skipping the hunk header if both old and new offset is 0 (this cannot happen for regular hunks, and we will use this as an indicator that we are looking at a special hunk). Likewise, we already prepare for hunk splitting by handling the absence of extra text in the hunk header gracefully: only the first split hunk will have that text, the others will not (indicated by an empty extra text start/end range). Preparing for hunk splitting already at this stage avoids an indentation change of the entire hunk header-printing block later, and is almost as easy to review as without that handling. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:51 +03:00
}
static int parse_range(const char **p,
unsigned long *offset, unsigned long *count)
{
char *pend;
*offset = strtoul(*p, &pend, 10);
if (pend == *p)
return -1;
if (*pend != ',') {
*count = 1;
*p = pend;
return 0;
}
*count = strtoul(pend + 1, (char **)p, 10);
return *p == pend + 1 ? -1 : 0;
}
static int parse_hunk_header(struct add_p_state *s, struct hunk *hunk)
{
struct hunk_header *header = &hunk->header;
const char *line = s->plain.buf + hunk->start, *p = line;
char *eol = memchr(p, '\n', s->plain.len - hunk->start);
if (!eol)
eol = s->plain.buf + s->plain.len;
if (!skip_prefix(p, "@@ -", &p) ||
parse_range(&p, &header->old_offset, &header->old_count) < 0 ||
!skip_prefix(p, " +", &p) ||
parse_range(&p, &header->new_offset, &header->new_count) < 0 ||
!skip_prefix(p, " @@", &p))
return error(_("could not parse hunk header '%.*s'"),
(int)(eol - line), line);
hunk->start = eol - s->plain.buf + (*eol == '\n');
header->extra_start = p - s->plain.buf;
header->extra_end = hunk->start;
if (!s->colored.len) {
header->colored_extra_start = header->colored_extra_end = 0;
return 0;
}
/* Now find the extra text in the colored diff */
line = s->colored.buf + hunk->colored_start;
eol = memchr(line, '\n', s->colored.len - hunk->colored_start);
if (!eol)
eol = s->colored.buf + s->colored.len;
p = memmem(line, eol - line, "@@ -", 4);
if (!p)
return error(_("could not parse colored hunk header '%.*s'"),
(int)(eol - line), line);
p = memmem(p + 4, eol - p - 4, " @@", 3);
if (!p)
return error(_("could not parse colored hunk header '%.*s'"),
(int)(eol - line), line);
hunk->colored_start = eol - s->colored.buf + (*eol == '\n');
header->colored_extra_start = p + 3 - s->colored.buf;
header->colored_extra_end = hunk->colored_start;
return 0;
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
}
static int is_octal(const char *p, size_t len)
{
if (!len)
return 0;
while (len--)
if (*p < '0' || *(p++) > '7')
return 0;
return 1;
}
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
static int parse_diff(struct add_p_state *s, const struct pathspec *ps)
{
struct strvec args = STRVEC_INIT;
const char *diff_algorithm = s->s.interactive_diff_algorithm;
struct strbuf *plain = &s->plain, *colored = NULL;
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
struct child_process cp = CHILD_PROCESS_INIT;
char *p, *pend, *colored_p = NULL, *colored_pend = NULL, marker = '\0';
size_t file_diff_alloc = 0, i, color_arg_index;
struct file_diff *file_diff = NULL;
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
struct hunk *hunk = NULL;
int res;
strvec_pushv(&args, s->mode->diff_cmd);
if (diff_algorithm)
strvec_pushf(&args, "--diff-algorithm=%s", diff_algorithm);
if (s->revision) {
struct object_id oid;
strvec_push(&args,
/* could be on an unborn branch */
!strcmp("HEAD", s->revision) &&
get_oid("HEAD", &oid) ?
empty_tree_oid_hex() : s->revision);
}
color_arg_index = args.nr;
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
/* Use `--no-color` explicitly, just in case `diff.color = always`. */
strvec_pushl(&args, "--no-color", "-p", "--", NULL);
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
for (i = 0; i < ps->nr; i++)
strvec_push(&args, ps->items[i].original);
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
setup_child_process(s, &cp, NULL);
cp.argv = args.v;
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
res = capture_command(&cp, plain, 0);
if (res) {
strvec_clear(&args);
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
return error(_("could not parse diff"));
}
if (!plain->len) {
strvec_clear(&args);
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
return 0;
}
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
strbuf_complete_line(plain);
if (want_color_fd(1, -1)) {
struct child_process colored_cp = CHILD_PROCESS_INIT;
const char *diff_filter = s->s.interactive_diff_filter;
setup_child_process(s, &colored_cp, NULL);
xsnprintf((char *)args.v[color_arg_index], 8, "--color");
colored_cp.argv = args.v;
colored = &s->colored;
res = capture_command(&colored_cp, colored, 0);
strvec_clear(&args);
if (res)
return error(_("could not parse colored diff"));
if (diff_filter) {
struct child_process filter_cp = CHILD_PROCESS_INIT;
setup_child_process(s, &filter_cp,
diff_filter, NULL);
filter_cp.git_cmd = 0;
filter_cp.use_shell = 1;
strbuf_reset(&s->buf);
if (pipe_command(&filter_cp,
colored->buf, colored->len,
&s->buf, colored->len,
NULL, 0) < 0)
return error(_("failed to run '%s'"),
diff_filter);
strbuf_swap(colored, &s->buf);
}
strbuf_complete_line(colored);
colored_p = colored->buf;
colored_pend = colored_p + colored->len;
}
strvec_clear(&args);
/* parse files and hunks */
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
p = plain->buf;
pend = p + plain->len;
while (p != pend) {
char *eol = memchr(p, '\n', pend - p);
add -p: fix editing of intent-to-add paths A popular way of partially staging a new file is to run `git add -N <path>` and then use the hunk editing of `git add -p` to select the part of the file that the user wishes to stage. Since 85953a3187 ("diff-files --raw: show correct post-image of intent-to-add files", 2020-07-01) this has stopped working as intent-to-add paths are now show as new files rather than changes to an empty blob and `git apply` refused to apply a creation patch for a path that was marked as intent-to-add. 7cfde3fa0f ("apply: allow "new file" patches on i-t-a entries", 2020-08-06) fixed the problem with apply but it still wasn't possible to edit the added hunk properly. 2c8bd8471a ("checkout -p: handle new files correctly", 2020-05-27) had previously changed `add -p` to handle new files but it did not implement patch editing correctly. The perl version simply forbade editing and the C version opened the editor with the full diff rather that just the hunk which meant that the user had to edit the hunk header manually to get it to work. The root cause of the problem is that added files store the diff header with the hunk data rather than separating the two as we do for other changes. Changing added files to store the diff header separately fixes the editing problem at the expense of having to special case empty additions as they no longer have any hunks associated with them, only the diff header. The changes move some existing code into a conditional changing the indentation, they are best viewed with --color-moved-ws=allow-indentation-change (or --ignore-space-change works well to get an overview of the changes) Signed-off-by: Phillip Wood <phillip.wood@dunelm.org.uk> Reported-by: Thomas Sullivan <tom@msbit.com.au> Reported-by: Yuchen Ying <ych@google.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-09-09 16:58:52 +03:00
const char *deleted = NULL, *mode_change = NULL;
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
if (!eol)
eol = pend;
if (starts_with(p, "diff ")) {
ALLOC_GROW_BY(s->file_diff, s->file_diff_nr, 1,
file_diff_alloc);
file_diff = s->file_diff + s->file_diff_nr - 1;
hunk = &file_diff->head;
hunk->start = p - plain->buf;
if (colored_p)
hunk->colored_start = colored_p - colored->buf;
marker = '\0';
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
} else if (p == plain->buf)
BUG("diff starts with unexpected line:\n"
"%.*s\n", (int)(eol - p), p);
add -p: fix editing of intent-to-add paths A popular way of partially staging a new file is to run `git add -N <path>` and then use the hunk editing of `git add -p` to select the part of the file that the user wishes to stage. Since 85953a3187 ("diff-files --raw: show correct post-image of intent-to-add files", 2020-07-01) this has stopped working as intent-to-add paths are now show as new files rather than changes to an empty blob and `git apply` refused to apply a creation patch for a path that was marked as intent-to-add. 7cfde3fa0f ("apply: allow "new file" patches on i-t-a entries", 2020-08-06) fixed the problem with apply but it still wasn't possible to edit the added hunk properly. 2c8bd8471a ("checkout -p: handle new files correctly", 2020-05-27) had previously changed `add -p` to handle new files but it did not implement patch editing correctly. The perl version simply forbade editing and the C version opened the editor with the full diff rather that just the hunk which meant that the user had to edit the hunk header manually to get it to work. The root cause of the problem is that added files store the diff header with the hunk data rather than separating the two as we do for other changes. Changing added files to store the diff header separately fixes the editing problem at the expense of having to special case empty additions as they no longer have any hunks associated with them, only the diff header. The changes move some existing code into a conditional changing the indentation, they are best viewed with --color-moved-ws=allow-indentation-change (or --ignore-space-change works well to get an overview of the changes) Signed-off-by: Phillip Wood <phillip.wood@dunelm.org.uk> Reported-by: Thomas Sullivan <tom@msbit.com.au> Reported-by: Yuchen Ying <ych@google.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-09-09 16:58:52 +03:00
else if (file_diff->deleted)
; /* keep the rest of the file in a single "hunk" */
else if (starts_with(p, "@@ ") ||
(hunk == &file_diff->head &&
add -p: fix editing of intent-to-add paths A popular way of partially staging a new file is to run `git add -N <path>` and then use the hunk editing of `git add -p` to select the part of the file that the user wishes to stage. Since 85953a3187 ("diff-files --raw: show correct post-image of intent-to-add files", 2020-07-01) this has stopped working as intent-to-add paths are now show as new files rather than changes to an empty blob and `git apply` refused to apply a creation patch for a path that was marked as intent-to-add. 7cfde3fa0f ("apply: allow "new file" patches on i-t-a entries", 2020-08-06) fixed the problem with apply but it still wasn't possible to edit the added hunk properly. 2c8bd8471a ("checkout -p: handle new files correctly", 2020-05-27) had previously changed `add -p` to handle new files but it did not implement patch editing correctly. The perl version simply forbade editing and the C version opened the editor with the full diff rather that just the hunk which meant that the user had to edit the hunk header manually to get it to work. The root cause of the problem is that added files store the diff header with the hunk data rather than separating the two as we do for other changes. Changing added files to store the diff header separately fixes the editing problem at the expense of having to special case empty additions as they no longer have any hunks associated with them, only the diff header. The changes move some existing code into a conditional changing the indentation, they are best viewed with --color-moved-ws=allow-indentation-change (or --ignore-space-change works well to get an overview of the changes) Signed-off-by: Phillip Wood <phillip.wood@dunelm.org.uk> Reported-by: Thomas Sullivan <tom@msbit.com.au> Reported-by: Yuchen Ying <ych@google.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-09-09 16:58:52 +03:00
(skip_prefix(p, "deleted file", &deleted)))) {
if (marker == '-' || marker == '+')
/*
* Should not happen; previous hunk did not end
* in a context line? Handle it anyway.
*/
hunk->splittable_into++;
ALLOC_GROW_BY(file_diff->hunk, file_diff->hunk_nr, 1,
file_diff->hunk_alloc);
hunk = file_diff->hunk + file_diff->hunk_nr - 1;
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
hunk->start = p - plain->buf;
if (colored)
hunk->colored_start = colored_p - colored->buf;
built-in add -p: adjust hunk headers as needed When skipping a hunk that adds a different number of lines than it removes, we need to adjust the subsequent hunk headers of non-skipped hunks: in pathological cases, the context is not enough to determine precisely where the patch should be applied. This problem was identified in 23fea4c240 (t3701: add failing test for pathological context lines, 2018-03-01) and fixed in the Perl version in fecc6f3a68 (add -p: adjust offsets of subsequent hunks when one is skipped, 2018-03-01). And this patch fixes it in the C version of `git add -p`. In contrast to the Perl version, we try to keep the extra text on the hunk header (which typically contains the signature of the function whose code is changed in the hunk) intact. Note: while the C version does not support staging mode changes at this stage, we already prepare for this by simply skipping the hunk header if both old and new offset is 0 (this cannot happen for regular hunks, and we will use this as an indicator that we are looking at a special hunk). Likewise, we already prepare for hunk splitting by handling the absence of extra text in the hunk header gracefully: only the first split hunk will have that text, the others will not (indicated by an empty extra text start/end range). Preparing for hunk splitting already at this stage avoids an indentation change of the entire hunk header-printing block later, and is almost as easy to review as without that handling. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:51 +03:00
if (deleted)
file_diff->deleted = 1;
else if (parse_hunk_header(s, hunk) < 0)
built-in add -p: adjust hunk headers as needed When skipping a hunk that adds a different number of lines than it removes, we need to adjust the subsequent hunk headers of non-skipped hunks: in pathological cases, the context is not enough to determine precisely where the patch should be applied. This problem was identified in 23fea4c240 (t3701: add failing test for pathological context lines, 2018-03-01) and fixed in the Perl version in fecc6f3a68 (add -p: adjust offsets of subsequent hunks when one is skipped, 2018-03-01). And this patch fixes it in the C version of `git add -p`. In contrast to the Perl version, we try to keep the extra text on the hunk header (which typically contains the signature of the function whose code is changed in the hunk) intact. Note: while the C version does not support staging mode changes at this stage, we already prepare for this by simply skipping the hunk header if both old and new offset is 0 (this cannot happen for regular hunks, and we will use this as an indicator that we are looking at a special hunk). Likewise, we already prepare for hunk splitting by handling the absence of extra text in the hunk header gracefully: only the first split hunk will have that text, the others will not (indicated by an empty extra text start/end range). Preparing for hunk splitting already at this stage avoids an indentation change of the entire hunk header-printing block later, and is almost as easy to review as without that handling. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:51 +03:00
return -1;
/*
* Start counting into how many hunks this one can be
* split
*/
marker = *p;
add -p: fix editing of intent-to-add paths A popular way of partially staging a new file is to run `git add -N <path>` and then use the hunk editing of `git add -p` to select the part of the file that the user wishes to stage. Since 85953a3187 ("diff-files --raw: show correct post-image of intent-to-add files", 2020-07-01) this has stopped working as intent-to-add paths are now show as new files rather than changes to an empty blob and `git apply` refused to apply a creation patch for a path that was marked as intent-to-add. 7cfde3fa0f ("apply: allow "new file" patches on i-t-a entries", 2020-08-06) fixed the problem with apply but it still wasn't possible to edit the added hunk properly. 2c8bd8471a ("checkout -p: handle new files correctly", 2020-05-27) had previously changed `add -p` to handle new files but it did not implement patch editing correctly. The perl version simply forbade editing and the C version opened the editor with the full diff rather that just the hunk which meant that the user had to edit the hunk header manually to get it to work. The root cause of the problem is that added files store the diff header with the hunk data rather than separating the two as we do for other changes. Changing added files to store the diff header separately fixes the editing problem at the expense of having to special case empty additions as they no longer have any hunks associated with them, only the diff header. The changes move some existing code into a conditional changing the indentation, they are best viewed with --color-moved-ws=allow-indentation-change (or --ignore-space-change works well to get an overview of the changes) Signed-off-by: Phillip Wood <phillip.wood@dunelm.org.uk> Reported-by: Thomas Sullivan <tom@msbit.com.au> Reported-by: Yuchen Ying <ych@google.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-09-09 16:58:52 +03:00
} else if (hunk == &file_diff->head &&
starts_with(p, "new file")) {
file_diff->added = 1;
} else if (hunk == &file_diff->head &&
skip_prefix(p, "old mode ", &mode_change) &&
is_octal(mode_change, eol - mode_change)) {
if (file_diff->mode_change)
BUG("double mode change?\n\n%.*s",
(int)(eol - plain->buf), plain->buf);
if (file_diff->hunk_nr)
BUG("mode change in the middle?\n\n%.*s",
(int)(eol - plain->buf), plain->buf);
/*
* Do *not* change `hunk`: the mode change pseudo-hunk
* is _part of_ the header "hunk".
*/
file_diff->mode_change = 1;
ALLOC_GROW_BY(file_diff->hunk, file_diff->hunk_nr, 1,
file_diff->hunk_alloc);
file_diff->hunk->start = p - plain->buf;
if (colored_p)
file_diff->hunk->colored_start =
colored_p - colored->buf;
} else if (hunk == &file_diff->head &&
skip_prefix(p, "new mode ", &mode_change) &&
is_octal(mode_change, eol - mode_change)) {
/*
* Extend the "mode change" pseudo-hunk to include also
* the "new mode" line.
*/
if (!file_diff->mode_change)
BUG("'new mode' without 'old mode'?\n\n%.*s",
(int)(eol - plain->buf), plain->buf);
if (file_diff->hunk_nr != 1)
BUG("mode change in the middle?\n\n%.*s",
(int)(eol - plain->buf), plain->buf);
if (p - plain->buf != file_diff->hunk->end)
BUG("'new mode' does not immediately follow "
"'old mode'?\n\n%.*s",
(int)(eol - plain->buf), plain->buf);
} else if (hunk == &file_diff->head &&
starts_with(p, "Binary files "))
file_diff->binary = 1;
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
if (!!file_diff->deleted + !!file_diff->added +
!!file_diff->mode_change > 1)
BUG("diff can only contain delete *or* add *or* a "
"mode change?!?\n%.*s",
(int)(eol - (plain->buf + file_diff->head.start)),
plain->buf + file_diff->head.start);
if ((marker == '-' || marker == '+') && *p == ' ')
hunk->splittable_into++;
if (marker && *p != '\\')
marker = *p;
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
p = eol == pend ? pend : eol + 1;
hunk->end = p - plain->buf;
if (colored) {
char *colored_eol = memchr(colored_p, '\n',
colored_pend - colored_p);
if (colored_eol)
colored_p = colored_eol + 1;
else if (p != pend)
/* colored shorter than non-colored? */
goto mismatched_output;
else
colored_p = colored_pend;
hunk->colored_end = colored_p - colored->buf;
}
if (mode_change) {
if (file_diff->hunk_nr != 1)
BUG("mode change in hunk #%d???",
(int)file_diff->hunk_nr);
/* Adjust the end of the "mode change" pseudo-hunk */
file_diff->hunk->end = hunk->end;
if (colored)
file_diff->hunk->colored_end = hunk->colored_end;
}
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
}
if (marker == '-' || marker == '+')
/*
* Last hunk ended in non-context line (i.e. it appended lines
* to the file, so there are no trailing context lines).
*/
hunk->splittable_into++;
/* non-colored shorter than colored? */
if (colored_p != colored_pend) {
mismatched_output:
error(_("mismatched output from interactive.diffFilter"));
advise(_("Your filter must maintain a one-to-one correspondence\n"
"between its input and output lines."));
return -1;
}
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
return 0;
}
static size_t find_next_line(struct strbuf *sb, size_t offset)
{
char *eol;
if (offset >= sb->len)
BUG("looking for next line beyond buffer (%d >= %d)\n%s",
(int)offset, (int)sb->len, sb->buf);
eol = memchr(sb->buf + offset, '\n', sb->len - offset);
if (!eol)
return sb->len;
return eol - sb->buf + 1;
}
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
static void render_hunk(struct add_p_state *s, struct hunk *hunk,
built-in add -p: adjust hunk headers as needed When skipping a hunk that adds a different number of lines than it removes, we need to adjust the subsequent hunk headers of non-skipped hunks: in pathological cases, the context is not enough to determine precisely where the patch should be applied. This problem was identified in 23fea4c240 (t3701: add failing test for pathological context lines, 2018-03-01) and fixed in the Perl version in fecc6f3a68 (add -p: adjust offsets of subsequent hunks when one is skipped, 2018-03-01). And this patch fixes it in the C version of `git add -p`. In contrast to the Perl version, we try to keep the extra text on the hunk header (which typically contains the signature of the function whose code is changed in the hunk) intact. Note: while the C version does not support staging mode changes at this stage, we already prepare for this by simply skipping the hunk header if both old and new offset is 0 (this cannot happen for regular hunks, and we will use this as an indicator that we are looking at a special hunk). Likewise, we already prepare for hunk splitting by handling the absence of extra text in the hunk header gracefully: only the first split hunk will have that text, the others will not (indicated by an empty extra text start/end range). Preparing for hunk splitting already at this stage avoids an indentation change of the entire hunk header-printing block later, and is almost as easy to review as without that handling. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:51 +03:00
ssize_t delta, int colored, struct strbuf *out)
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
{
built-in add -p: adjust hunk headers as needed When skipping a hunk that adds a different number of lines than it removes, we need to adjust the subsequent hunk headers of non-skipped hunks: in pathological cases, the context is not enough to determine precisely where the patch should be applied. This problem was identified in 23fea4c240 (t3701: add failing test for pathological context lines, 2018-03-01) and fixed in the Perl version in fecc6f3a68 (add -p: adjust offsets of subsequent hunks when one is skipped, 2018-03-01). And this patch fixes it in the C version of `git add -p`. In contrast to the Perl version, we try to keep the extra text on the hunk header (which typically contains the signature of the function whose code is changed in the hunk) intact. Note: while the C version does not support staging mode changes at this stage, we already prepare for this by simply skipping the hunk header if both old and new offset is 0 (this cannot happen for regular hunks, and we will use this as an indicator that we are looking at a special hunk). Likewise, we already prepare for hunk splitting by handling the absence of extra text in the hunk header gracefully: only the first split hunk will have that text, the others will not (indicated by an empty extra text start/end range). Preparing for hunk splitting already at this stage avoids an indentation change of the entire hunk header-printing block later, and is almost as easy to review as without that handling. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:51 +03:00
struct hunk_header *header = &hunk->header;
if (hunk->header.old_offset != 0 || hunk->header.new_offset != 0) {
/*
* Generate the hunk header dynamically, except for special
* hunks (such as the diff header).
*/
const char *p;
size_t len;
unsigned long old_offset = header->old_offset;
unsigned long new_offset = header->new_offset;
if (!colored) {
p = s->plain.buf + header->extra_start;
len = header->extra_end - header->extra_start;
} else {
strbuf_addstr(out, s->s.fraginfo_color);
p = s->colored.buf + header->colored_extra_start;
len = header->colored_extra_end
- header->colored_extra_start;
}
if (s->mode->is_reverse)
old_offset -= delta;
else
new_offset += delta;
built-in add -p: adjust hunk headers as needed When skipping a hunk that adds a different number of lines than it removes, we need to adjust the subsequent hunk headers of non-skipped hunks: in pathological cases, the context is not enough to determine precisely where the patch should be applied. This problem was identified in 23fea4c240 (t3701: add failing test for pathological context lines, 2018-03-01) and fixed in the Perl version in fecc6f3a68 (add -p: adjust offsets of subsequent hunks when one is skipped, 2018-03-01). And this patch fixes it in the C version of `git add -p`. In contrast to the Perl version, we try to keep the extra text on the hunk header (which typically contains the signature of the function whose code is changed in the hunk) intact. Note: while the C version does not support staging mode changes at this stage, we already prepare for this by simply skipping the hunk header if both old and new offset is 0 (this cannot happen for regular hunks, and we will use this as an indicator that we are looking at a special hunk). Likewise, we already prepare for hunk splitting by handling the absence of extra text in the hunk header gracefully: only the first split hunk will have that text, the others will not (indicated by an empty extra text start/end range). Preparing for hunk splitting already at this stage avoids an indentation change of the entire hunk header-printing block later, and is almost as easy to review as without that handling. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:51 +03:00
strbuf_addf(out, "@@ -%lu", old_offset);
if (header->old_count != 1)
strbuf_addf(out, ",%lu", header->old_count);
strbuf_addf(out, " +%lu", new_offset);
if (header->new_count != 1)
strbuf_addf(out, ",%lu", header->new_count);
strbuf_addstr(out, " @@");
built-in add -p: adjust hunk headers as needed When skipping a hunk that adds a different number of lines than it removes, we need to adjust the subsequent hunk headers of non-skipped hunks: in pathological cases, the context is not enough to determine precisely where the patch should be applied. This problem was identified in 23fea4c240 (t3701: add failing test for pathological context lines, 2018-03-01) and fixed in the Perl version in fecc6f3a68 (add -p: adjust offsets of subsequent hunks when one is skipped, 2018-03-01). And this patch fixes it in the C version of `git add -p`. In contrast to the Perl version, we try to keep the extra text on the hunk header (which typically contains the signature of the function whose code is changed in the hunk) intact. Note: while the C version does not support staging mode changes at this stage, we already prepare for this by simply skipping the hunk header if both old and new offset is 0 (this cannot happen for regular hunks, and we will use this as an indicator that we are looking at a special hunk). Likewise, we already prepare for hunk splitting by handling the absence of extra text in the hunk header gracefully: only the first split hunk will have that text, the others will not (indicated by an empty extra text start/end range). Preparing for hunk splitting already at this stage avoids an indentation change of the entire hunk header-printing block later, and is almost as easy to review as without that handling. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:51 +03:00
if (len)
strbuf_add(out, p, len);
else if (colored)
strbuf_addf(out, "%s\n", s->s.reset_color);
built-in add -p: adjust hunk headers as needed When skipping a hunk that adds a different number of lines than it removes, we need to adjust the subsequent hunk headers of non-skipped hunks: in pathological cases, the context is not enough to determine precisely where the patch should be applied. This problem was identified in 23fea4c240 (t3701: add failing test for pathological context lines, 2018-03-01) and fixed in the Perl version in fecc6f3a68 (add -p: adjust offsets of subsequent hunks when one is skipped, 2018-03-01). And this patch fixes it in the C version of `git add -p`. In contrast to the Perl version, we try to keep the extra text on the hunk header (which typically contains the signature of the function whose code is changed in the hunk) intact. Note: while the C version does not support staging mode changes at this stage, we already prepare for this by simply skipping the hunk header if both old and new offset is 0 (this cannot happen for regular hunks, and we will use this as an indicator that we are looking at a special hunk). Likewise, we already prepare for hunk splitting by handling the absence of extra text in the hunk header gracefully: only the first split hunk will have that text, the others will not (indicated by an empty extra text start/end range). Preparing for hunk splitting already at this stage avoids an indentation change of the entire hunk header-printing block later, and is almost as easy to review as without that handling. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:51 +03:00
else
strbuf_addch(out, '\n');
}
if (colored)
strbuf_add(out, s->colored.buf + hunk->colored_start,
hunk->colored_end - hunk->colored_start);
else
strbuf_add(out, s->plain.buf + hunk->start,
hunk->end - hunk->start);
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
}
static void render_diff_header(struct add_p_state *s,
struct file_diff *file_diff, int colored,
struct strbuf *out)
{
/*
* If there was a mode change, the first hunk is a pseudo hunk that
* corresponds to the mode line in the header. If the user did not want
* to stage that "hunk", we actually have to cut it out from the header.
*/
int skip_mode_change =
file_diff->mode_change && file_diff->hunk->use != USE_HUNK;
struct hunk *head = &file_diff->head, *first = file_diff->hunk;
if (!skip_mode_change) {
render_hunk(s, head, 0, colored, out);
return;
}
if (colored) {
const char *p = s->colored.buf;
strbuf_add(out, p + head->colored_start,
first->colored_start - head->colored_start);
strbuf_add(out, p + first->colored_end,
head->colored_end - first->colored_end);
} else {
const char *p = s->plain.buf;
strbuf_add(out, p + head->start, first->start - head->start);
strbuf_add(out, p + first->end, head->end - first->end);
}
}
/* Coalesce hunks again that were split */
static int merge_hunks(struct add_p_state *s, struct file_diff *file_diff,
size_t *hunk_index, int use_all, struct hunk *merged)
{
size_t i = *hunk_index, delta;
struct hunk *hunk = file_diff->hunk + i;
/* `header` corresponds to the merged hunk */
struct hunk_header *header = &merged->header, *next;
if (!use_all && hunk->use != USE_HUNK)
return 0;
*merged = *hunk;
/* We simply skip the colored part (if any) when merging hunks */
merged->colored_start = merged->colored_end = 0;
for (; i + 1 < file_diff->hunk_nr; i++) {
hunk++;
next = &hunk->header;
/*
* Stop merging hunks when:
*
* - the hunk is not selected for use, or
* - the hunk does not overlap with the already-merged hunk(s)
*/
if ((!use_all && hunk->use != USE_HUNK) ||
header->new_offset >= next->new_offset + merged->delta ||
header->new_offset + header->new_count
< next->new_offset + merged->delta)
break;
/*
* If the hunks were not edited, and overlap, we can simply
* extend the line range.
*/
if (merged->start < hunk->start && merged->end > hunk->start) {
merged->end = hunk->end;
merged->colored_end = hunk->colored_end;
delta = 0;
} else {
const char *plain = s->plain.buf;
size_t overlapping_line_count = header->new_offset
+ header->new_count - merged->delta
- next->new_offset;
size_t overlap_end = hunk->start;
size_t overlap_start = overlap_end;
size_t overlap_next, len, j;
/*
* One of the hunks was edited: the modified hunk was
* appended to the strbuf `s->plain`.
*
* Let's ensure that at least the last context line of
* the first hunk overlaps with the corresponding line
* of the second hunk, and then merge.
*/
for (j = 0; j < overlapping_line_count; j++) {
overlap_next = find_next_line(&s->plain,
overlap_end);
if (overlap_next > hunk->end)
BUG("failed to find %d context lines "
"in:\n%.*s",
(int)overlapping_line_count,
(int)(hunk->end - hunk->start),
plain + hunk->start);
if (plain[overlap_end] != ' ')
return error(_("expected context line "
"#%d in\n%.*s"),
(int)(j + 1),
(int)(hunk->end
- hunk->start),
plain + hunk->start);
overlap_start = overlap_end;
overlap_end = overlap_next;
}
len = overlap_end - overlap_start;
if (len > merged->end - merged->start ||
memcmp(plain + merged->end - len,
plain + overlap_start, len))
return error(_("hunks do not overlap:\n%.*s\n"
"\tdoes not end with:\n%.*s"),
(int)(merged->end - merged->start),
plain + merged->start,
(int)len, plain + overlap_start);
/*
* Since the start-end ranges are not adjacent, we
* cannot simply take the union of the ranges. To
* address that, we temporarily append the union of the
* lines to the `plain` strbuf.
*/
if (merged->end != s->plain.len) {
size_t start = s->plain.len;
strbuf_add(&s->plain, plain + merged->start,
merged->end - merged->start);
plain = s->plain.buf;
merged->start = start;
merged->end = s->plain.len;
}
strbuf_add(&s->plain,
plain + overlap_end,
hunk->end - overlap_end);
merged->end = s->plain.len;
merged->splittable_into += hunk->splittable_into;
delta = merged->delta;
merged->delta += hunk->delta;
}
header->old_count = next->old_offset + next->old_count
- header->old_offset;
header->new_count = next->new_offset + delta
+ next->new_count - header->new_offset;
}
if (i == *hunk_index)
return 0;
*hunk_index = i;
return 1;
}
static void reassemble_patch(struct add_p_state *s,
struct file_diff *file_diff, int use_all,
struct strbuf *out)
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
{
struct hunk *hunk;
size_t save_len = s->plain.len, i;
built-in add -p: adjust hunk headers as needed When skipping a hunk that adds a different number of lines than it removes, we need to adjust the subsequent hunk headers of non-skipped hunks: in pathological cases, the context is not enough to determine precisely where the patch should be applied. This problem was identified in 23fea4c240 (t3701: add failing test for pathological context lines, 2018-03-01) and fixed in the Perl version in fecc6f3a68 (add -p: adjust offsets of subsequent hunks when one is skipped, 2018-03-01). And this patch fixes it in the C version of `git add -p`. In contrast to the Perl version, we try to keep the extra text on the hunk header (which typically contains the signature of the function whose code is changed in the hunk) intact. Note: while the C version does not support staging mode changes at this stage, we already prepare for this by simply skipping the hunk header if both old and new offset is 0 (this cannot happen for regular hunks, and we will use this as an indicator that we are looking at a special hunk). Likewise, we already prepare for hunk splitting by handling the absence of extra text in the hunk header gracefully: only the first split hunk will have that text, the others will not (indicated by an empty extra text start/end range). Preparing for hunk splitting already at this stage avoids an indentation change of the entire hunk header-printing block later, and is almost as easy to review as without that handling. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:51 +03:00
ssize_t delta = 0;
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
render_diff_header(s, file_diff, 0, out);
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
for (i = file_diff->mode_change; i < file_diff->hunk_nr; i++) {
struct hunk merged = { 0 };
hunk = file_diff->hunk + i;
if (!use_all && hunk->use != USE_HUNK)
built-in add -p: adjust hunk headers as needed When skipping a hunk that adds a different number of lines than it removes, we need to adjust the subsequent hunk headers of non-skipped hunks: in pathological cases, the context is not enough to determine precisely where the patch should be applied. This problem was identified in 23fea4c240 (t3701: add failing test for pathological context lines, 2018-03-01) and fixed in the Perl version in fecc6f3a68 (add -p: adjust offsets of subsequent hunks when one is skipped, 2018-03-01). And this patch fixes it in the C version of `git add -p`. In contrast to the Perl version, we try to keep the extra text on the hunk header (which typically contains the signature of the function whose code is changed in the hunk) intact. Note: while the C version does not support staging mode changes at this stage, we already prepare for this by simply skipping the hunk header if both old and new offset is 0 (this cannot happen for regular hunks, and we will use this as an indicator that we are looking at a special hunk). Likewise, we already prepare for hunk splitting by handling the absence of extra text in the hunk header gracefully: only the first split hunk will have that text, the others will not (indicated by an empty extra text start/end range). Preparing for hunk splitting already at this stage avoids an indentation change of the entire hunk header-printing block later, and is almost as easy to review as without that handling. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:51 +03:00
delta += hunk->header.old_count
- hunk->header.new_count;
else {
/* merge overlapping hunks into a temporary hunk */
if (merge_hunks(s, file_diff, &i, use_all, &merged))
hunk = &merged;
built-in add -p: adjust hunk headers as needed When skipping a hunk that adds a different number of lines than it removes, we need to adjust the subsequent hunk headers of non-skipped hunks: in pathological cases, the context is not enough to determine precisely where the patch should be applied. This problem was identified in 23fea4c240 (t3701: add failing test for pathological context lines, 2018-03-01) and fixed in the Perl version in fecc6f3a68 (add -p: adjust offsets of subsequent hunks when one is skipped, 2018-03-01). And this patch fixes it in the C version of `git add -p`. In contrast to the Perl version, we try to keep the extra text on the hunk header (which typically contains the signature of the function whose code is changed in the hunk) intact. Note: while the C version does not support staging mode changes at this stage, we already prepare for this by simply skipping the hunk header if both old and new offset is 0 (this cannot happen for regular hunks, and we will use this as an indicator that we are looking at a special hunk). Likewise, we already prepare for hunk splitting by handling the absence of extra text in the hunk header gracefully: only the first split hunk will have that text, the others will not (indicated by an empty extra text start/end range). Preparing for hunk splitting already at this stage avoids an indentation change of the entire hunk header-printing block later, and is almost as easy to review as without that handling. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:51 +03:00
render_hunk(s, hunk, delta, 0, out);
/*
* In case `merge_hunks()` used `plain` as a scratch
* pad (this happens when an edited hunk had to be
* coalesced with another hunk).
*/
strbuf_setlen(&s->plain, save_len);
delta += hunk->delta;
}
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
}
}
static int split_hunk(struct add_p_state *s, struct file_diff *file_diff,
size_t hunk_index)
{
int colored = !!s->colored.len, first = 1;
struct hunk *hunk = file_diff->hunk + hunk_index;
size_t splittable_into;
size_t end, colored_end, current, colored_current = 0, context_line_count;
struct hunk_header remaining, *header;
char marker, ch;
if (hunk_index >= file_diff->hunk_nr)
BUG("invalid hunk index: %d (must be >= 0 and < %d)",
(int)hunk_index, (int)file_diff->hunk_nr);
if (hunk->splittable_into < 2)
return 0;
splittable_into = hunk->splittable_into;
end = hunk->end;
colored_end = hunk->colored_end;
remaining = hunk->header;
file_diff->hunk_nr += splittable_into - 1;
ALLOC_GROW(file_diff->hunk, file_diff->hunk_nr, file_diff->hunk_alloc);
if (hunk_index + splittable_into < file_diff->hunk_nr)
memmove(file_diff->hunk + hunk_index + splittable_into,
file_diff->hunk + hunk_index + 1,
(file_diff->hunk_nr - hunk_index - splittable_into)
* sizeof(*hunk));
hunk = file_diff->hunk + hunk_index;
hunk->splittable_into = 1;
memset(hunk + 1, 0, (splittable_into - 1) * sizeof(*hunk));
header = &hunk->header;
header->old_count = header->new_count = 0;
current = hunk->start;
if (colored)
colored_current = hunk->colored_start;
marker = '\0';
context_line_count = 0;
while (splittable_into > 1) {
ch = s->plain.buf[current];
if (!ch)
BUG("buffer overrun while splitting hunks");
/*
* Is this the first context line after a chain of +/- lines?
* Then record the start of the next split hunk.
*/
if ((marker == '-' || marker == '+') && ch == ' ') {
first = 0;
hunk[1].start = current;
if (colored)
hunk[1].colored_start = colored_current;
context_line_count = 0;
}
/*
* Was the previous line a +/- one? Alternatively, is this the
* first line (and not a +/- one)?
*
* Then just increment the appropriate counter and continue
* with the next line.
*/
if (marker != ' ' || (ch != '-' && ch != '+')) {
next_hunk_line:
/* Comment lines are attached to the previous line */
if (ch == '\\')
ch = marker ? marker : ' ';
/* current hunk not done yet */
if (ch == ' ')
context_line_count++;
else if (ch == '-')
header->old_count++;
else if (ch == '+')
header->new_count++;
else
BUG("unhandled diff marker: '%c'", ch);
marker = ch;
current = find_next_line(&s->plain, current);
if (colored)
colored_current =
find_next_line(&s->colored,
colored_current);
continue;
}
/*
* We got us the start of a new hunk!
*
* This is a context line, so it is shared with the previous
* hunk, if any.
*/
if (first) {
if (header->old_count || header->new_count)
BUG("counts are off: %d/%d",
(int)header->old_count,
(int)header->new_count);
header->old_count = context_line_count;
header->new_count = context_line_count;
context_line_count = 0;
first = 0;
goto next_hunk_line;
}
remaining.old_offset += header->old_count;
remaining.old_count -= header->old_count;
remaining.new_offset += header->new_count;
remaining.new_count -= header->new_count;
/* initialize next hunk header's offsets */
hunk[1].header.old_offset =
header->old_offset + header->old_count;
hunk[1].header.new_offset =
header->new_offset + header->new_count;
/* add one split hunk */
header->old_count += context_line_count;
header->new_count += context_line_count;
hunk->end = current;
if (colored)
hunk->colored_end = colored_current;
hunk++;
hunk->splittable_into = 1;
hunk->use = hunk[-1].use;
header = &hunk->header;
header->old_count = header->new_count = context_line_count;
context_line_count = 0;
splittable_into--;
marker = ch;
}
/* last hunk simply gets the rest */
if (header->old_offset != remaining.old_offset)
BUG("miscounted old_offset: %lu != %lu",
header->old_offset, remaining.old_offset);
if (header->new_offset != remaining.new_offset)
BUG("miscounted new_offset: %lu != %lu",
header->new_offset, remaining.new_offset);
header->old_count = remaining.old_count;
header->new_count = remaining.new_count;
hunk->end = end;
if (colored)
hunk->colored_end = colored_end;
return 0;
}
static void recolor_hunk(struct add_p_state *s, struct hunk *hunk)
{
const char *plain = s->plain.buf;
size_t current, eol, next;
if (!s->colored.len)
return;
hunk->colored_start = s->colored.len;
for (current = hunk->start; current < hunk->end; ) {
for (eol = current; eol < hunk->end; eol++)
if (plain[eol] == '\n')
break;
next = eol + (eol < hunk->end);
if (eol > current && plain[eol - 1] == '\r')
eol--;
strbuf_addstr(&s->colored,
plain[current] == '-' ?
s->s.file_old_color :
plain[current] == '+' ?
s->s.file_new_color :
s->s.context_color);
strbuf_add(&s->colored, plain + current, eol - current);
strbuf_addstr(&s->colored, s->s.reset_color);
if (next > eol)
strbuf_add(&s->colored, plain + eol, next - eol);
current = next;
}
hunk->colored_end = s->colored.len;
}
static int edit_hunk_manually(struct add_p_state *s, struct hunk *hunk)
{
size_t i;
strbuf_reset(&s->buf);
strbuf_commented_addf(&s->buf, _("Manual hunk edit mode -- see bottom for "
"a quick guide.\n"));
render_hunk(s, hunk, 0, 0, &s->buf);
strbuf_commented_addf(&s->buf,
_("---\n"
"To remove '%c' lines, make them ' ' lines "
"(context).\n"
"To remove '%c' lines, delete them.\n"
"Lines starting with %c will be removed.\n"),
s->mode->is_reverse ? '+' : '-',
s->mode->is_reverse ? '-' : '+',
comment_line_char);
strbuf_commented_addf(&s->buf, "%s", _(s->mode->edit_hunk_hint));
/*
* TRANSLATORS: 'it' refers to the patch mentioned in the previous
* messages.
*/
strbuf_commented_addf(&s->buf,
_("If it does not apply cleanly, you will be "
"given an opportunity to\n"
"edit again. If all lines of the hunk are "
"removed, then the edit is\n"
"aborted and the hunk is left unchanged.\n"));
if (strbuf_edit_interactively(&s->buf, "addp-hunk-edit.diff", NULL) < 0)
return -1;
/* strip out commented lines */
hunk->start = s->plain.len;
for (i = 0; i < s->buf.len; ) {
size_t next = find_next_line(&s->buf, i);
if (s->buf.buf[i] != comment_line_char)
strbuf_add(&s->plain, s->buf.buf + i, next - i);
i = next;
}
hunk->end = s->plain.len;
if (hunk->end == hunk->start)
/* The user aborted editing by deleting everything */
return 0;
recolor_hunk(s, hunk);
/*
* If the hunk header is intact, parse it, otherwise simply use the
* hunk header prior to editing (which will adjust `hunk->start` to
* skip the hunk header).
*/
if (s->plain.buf[hunk->start] == '@' &&
parse_hunk_header(s, hunk) < 0)
return error(_("could not parse hunk header"));
return 1;
}
static ssize_t recount_edited_hunk(struct add_p_state *s, struct hunk *hunk,
size_t orig_old_count, size_t orig_new_count)
{
struct hunk_header *header = &hunk->header;
size_t i;
header->old_count = header->new_count = 0;
for (i = hunk->start; i < hunk->end; ) {
switch (s->plain.buf[i]) {
case '-':
header->old_count++;
break;
case '+':
header->new_count++;
break;
case ' ': case '\r': case '\n':
header->old_count++;
header->new_count++;
break;
}
i = find_next_line(&s->plain, i);
}
return orig_old_count - orig_new_count
- header->old_count + header->new_count;
}
static int run_apply_check(struct add_p_state *s,
struct file_diff *file_diff)
{
struct child_process cp = CHILD_PROCESS_INIT;
strbuf_reset(&s->buf);
reassemble_patch(s, file_diff, 1, &s->buf);
setup_child_process(s, &cp,
"apply", "--check", NULL);
strvec_pushv(&cp.args, s->mode->apply_check_args);
if (pipe_command(&cp, s->buf.buf, s->buf.len, NULL, 0, NULL, 0))
return error(_("'git apply --cached' failed"));
return 0;
}
static int read_single_character(struct add_p_state *s)
{
if (s->s.use_single_key) {
int res = read_key_without_echo(&s->answer);
printf("%s\n", res == EOF ? "" : s->answer.buf);
return res;
}
if (git_read_line_interactively(&s->answer) == EOF)
return EOF;
return 0;
}
static int prompt_yesno(struct add_p_state *s, const char *prompt)
{
for (;;) {
color_fprintf(stdout, s->s.prompt_color, "%s", _(prompt));
fflush(stdout);
if (read_single_character(s) == EOF)
return -1;
switch (tolower(s->answer.buf[0])) {
case 'n': return 0;
case 'y': return 1;
}
}
}
static int edit_hunk_loop(struct add_p_state *s,
struct file_diff *file_diff, struct hunk *hunk)
{
size_t plain_len = s->plain.len, colored_len = s->colored.len;
struct hunk backup;
backup = *hunk;
for (;;) {
int res = edit_hunk_manually(s, hunk);
if (res == 0) {
/* abandoned */
*hunk = backup;
return -1;
}
if (res > 0) {
hunk->delta +=
recount_edited_hunk(s, hunk,
backup.header.old_count,
backup.header.new_count);
if (!run_apply_check(s, file_diff))
return 0;
}
/* Drop edits (they were appended to s->plain) */
strbuf_setlen(&s->plain, plain_len);
strbuf_setlen(&s->colored, colored_len);
*hunk = backup;
/*
* TRANSLATORS: do not translate [y/n]
* The program will only accept that input at this point.
* Consider translating (saying "no" discards!) as
* (saying "n" for "no" discards!) if the translation
* of the word "no" does not start with n.
*/
res = prompt_yesno(s, _("Your edited hunk does not apply. "
"Edit again (saying \"no\" discards!) "
"[y/n]? "));
if (res < 1)
return -1;
}
}
static int apply_for_checkout(struct add_p_state *s, struct strbuf *diff,
int is_reverse)
{
const char *reverse = is_reverse ? "-R" : NULL;
struct child_process check_index = CHILD_PROCESS_INIT;
struct child_process check_worktree = CHILD_PROCESS_INIT;
struct child_process apply_index = CHILD_PROCESS_INIT;
struct child_process apply_worktree = CHILD_PROCESS_INIT;
int applies_index, applies_worktree;
setup_child_process(s, &check_index,
"apply", "--cached", "--check", reverse, NULL);
applies_index = !pipe_command(&check_index, diff->buf, diff->len,
NULL, 0, NULL, 0);
setup_child_process(s, &check_worktree,
"apply", "--check", reverse, NULL);
applies_worktree = !pipe_command(&check_worktree, diff->buf, diff->len,
NULL, 0, NULL, 0);
if (applies_worktree && applies_index) {
setup_child_process(s, &apply_index,
"apply", "--cached", reverse, NULL);
pipe_command(&apply_index, diff->buf, diff->len,
NULL, 0, NULL, 0);
setup_child_process(s, &apply_worktree,
"apply", reverse, NULL);
pipe_command(&apply_worktree, diff->buf, diff->len,
NULL, 0, NULL, 0);
return 1;
}
if (!applies_index) {
err(s, _("The selected hunks do not apply to the index!"));
if (prompt_yesno(s, _("Apply them to the worktree "
"anyway? ")) > 0) {
setup_child_process(s, &apply_worktree,
"apply", reverse, NULL);
return pipe_command(&apply_worktree, diff->buf,
diff->len, NULL, 0, NULL, 0);
}
err(s, _("Nothing was applied.\n"));
} else
/* As a last resort, show the diff to the user */
fwrite(diff->buf, diff->len, 1, stderr);
return 0;
}
#define SUMMARY_HEADER_WIDTH 20
#define SUMMARY_LINE_WIDTH 80
static void summarize_hunk(struct add_p_state *s, struct hunk *hunk,
struct strbuf *out)
{
struct hunk_header *header = &hunk->header;
struct strbuf *plain = &s->plain;
size_t len = out->len, i;
strbuf_addf(out, " -%lu,%lu +%lu,%lu ",
header->old_offset, header->old_count,
header->new_offset, header->new_count);
if (out->len - len < SUMMARY_HEADER_WIDTH)
strbuf_addchars(out, ' ',
SUMMARY_HEADER_WIDTH + len - out->len);
for (i = hunk->start; i < hunk->end; i = find_next_line(plain, i))
if (plain->buf[i] != ' ')
break;
if (i < hunk->end)
strbuf_add(out, plain->buf + i, find_next_line(plain, i) - i);
if (out->len - len > SUMMARY_LINE_WIDTH)
strbuf_setlen(out, len + SUMMARY_LINE_WIDTH);
strbuf_complete_line(out);
}
#define DISPLAY_HUNKS_LINES 20
static size_t display_hunks(struct add_p_state *s,
struct file_diff *file_diff, size_t start_index)
{
size_t end_index = start_index + DISPLAY_HUNKS_LINES;
if (end_index > file_diff->hunk_nr)
end_index = file_diff->hunk_nr;
while (start_index < end_index) {
struct hunk *hunk = file_diff->hunk + start_index++;
strbuf_reset(&s->buf);
strbuf_addf(&s->buf, "%c%2d: ", hunk->use == USE_HUNK ? '+'
: hunk->use == SKIP_HUNK ? '-' : ' ',
(int)start_index);
summarize_hunk(s, hunk, &s->buf);
fputs(s->buf.buf, stdout);
}
return end_index;
}
static const char help_patch_remainder[] =
N_("j - leave this hunk undecided, see next undecided hunk\n"
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
"J - leave this hunk undecided, see next hunk\n"
"k - leave this hunk undecided, see previous undecided hunk\n"
"K - leave this hunk undecided, see previous hunk\n"
"g - select a hunk to go to\n"
"/ - search for a hunk matching the given regex\n"
"s - split the current hunk into smaller hunks\n"
"e - manually edit the current hunk\n"
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
"? - print help\n");
static int patch_update_file(struct add_p_state *s,
struct file_diff *file_diff)
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
{
size_t hunk_index = 0;
ssize_t i, undecided_previous, undecided_next;
struct hunk *hunk;
char ch;
struct child_process cp = CHILD_PROCESS_INIT;
int colored = !!s->colored.len, quit = 0;
enum prompt_mode_type prompt_mode_type;
enum {
ALLOW_GOTO_PREVIOUS_HUNK = 1 << 0,
ALLOW_GOTO_PREVIOUS_UNDECIDED_HUNK = 1 << 1,
ALLOW_GOTO_NEXT_HUNK = 1 << 2,
ALLOW_GOTO_NEXT_UNDECIDED_HUNK = 1 << 3,
ALLOW_SEARCH_AND_GOTO = 1 << 4,
ALLOW_SPLIT = 1 << 5,
ALLOW_EDIT = 1 << 6
} permitted = 0;
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
add -p: fix editing of intent-to-add paths A popular way of partially staging a new file is to run `git add -N <path>` and then use the hunk editing of `git add -p` to select the part of the file that the user wishes to stage. Since 85953a3187 ("diff-files --raw: show correct post-image of intent-to-add files", 2020-07-01) this has stopped working as intent-to-add paths are now show as new files rather than changes to an empty blob and `git apply` refused to apply a creation patch for a path that was marked as intent-to-add. 7cfde3fa0f ("apply: allow "new file" patches on i-t-a entries", 2020-08-06) fixed the problem with apply but it still wasn't possible to edit the added hunk properly. 2c8bd8471a ("checkout -p: handle new files correctly", 2020-05-27) had previously changed `add -p` to handle new files but it did not implement patch editing correctly. The perl version simply forbade editing and the C version opened the editor with the full diff rather that just the hunk which meant that the user had to edit the hunk header manually to get it to work. The root cause of the problem is that added files store the diff header with the hunk data rather than separating the two as we do for other changes. Changing added files to store the diff header separately fixes the editing problem at the expense of having to special case empty additions as they no longer have any hunks associated with them, only the diff header. The changes move some existing code into a conditional changing the indentation, they are best viewed with --color-moved-ws=allow-indentation-change (or --ignore-space-change works well to get an overview of the changes) Signed-off-by: Phillip Wood <phillip.wood@dunelm.org.uk> Reported-by: Thomas Sullivan <tom@msbit.com.au> Reported-by: Yuchen Ying <ych@google.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-09-09 16:58:52 +03:00
/* Empty added files have no hunks */
if (!file_diff->hunk_nr && !file_diff->added)
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
return 0;
strbuf_reset(&s->buf);
render_diff_header(s, file_diff, colored, &s->buf);
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
fputs(s->buf.buf, stdout);
for (;;) {
if (hunk_index >= file_diff->hunk_nr)
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
hunk_index = 0;
add -p: fix editing of intent-to-add paths A popular way of partially staging a new file is to run `git add -N <path>` and then use the hunk editing of `git add -p` to select the part of the file that the user wishes to stage. Since 85953a3187 ("diff-files --raw: show correct post-image of intent-to-add files", 2020-07-01) this has stopped working as intent-to-add paths are now show as new files rather than changes to an empty blob and `git apply` refused to apply a creation patch for a path that was marked as intent-to-add. 7cfde3fa0f ("apply: allow "new file" patches on i-t-a entries", 2020-08-06) fixed the problem with apply but it still wasn't possible to edit the added hunk properly. 2c8bd8471a ("checkout -p: handle new files correctly", 2020-05-27) had previously changed `add -p` to handle new files but it did not implement patch editing correctly. The perl version simply forbade editing and the C version opened the editor with the full diff rather that just the hunk which meant that the user had to edit the hunk header manually to get it to work. The root cause of the problem is that added files store the diff header with the hunk data rather than separating the two as we do for other changes. Changing added files to store the diff header separately fixes the editing problem at the expense of having to special case empty additions as they no longer have any hunks associated with them, only the diff header. The changes move some existing code into a conditional changing the indentation, they are best viewed with --color-moved-ws=allow-indentation-change (or --ignore-space-change works well to get an overview of the changes) Signed-off-by: Phillip Wood <phillip.wood@dunelm.org.uk> Reported-by: Thomas Sullivan <tom@msbit.com.au> Reported-by: Yuchen Ying <ych@google.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-09-09 16:58:52 +03:00
hunk = file_diff->hunk_nr
? file_diff->hunk + hunk_index
: &file_diff->head;
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
undecided_previous = -1;
undecided_next = -1;
add -p: fix editing of intent-to-add paths A popular way of partially staging a new file is to run `git add -N <path>` and then use the hunk editing of `git add -p` to select the part of the file that the user wishes to stage. Since 85953a3187 ("diff-files --raw: show correct post-image of intent-to-add files", 2020-07-01) this has stopped working as intent-to-add paths are now show as new files rather than changes to an empty blob and `git apply` refused to apply a creation patch for a path that was marked as intent-to-add. 7cfde3fa0f ("apply: allow "new file" patches on i-t-a entries", 2020-08-06) fixed the problem with apply but it still wasn't possible to edit the added hunk properly. 2c8bd8471a ("checkout -p: handle new files correctly", 2020-05-27) had previously changed `add -p` to handle new files but it did not implement patch editing correctly. The perl version simply forbade editing and the C version opened the editor with the full diff rather that just the hunk which meant that the user had to edit the hunk header manually to get it to work. The root cause of the problem is that added files store the diff header with the hunk data rather than separating the two as we do for other changes. Changing added files to store the diff header separately fixes the editing problem at the expense of having to special case empty additions as they no longer have any hunks associated with them, only the diff header. The changes move some existing code into a conditional changing the indentation, they are best viewed with --color-moved-ws=allow-indentation-change (or --ignore-space-change works well to get an overview of the changes) Signed-off-by: Phillip Wood <phillip.wood@dunelm.org.uk> Reported-by: Thomas Sullivan <tom@msbit.com.au> Reported-by: Yuchen Ying <ych@google.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-09-09 16:58:52 +03:00
if (file_diff->hunk_nr) {
for (i = hunk_index - 1; i >= 0; i--)
if (file_diff->hunk[i].use == UNDECIDED_HUNK) {
undecided_previous = i;
break;
}
for (i = hunk_index + 1; i < file_diff->hunk_nr; i++)
if (file_diff->hunk[i].use == UNDECIDED_HUNK) {
undecided_next = i;
break;
}
}
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
/* Everything decided? */
if (undecided_previous < 0 && undecided_next < 0 &&
hunk->use != UNDECIDED_HUNK)
break;
strbuf_reset(&s->buf);
add -p: fix editing of intent-to-add paths A popular way of partially staging a new file is to run `git add -N <path>` and then use the hunk editing of `git add -p` to select the part of the file that the user wishes to stage. Since 85953a3187 ("diff-files --raw: show correct post-image of intent-to-add files", 2020-07-01) this has stopped working as intent-to-add paths are now show as new files rather than changes to an empty blob and `git apply` refused to apply a creation patch for a path that was marked as intent-to-add. 7cfde3fa0f ("apply: allow "new file" patches on i-t-a entries", 2020-08-06) fixed the problem with apply but it still wasn't possible to edit the added hunk properly. 2c8bd8471a ("checkout -p: handle new files correctly", 2020-05-27) had previously changed `add -p` to handle new files but it did not implement patch editing correctly. The perl version simply forbade editing and the C version opened the editor with the full diff rather that just the hunk which meant that the user had to edit the hunk header manually to get it to work. The root cause of the problem is that added files store the diff header with the hunk data rather than separating the two as we do for other changes. Changing added files to store the diff header separately fixes the editing problem at the expense of having to special case empty additions as they no longer have any hunks associated with them, only the diff header. The changes move some existing code into a conditional changing the indentation, they are best viewed with --color-moved-ws=allow-indentation-change (or --ignore-space-change works well to get an overview of the changes) Signed-off-by: Phillip Wood <phillip.wood@dunelm.org.uk> Reported-by: Thomas Sullivan <tom@msbit.com.au> Reported-by: Yuchen Ying <ych@google.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-09-09 16:58:52 +03:00
if (file_diff->hunk_nr) {
render_hunk(s, hunk, 0, colored, &s->buf);
fputs(s->buf.buf, stdout);
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
add -p: fix editing of intent-to-add paths A popular way of partially staging a new file is to run `git add -N <path>` and then use the hunk editing of `git add -p` to select the part of the file that the user wishes to stage. Since 85953a3187 ("diff-files --raw: show correct post-image of intent-to-add files", 2020-07-01) this has stopped working as intent-to-add paths are now show as new files rather than changes to an empty blob and `git apply` refused to apply a creation patch for a path that was marked as intent-to-add. 7cfde3fa0f ("apply: allow "new file" patches on i-t-a entries", 2020-08-06) fixed the problem with apply but it still wasn't possible to edit the added hunk properly. 2c8bd8471a ("checkout -p: handle new files correctly", 2020-05-27) had previously changed `add -p` to handle new files but it did not implement patch editing correctly. The perl version simply forbade editing and the C version opened the editor with the full diff rather that just the hunk which meant that the user had to edit the hunk header manually to get it to work. The root cause of the problem is that added files store the diff header with the hunk data rather than separating the two as we do for other changes. Changing added files to store the diff header separately fixes the editing problem at the expense of having to special case empty additions as they no longer have any hunks associated with them, only the diff header. The changes move some existing code into a conditional changing the indentation, they are best viewed with --color-moved-ws=allow-indentation-change (or --ignore-space-change works well to get an overview of the changes) Signed-off-by: Phillip Wood <phillip.wood@dunelm.org.uk> Reported-by: Thomas Sullivan <tom@msbit.com.au> Reported-by: Yuchen Ying <ych@google.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-09-09 16:58:52 +03:00
strbuf_reset(&s->buf);
if (undecided_previous >= 0) {
permitted |= ALLOW_GOTO_PREVIOUS_UNDECIDED_HUNK;
strbuf_addstr(&s->buf, ",k");
}
if (hunk_index) {
permitted |= ALLOW_GOTO_PREVIOUS_HUNK;
strbuf_addstr(&s->buf, ",K");
}
if (undecided_next >= 0) {
permitted |= ALLOW_GOTO_NEXT_UNDECIDED_HUNK;
strbuf_addstr(&s->buf, ",j");
}
if (hunk_index + 1 < file_diff->hunk_nr) {
permitted |= ALLOW_GOTO_NEXT_HUNK;
strbuf_addstr(&s->buf, ",J");
}
if (file_diff->hunk_nr > 1) {
permitted |= ALLOW_SEARCH_AND_GOTO;
strbuf_addstr(&s->buf, ",g,/");
}
if (hunk->splittable_into > 1) {
permitted |= ALLOW_SPLIT;
strbuf_addstr(&s->buf, ",s");
}
if (hunk_index + 1 > file_diff->mode_change &&
!file_diff->deleted) {
permitted |= ALLOW_EDIT;
strbuf_addstr(&s->buf, ",e");
}
}
if (file_diff->deleted)
prompt_mode_type = PROMPT_DELETION;
else if (file_diff->added)
prompt_mode_type = PROMPT_ADDITION;
else if (file_diff->mode_change && !hunk_index)
prompt_mode_type = PROMPT_MODE_CHANGE;
else
prompt_mode_type = PROMPT_HUNK;
printf("%s(%"PRIuMAX"/%"PRIuMAX") ", s->s.prompt_color,
(uintmax_t)hunk_index + 1,
add -p: fix editing of intent-to-add paths A popular way of partially staging a new file is to run `git add -N <path>` and then use the hunk editing of `git add -p` to select the part of the file that the user wishes to stage. Since 85953a3187 ("diff-files --raw: show correct post-image of intent-to-add files", 2020-07-01) this has stopped working as intent-to-add paths are now show as new files rather than changes to an empty blob and `git apply` refused to apply a creation patch for a path that was marked as intent-to-add. 7cfde3fa0f ("apply: allow "new file" patches on i-t-a entries", 2020-08-06) fixed the problem with apply but it still wasn't possible to edit the added hunk properly. 2c8bd8471a ("checkout -p: handle new files correctly", 2020-05-27) had previously changed `add -p` to handle new files but it did not implement patch editing correctly. The perl version simply forbade editing and the C version opened the editor with the full diff rather that just the hunk which meant that the user had to edit the hunk header manually to get it to work. The root cause of the problem is that added files store the diff header with the hunk data rather than separating the two as we do for other changes. Changing added files to store the diff header separately fixes the editing problem at the expense of having to special case empty additions as they no longer have any hunks associated with them, only the diff header. The changes move some existing code into a conditional changing the indentation, they are best viewed with --color-moved-ws=allow-indentation-change (or --ignore-space-change works well to get an overview of the changes) Signed-off-by: Phillip Wood <phillip.wood@dunelm.org.uk> Reported-by: Thomas Sullivan <tom@msbit.com.au> Reported-by: Yuchen Ying <ych@google.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-09-09 16:58:52 +03:00
(uintmax_t)(file_diff->hunk_nr
? file_diff->hunk_nr
: 1));
printf(_(s->mode->prompt_mode[prompt_mode_type]),
s->buf.buf);
if (*s->s.reset_color)
fputs(s->s.reset_color, stdout);
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
fflush(stdout);
if (read_single_character(s) == EOF)
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
break;
if (!s->answer.len)
continue;
ch = tolower(s->answer.buf[0]);
if (ch == 'y') {
hunk->use = USE_HUNK;
soft_increment:
hunk_index = undecided_next < 0 ?
file_diff->hunk_nr : undecided_next;
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
} else if (ch == 'n') {
hunk->use = SKIP_HUNK;
goto soft_increment;
} else if (ch == 'a') {
add -p: fix editing of intent-to-add paths A popular way of partially staging a new file is to run `git add -N <path>` and then use the hunk editing of `git add -p` to select the part of the file that the user wishes to stage. Since 85953a3187 ("diff-files --raw: show correct post-image of intent-to-add files", 2020-07-01) this has stopped working as intent-to-add paths are now show as new files rather than changes to an empty blob and `git apply` refused to apply a creation patch for a path that was marked as intent-to-add. 7cfde3fa0f ("apply: allow "new file" patches on i-t-a entries", 2020-08-06) fixed the problem with apply but it still wasn't possible to edit the added hunk properly. 2c8bd8471a ("checkout -p: handle new files correctly", 2020-05-27) had previously changed `add -p` to handle new files but it did not implement patch editing correctly. The perl version simply forbade editing and the C version opened the editor with the full diff rather that just the hunk which meant that the user had to edit the hunk header manually to get it to work. The root cause of the problem is that added files store the diff header with the hunk data rather than separating the two as we do for other changes. Changing added files to store the diff header separately fixes the editing problem at the expense of having to special case empty additions as they no longer have any hunks associated with them, only the diff header. The changes move some existing code into a conditional changing the indentation, they are best viewed with --color-moved-ws=allow-indentation-change (or --ignore-space-change works well to get an overview of the changes) Signed-off-by: Phillip Wood <phillip.wood@dunelm.org.uk> Reported-by: Thomas Sullivan <tom@msbit.com.au> Reported-by: Yuchen Ying <ych@google.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-09-09 16:58:52 +03:00
if (file_diff->hunk_nr) {
for (; hunk_index < file_diff->hunk_nr; hunk_index++) {
hunk = file_diff->hunk + hunk_index;
if (hunk->use == UNDECIDED_HUNK)
hunk->use = USE_HUNK;
}
} else if (hunk->use == UNDECIDED_HUNK) {
hunk->use = USE_HUNK;
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
}
} else if (ch == 'd' || ch == 'q') {
add -p: fix editing of intent-to-add paths A popular way of partially staging a new file is to run `git add -N <path>` and then use the hunk editing of `git add -p` to select the part of the file that the user wishes to stage. Since 85953a3187 ("diff-files --raw: show correct post-image of intent-to-add files", 2020-07-01) this has stopped working as intent-to-add paths are now show as new files rather than changes to an empty blob and `git apply` refused to apply a creation patch for a path that was marked as intent-to-add. 7cfde3fa0f ("apply: allow "new file" patches on i-t-a entries", 2020-08-06) fixed the problem with apply but it still wasn't possible to edit the added hunk properly. 2c8bd8471a ("checkout -p: handle new files correctly", 2020-05-27) had previously changed `add -p` to handle new files but it did not implement patch editing correctly. The perl version simply forbade editing and the C version opened the editor with the full diff rather that just the hunk which meant that the user had to edit the hunk header manually to get it to work. The root cause of the problem is that added files store the diff header with the hunk data rather than separating the two as we do for other changes. Changing added files to store the diff header separately fixes the editing problem at the expense of having to special case empty additions as they no longer have any hunks associated with them, only the diff header. The changes move some existing code into a conditional changing the indentation, they are best viewed with --color-moved-ws=allow-indentation-change (or --ignore-space-change works well to get an overview of the changes) Signed-off-by: Phillip Wood <phillip.wood@dunelm.org.uk> Reported-by: Thomas Sullivan <tom@msbit.com.au> Reported-by: Yuchen Ying <ych@google.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-09-09 16:58:52 +03:00
if (file_diff->hunk_nr) {
for (; hunk_index < file_diff->hunk_nr; hunk_index++) {
hunk = file_diff->hunk + hunk_index;
if (hunk->use == UNDECIDED_HUNK)
hunk->use = SKIP_HUNK;
}
} else if (hunk->use == UNDECIDED_HUNK) {
hunk->use = SKIP_HUNK;
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
}
if (ch == 'q') {
quit = 1;
break;
}
} else if (s->answer.buf[0] == 'K') {
if (permitted & ALLOW_GOTO_PREVIOUS_HUNK)
hunk_index--;
else
err(s, _("No previous hunk"));
} else if (s->answer.buf[0] == 'J') {
if (permitted & ALLOW_GOTO_NEXT_HUNK)
hunk_index++;
else
err(s, _("No next hunk"));
} else if (s->answer.buf[0] == 'k') {
if (permitted & ALLOW_GOTO_PREVIOUS_UNDECIDED_HUNK)
hunk_index = undecided_previous;
else
err(s, _("No previous hunk"));
} else if (s->answer.buf[0] == 'j') {
if (permitted & ALLOW_GOTO_NEXT_UNDECIDED_HUNK)
hunk_index = undecided_next;
else
err(s, _("No next hunk"));
} else if (s->answer.buf[0] == 'g') {
char *pend;
unsigned long response;
if (!(permitted & ALLOW_SEARCH_AND_GOTO)) {
err(s, _("No other hunks to goto"));
continue;
}
strbuf_remove(&s->answer, 0, 1);
strbuf_trim(&s->answer);
i = hunk_index - DISPLAY_HUNKS_LINES / 2;
if (i < file_diff->mode_change)
i = file_diff->mode_change;
while (s->answer.len == 0) {
i = display_hunks(s, file_diff, i);
printf("%s", i < file_diff->hunk_nr ?
_("go to which hunk (<ret> to see "
"more)? ") : _("go to which hunk? "));
fflush(stdout);
if (strbuf_getline(&s->answer,
stdin) == EOF)
break;
strbuf_trim_trailing_newline(&s->answer);
}
strbuf_trim(&s->answer);
response = strtoul(s->answer.buf, &pend, 10);
if (*pend || pend == s->answer.buf)
err(s, _("Invalid number: '%s'"),
s->answer.buf);
else if (0 < response && response <= file_diff->hunk_nr)
hunk_index = response - 1;
else
err(s, Q_("Sorry, only %d hunk available.",
"Sorry, only %d hunks available.",
file_diff->hunk_nr),
(int)file_diff->hunk_nr);
} else if (s->answer.buf[0] == '/') {
regex_t regex;
int ret;
if (!(permitted & ALLOW_SEARCH_AND_GOTO)) {
err(s, _("No other hunks to search"));
continue;
}
strbuf_remove(&s->answer, 0, 1);
strbuf_trim_trailing_newline(&s->answer);
if (s->answer.len == 0) {
printf("%s", _("search for regex? "));
fflush(stdout);
if (strbuf_getline(&s->answer,
stdin) == EOF)
break;
strbuf_trim_trailing_newline(&s->answer);
if (s->answer.len == 0)
continue;
}
ret = regcomp(&regex, s->answer.buf,
REG_EXTENDED | REG_NOSUB | REG_NEWLINE);
if (ret) {
char errbuf[1024];
regerror(ret, &regex, errbuf, sizeof(errbuf));
err(s, _("Malformed search regexp %s: %s"),
s->answer.buf, errbuf);
continue;
}
i = hunk_index;
for (;;) {
/* render the hunk into a scratch buffer */
render_hunk(s, file_diff->hunk + i, 0, 0,
&s->buf);
if (regexec(&regex, s->buf.buf, 0, NULL, 0)
!= REG_NOMATCH)
break;
i++;
if (i == file_diff->hunk_nr)
i = 0;
if (i != hunk_index)
continue;
err(s, _("No hunk matches the given pattern"));
break;
}
hunk_index = i;
} else if (s->answer.buf[0] == 's') {
size_t splittable_into = hunk->splittable_into;
if (!(permitted & ALLOW_SPLIT))
err(s, _("Sorry, cannot split this hunk"));
else if (!split_hunk(s, file_diff,
hunk - file_diff->hunk))
color_fprintf_ln(stdout, s->s.header_color,
_("Split into %d hunks."),
(int)splittable_into);
} else if (s->answer.buf[0] == 'e') {
if (!(permitted & ALLOW_EDIT))
err(s, _("Sorry, cannot edit this hunk"));
else if (edit_hunk_loop(s, file_diff, hunk) >= 0) {
hunk->use = USE_HUNK;
goto soft_increment;
}
} else {
const char *p = _(help_patch_remainder), *eol = p;
color_fprintf(stdout, s->s.help_color, "%s",
_(s->mode->help_patch_text));
/*
* Show only those lines of the remainder that are
* actually applicable with the current hunk.
*/
for (; *p; p = eol + (*eol == '\n')) {
eol = strchrnul(p, '\n');
/*
* `s->buf` still contains the part of the
* commands shown in the prompt that are not
* always available.
*/
if (*p != '?' && !strchr(s->buf.buf, *p))
continue;
color_fprintf_ln(stdout, s->s.help_color,
"%.*s", (int)(eol - p), p);
}
}
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
}
/* Any hunk to be used? */
for (i = 0; i < file_diff->hunk_nr; i++)
if (file_diff->hunk[i].use == USE_HUNK)
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
break;
add -p: fix editing of intent-to-add paths A popular way of partially staging a new file is to run `git add -N <path>` and then use the hunk editing of `git add -p` to select the part of the file that the user wishes to stage. Since 85953a3187 ("diff-files --raw: show correct post-image of intent-to-add files", 2020-07-01) this has stopped working as intent-to-add paths are now show as new files rather than changes to an empty blob and `git apply` refused to apply a creation patch for a path that was marked as intent-to-add. 7cfde3fa0f ("apply: allow "new file" patches on i-t-a entries", 2020-08-06) fixed the problem with apply but it still wasn't possible to edit the added hunk properly. 2c8bd8471a ("checkout -p: handle new files correctly", 2020-05-27) had previously changed `add -p` to handle new files but it did not implement patch editing correctly. The perl version simply forbade editing and the C version opened the editor with the full diff rather that just the hunk which meant that the user had to edit the hunk header manually to get it to work. The root cause of the problem is that added files store the diff header with the hunk data rather than separating the two as we do for other changes. Changing added files to store the diff header separately fixes the editing problem at the expense of having to special case empty additions as they no longer have any hunks associated with them, only the diff header. The changes move some existing code into a conditional changing the indentation, they are best viewed with --color-moved-ws=allow-indentation-change (or --ignore-space-change works well to get an overview of the changes) Signed-off-by: Phillip Wood <phillip.wood@dunelm.org.uk> Reported-by: Thomas Sullivan <tom@msbit.com.au> Reported-by: Yuchen Ying <ych@google.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-09-09 16:58:52 +03:00
if (i < file_diff->hunk_nr ||
(!file_diff->hunk_nr && file_diff->head.use == USE_HUNK)) {
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
/* At least one hunk selected: apply */
strbuf_reset(&s->buf);
reassemble_patch(s, file_diff, 0, &s->buf);
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
built-in add -p: adjust hunk headers as needed When skipping a hunk that adds a different number of lines than it removes, we need to adjust the subsequent hunk headers of non-skipped hunks: in pathological cases, the context is not enough to determine precisely where the patch should be applied. This problem was identified in 23fea4c240 (t3701: add failing test for pathological context lines, 2018-03-01) and fixed in the Perl version in fecc6f3a68 (add -p: adjust offsets of subsequent hunks when one is skipped, 2018-03-01). And this patch fixes it in the C version of `git add -p`. In contrast to the Perl version, we try to keep the extra text on the hunk header (which typically contains the signature of the function whose code is changed in the hunk) intact. Note: while the C version does not support staging mode changes at this stage, we already prepare for this by simply skipping the hunk header if both old and new offset is 0 (this cannot happen for regular hunks, and we will use this as an indicator that we are looking at a special hunk). Likewise, we already prepare for hunk splitting by handling the absence of extra text in the hunk header gracefully: only the first split hunk will have that text, the others will not (indicated by an empty extra text start/end range). Preparing for hunk splitting already at this stage avoids an indentation change of the entire hunk header-printing block later, and is almost as easy to review as without that handling. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:51 +03:00
discard_index(s->s.r->index);
if (s->mode->apply_for_checkout)
apply_for_checkout(s, &s->buf,
s->mode->is_reverse);
else {
setup_child_process(s, &cp, "apply", NULL);
strvec_pushv(&cp.args, s->mode->apply_args);
if (pipe_command(&cp, s->buf.buf, s->buf.len,
NULL, 0, NULL, 0))
error(_("'git apply' failed"));
}
if (repo_read_index(s->s.r) >= 0)
built-in add -p: adjust hunk headers as needed When skipping a hunk that adds a different number of lines than it removes, we need to adjust the subsequent hunk headers of non-skipped hunks: in pathological cases, the context is not enough to determine precisely where the patch should be applied. This problem was identified in 23fea4c240 (t3701: add failing test for pathological context lines, 2018-03-01) and fixed in the Perl version in fecc6f3a68 (add -p: adjust offsets of subsequent hunks when one is skipped, 2018-03-01). And this patch fixes it in the C version of `git add -p`. In contrast to the Perl version, we try to keep the extra text on the hunk header (which typically contains the signature of the function whose code is changed in the hunk) intact. Note: while the C version does not support staging mode changes at this stage, we already prepare for this by simply skipping the hunk header if both old and new offset is 0 (this cannot happen for regular hunks, and we will use this as an indicator that we are looking at a special hunk). Likewise, we already prepare for hunk splitting by handling the absence of extra text in the hunk header gracefully: only the first split hunk will have that text, the others will not (indicated by an empty extra text start/end range). Preparing for hunk splitting already at this stage avoids an indentation change of the entire hunk header-printing block later, and is almost as easy to review as without that handling. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:51 +03:00
repo_refresh_and_write_index(s->s.r, REFRESH_QUIET, 0,
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
1, NULL, NULL, NULL);
}
putchar('\n');
return quit;
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
}
int run_add_p(struct repository *r, enum add_p_mode mode,
const char *revision, const struct pathspec *ps)
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
{
built-in add -p: adjust hunk headers as needed When skipping a hunk that adds a different number of lines than it removes, we need to adjust the subsequent hunk headers of non-skipped hunks: in pathological cases, the context is not enough to determine precisely where the patch should be applied. This problem was identified in 23fea4c240 (t3701: add failing test for pathological context lines, 2018-03-01) and fixed in the Perl version in fecc6f3a68 (add -p: adjust offsets of subsequent hunks when one is skipped, 2018-03-01). And this patch fixes it in the C version of `git add -p`. In contrast to the Perl version, we try to keep the extra text on the hunk header (which typically contains the signature of the function whose code is changed in the hunk) intact. Note: while the C version does not support staging mode changes at this stage, we already prepare for this by simply skipping the hunk header if both old and new offset is 0 (this cannot happen for regular hunks, and we will use this as an indicator that we are looking at a special hunk). Likewise, we already prepare for hunk splitting by handling the absence of extra text in the hunk header gracefully: only the first split hunk will have that text, the others will not (indicated by an empty extra text start/end range). Preparing for hunk splitting already at this stage avoids an indentation change of the entire hunk header-printing block later, and is almost as easy to review as without that handling. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:51 +03:00
struct add_p_state s = {
{ r }, STRBUF_INIT, STRBUF_INIT, STRBUF_INIT, STRBUF_INIT
};
size_t i, binary_count = 0;
built-in add -p: adjust hunk headers as needed When skipping a hunk that adds a different number of lines than it removes, we need to adjust the subsequent hunk headers of non-skipped hunks: in pathological cases, the context is not enough to determine precisely where the patch should be applied. This problem was identified in 23fea4c240 (t3701: add failing test for pathological context lines, 2018-03-01) and fixed in the Perl version in fecc6f3a68 (add -p: adjust offsets of subsequent hunks when one is skipped, 2018-03-01). And this patch fixes it in the C version of `git add -p`. In contrast to the Perl version, we try to keep the extra text on the hunk header (which typically contains the signature of the function whose code is changed in the hunk) intact. Note: while the C version does not support staging mode changes at this stage, we already prepare for this by simply skipping the hunk header if both old and new offset is 0 (this cannot happen for regular hunks, and we will use this as an indicator that we are looking at a special hunk). Likewise, we already prepare for hunk splitting by handling the absence of extra text in the hunk header gracefully: only the first split hunk will have that text, the others will not (indicated by an empty extra text start/end range). Preparing for hunk splitting already at this stage avoids an indentation change of the entire hunk header-printing block later, and is almost as easy to review as without that handling. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:51 +03:00
init_add_i_state(&s.s, r);
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
if (mode == ADD_P_STASH)
s.mode = &patch_mode_stash;
else if (mode == ADD_P_RESET) {
/*
* NEEDSWORK: Instead of comparing to the literal "HEAD",
* compare the commit objects instead so that other ways of
* saying the same thing (such as "@") are also handled
* appropriately.
*
* This applies to the cases below too.
*/
if (!revision || !strcmp(revision, "HEAD"))
s.mode = &patch_mode_reset_head;
else
s.mode = &patch_mode_reset_nothead;
} else if (mode == ADD_P_CHECKOUT) {
if (!revision)
s.mode = &patch_mode_checkout_index;
else if (!strcmp(revision, "HEAD"))
s.mode = &patch_mode_checkout_head;
else
s.mode = &patch_mode_checkout_nothead;
} else if (mode == ADD_P_WORKTREE) {
if (!revision)
s.mode = &patch_mode_checkout_index;
else if (!strcmp(revision, "HEAD"))
s.mode = &patch_mode_worktree_head;
else
s.mode = &patch_mode_worktree_nothead;
} else
s.mode = &patch_mode_add;
s.revision = revision;
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
if (discard_index(r->index) < 0 || repo_read_index(r) < 0 ||
(!s.mode->index_only &&
repo_refresh_and_write_index(r, REFRESH_QUIET, 0, 1,
NULL, NULL, NULL) < 0) ||
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
parse_diff(&s, ps) < 0) {
add_p_state_clear(&s);
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
return -1;
}
for (i = 0; i < s.file_diff_nr; i++)
if (s.file_diff[i].binary && !s.file_diff[i].hunk_nr)
binary_count++;
else if (patch_update_file(&s, s.file_diff + i))
break;
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
if (s.file_diff_nr == 0)
fprintf(stderr, _("No changes.\n"));
else if (binary_count == s.file_diff_nr)
fprintf(stderr, _("Only binary files changed.\n"));
add_p_state_clear(&s);
built-in add -i: start implementing the `patch` functionality in C In the previous steps, we re-implemented the main loop of `git add -i` in C, and most of the commands. Notably, we left out the actual functionality of `patch`, as the relevant code makes up more than half of `git-add--interactive.perl`, and is actually pretty independent of the rest of the commands. With this commit, we start to tackle that `patch` part. For better separation of concerns, we keep the code in a separate file, `add-patch.c`. The new code is still guarded behind the `add.interactive.useBuiltin` config setting, and for the moment, it can only be called via `git add -p`. The actual functionality follows the original implementation of 5cde71d64aff (git-add --interactive, 2006-12-10), but not too closely (for example, we use string offsets rather than copying strings around, and after seeing whether the `k` and `j` commands are applicable, in the C version we remember which previous/next hunk was undecided, and use it rather than looking again when the user asked to jump). As a further deviation from that commit, We also use a comma instead of a slash to separate the available commands in the prompt, as the current version of the Perl script does this, and we also add a line about the question mark ("print help") to the help text. While it is tempting to use this conversion of `git add -p` as an excuse to work on `apply_all_patches()` so that it does _not_ want to read a file from `stdin` or from a file, but accepts, say, an `strbuf` instead, we will refrain from this particular rabbit hole at this stage. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-12-13 11:07:48 +03:00
return 0;
}