git/remote.h

413 строки
11 KiB
C
Исходник Обычный вид История

#ifndef REMOTE_H
#define REMOTE_H
#include "cache.h"
remote.c: add command line option parser for "--force-with-lease" Update "git push" and "git send-pack" to parse this commnd line option. The intended sematics is: * "--force-with-lease" alone, without specifying the details, will protect _all_ remote refs that are going to be updated by requiring their current value to be the same as some reasonable default, unless otherwise specified; * "--force-with-lease=refname", without specifying the expected value, will protect that refname, if it is going to be updated, by requiring its current value to be the same as some reasonable default. * "--force-with-lease=refname:value" will protect that refname, if it is going to be updated, by requiring its current value to be the same as the specified value; and * "--no-force-with-lease" will cancel all the previous --force-with-lease on the command line. For now, "some reasonable default" is tentatively defined as "the value of the remote-tracking branch we have for the ref of the remote being updated", and it is an error if we do not have such a remote-tracking branch. But this is known to be fragile, its use is not yet recommended, and hopefully we will find more reasonable default as we gain experience with this feature. The manual marks the feature as experimental unless the expected value is specified explicitly for this reason. Because the command line options are parsed _before_ we know which remote we are pushing to, there needs further processing to the parsed data after we instantiate the transport object to: * expand "refname" given by the user to a full refname to be matched with the list of "struct ref" used in match_push_refs() and set_ref_status_for_push(); and * learning the actual local ref that is the remote-tracking branch for the specified remote ref. Further, some processing need to be deferred until we find the set of remote refs and match_push_refs() returns in order to find the ones that need to be checked after explicit ones have been processed for "--force-with-lease" (no specific details). These post-processing will be the topic of the next patch. This option was originally called "cas" (for "compare and swap"), the name which nobody liked because it was too technical. The second attempt called it "lockref" (because it is conceptually like pushing after taking a lock) but the word "lock" was hated because it implied that it may reject push by others, which is not the way this option works. This round calls it "force-with-lease". You assume you took the lease on the ref when you fetched to decide what the rebased history should be, and you can push back only if the lease has not been broken. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-07-09 02:34:36 +04:00
#include "parse-options.h"
#include "hashmap.h"
#include "refspec.h"
remote.c: add command line option parser for "--force-with-lease" Update "git push" and "git send-pack" to parse this commnd line option. The intended sematics is: * "--force-with-lease" alone, without specifying the details, will protect _all_ remote refs that are going to be updated by requiring their current value to be the same as some reasonable default, unless otherwise specified; * "--force-with-lease=refname", without specifying the expected value, will protect that refname, if it is going to be updated, by requiring its current value to be the same as some reasonable default. * "--force-with-lease=refname:value" will protect that refname, if it is going to be updated, by requiring its current value to be the same as the specified value; and * "--no-force-with-lease" will cancel all the previous --force-with-lease on the command line. For now, "some reasonable default" is tentatively defined as "the value of the remote-tracking branch we have for the ref of the remote being updated", and it is an error if we do not have such a remote-tracking branch. But this is known to be fragile, its use is not yet recommended, and hopefully we will find more reasonable default as we gain experience with this feature. The manual marks the feature as experimental unless the expected value is specified explicitly for this reason. Because the command line options are parsed _before_ we know which remote we are pushing to, there needs further processing to the parsed data after we instantiate the transport object to: * expand "refname" given by the user to a full refname to be matched with the list of "struct ref" used in match_push_refs() and set_ref_status_for_push(); and * learning the actual local ref that is the remote-tracking branch for the specified remote ref. Further, some processing need to be deferred until we find the set of remote refs and match_push_refs() returns in order to find the ones that need to be checked after explicit ones have been processed for "--force-with-lease" (no specific details). These post-processing will be the topic of the next patch. This option was originally called "cas" (for "compare and swap"), the name which nobody liked because it was too technical. The second attempt called it "lockref" (because it is conceptually like pushing after taking a lock) but the word "lock" was hated because it implied that it may reject push by others, which is not the way this option works. This round calls it "force-with-lease". You assume you took the lease on the ref when you fetched to decide what the rebased history should be, and you can push back only if the lease has not been broken. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-07-09 02:34:36 +04:00
struct transport_ls_refs_options;
/**
* The API gives access to the configuration related to remotes. It handles
* all three configuration mechanisms historically and currently used by Git,
* and presents the information in a uniform fashion. Note that the code also
* handles plain URLs without any configuration, giving them just the default
* information.
*/
enum {
REMOTE_UNCONFIGURED = 0,
REMOTE_CONFIG,
REMOTE_REMOTES,
REMOTE_BRANCHES
};
struct rewrite {
const char *base;
size_t baselen;
struct counted_string *instead_of;
int instead_of_nr;
int instead_of_alloc;
};
struct rewrites {
struct rewrite **rewrite;
int rewrite_alloc;
int rewrite_nr;
};
struct remote_state {
struct remote **remotes;
int remotes_alloc;
int remotes_nr;
struct hashmap remotes_hash;
struct hashmap branches_hash;
struct branch *current_branch;
const char *pushremote_name;
struct rewrites rewrites;
struct rewrites rewrites_push;
int initialized;
};
void remote_state_clear(struct remote_state *remote_state);
struct remote_state *remote_state_new(void);
struct remote {
struct hashmap_entry ent;
/* The user's nickname for the remote */
const char *name;
remote rename: more carefully determine whether a remote is configured One of the really nice features of the ~/.gitconfig file is that users can override defaults by their own preferred settings for all of their repositories. One such default that some users like to override is whether the "origin" remote gets auto-pruned or not. The user would simply call git config --global remote.origin.prune true and from now on all "origin" remotes would be pruned automatically when fetching into the local repository. There is just one catch: now Git thinks that the "origin" remote is configured, even if the repository config has no [remote "origin"] section at all, as it does not realize that the "prune" setting was configured globally and that there really is no "origin" remote configured in this repository. That is a problem e.g. when renaming a remote to a new name, when Git may be fooled into thinking that there is already a remote of that new name. Let's fix this by paying more attention to *where* the remote settings came from: if they are configured in the local repository config, we must not overwrite them. If they were configured elsewhere, we cannot overwrite them to begin with, as we only write the repository config. There is only one caller of remote_is_configured() (in `git fetch`) that may want to take remotes into account even if they were configured outside the repository config; all other callers essentially try to prevent the Git command from overwriting settings in the repository config. To accommodate that fact, the remote_is_configured() function now requires a parameter that states whether the caller is interested in all remotes, or only in those that were configured in the repository config. Many thanks to Jeff King whose tireless review helped with settling for nothing less than the current strategy. This fixes https://github.com/git-for-windows/git/issues/888 Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2017-01-20 00:20:02 +03:00
int origin, configured_in_repo;
const char *foreign_vcs;
/* An array of all of the url_nr URLs configured for the remote */
const char **url;
int url_nr;
int url_alloc;
/* An array of all of the pushurl_nr push URLs configured for the remote */
const char **pushurl;
int pushurl_nr;
int pushurl_alloc;
struct refspec push;
struct refspec fetch;
/*
* The setting for whether to fetch tags (as a separate rule from the
* configured refspecs);
* -1 to never fetch tags
* 0 to auto-follow tags on heuristic (default)
* 1 to always auto-follow tags
* 2 to always fetch tags
*/
int fetch_tags;
int skip_default_update;
int mirror;
int prune;
fetch: add a --prune-tags option and fetch.pruneTags config Add a --prune-tags option to git-fetch, along with fetch.pruneTags config option and a -P shorthand (-p is --prune). This allows for doing any of: git fetch -p -P git fetch --prune --prune-tags git fetch -p -P origin git fetch --prune --prune-tags origin Or simply: git config fetch.prune true && git config fetch.pruneTags true && git fetch Instead of the much more verbose: git fetch --prune origin 'refs/tags/*:refs/tags/*' '+refs/heads/*:refs/remotes/origin/*' Before this feature it was painful to support the use-case of pulling from a repo which is having both its branches *and* tags deleted regularly, and have our local references to reflect upstream. At work we create deployment tags in the repo for each rollout, and there's *lots* of those, so they're archived within weeks for performance reasons. Without this change it's hard to centrally configure such repos in /etc/gitconfig (on servers that are only used for working with them). You need to set fetch.prune=true globally, and then for each repo: git -C {} config --replace-all remote.origin.fetch "refs/tags/*:refs/tags/*" "^\+*refs/tags/\*:refs/tags/\*$" Now I can simply set fetch.pruneTags=true in /etc/gitconfig as well, and users running "git pull" will automatically get the pruning semantics I want. Even though "git remote" has corresponding "prune" and "update --prune" subcommands I'm intentionally not adding a corresponding prune-tags or "update --prune --prune-tags" mode to that command. It's advertised (as noted in my recent "git remote doc: correct dangerous lies about what prune does") as only modifying remote tracking references, whereas any --prune-tags option is always going to modify what from the user's perspective is a local copy of the tag, since there's no such thing as a remote tracking tag. Ideally add_prune_tags_to_fetch_refspec() would be something that would use ALLOC_GROW() to grow the 'fetch` member of the 'remote' struct. Instead I'm realloc-ing remote->fetch and adding the tag_refspec to the end. The reason is that parse_{fetch,push}_refspec which allocate the refspec (ultimately remote->fetch) struct are called many places that don't have access to a 'remote' struct. It would be hard to change all their callsites to be amenable to carry around the bookkeeping variables required for dynamic allocation. All the other callers of the API first incrementally construct the string version of the refspec in remote->fetch_refspec via add_fetch_refspec(), before finally calling parse_fetch_refspec() via some variation of remote_get(). It's less of a pain to deal with the one special case that needs to modify already constructed refspecs than to chase down and change all the other callsites. The API I'm adding is intentionally not generalized because if we add more of these we'd probably want to re-visit how this is done. See my "Re: [BUG] git remote prune removes local tags, depending on fetch config" (87po6ahx87.fsf@evledraar.gmail.com; https://public-inbox.org/git/87po6ahx87.fsf@evledraar.gmail.com/) for more background info. Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-02-09 23:32:15 +03:00
int prune_tags;
/**
* The configured helper programs to run on the remote side, for
* Git-native protocols.
*/
const char *receivepack;
const char *uploadpack;
/* The proxy to use for curl (http, https, ftp, etc.) URLs. */
char *http_proxy;
/* The method used for authenticating against `http_proxy`. */
char *http_proxy_authmethod;
};
/**
* struct remotes can be found by name with remote_get().
* remote_get(NULL) will return the default remote, given the current branch
* and configuration.
*/
struct remote *remote_get(const char *name);
struct remote *pushremote_get(const char *name);
remote rename: more carefully determine whether a remote is configured One of the really nice features of the ~/.gitconfig file is that users can override defaults by their own preferred settings for all of their repositories. One such default that some users like to override is whether the "origin" remote gets auto-pruned or not. The user would simply call git config --global remote.origin.prune true and from now on all "origin" remotes would be pruned automatically when fetching into the local repository. There is just one catch: now Git thinks that the "origin" remote is configured, even if the repository config has no [remote "origin"] section at all, as it does not realize that the "prune" setting was configured globally and that there really is no "origin" remote configured in this repository. That is a problem e.g. when renaming a remote to a new name, when Git may be fooled into thinking that there is already a remote of that new name. Let's fix this by paying more attention to *where* the remote settings came from: if they are configured in the local repository config, we must not overwrite them. If they were configured elsewhere, we cannot overwrite them to begin with, as we only write the repository config. There is only one caller of remote_is_configured() (in `git fetch`) that may want to take remotes into account even if they were configured outside the repository config; all other callers essentially try to prevent the Git command from overwriting settings in the repository config. To accommodate that fact, the remote_is_configured() function now requires a parameter that states whether the caller is interested in all remotes, or only in those that were configured in the repository config. Many thanks to Jeff King whose tireless review helped with settling for nothing less than the current strategy. This fixes https://github.com/git-for-windows/git/issues/888 Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2017-01-20 00:20:02 +03:00
int remote_is_configured(struct remote *remote, int in_repo);
typedef int each_remote_fn(struct remote *remote, void *priv);
/* iterate through struct remotes */
int for_each_remote(each_remote_fn fn, void *priv);
int remote_has_url(struct remote *remote, const char *url);
receive-pack: add new proc-receive hook Git calls an internal `execute_commands` function to handle commands sent from client to `git-receive-pack`. Regardless of what references the user pushes, git creates or updates the corresponding references if the user has write-permission. A contributor who has no write-permission, cannot push to the repository directly. So, the contributor has to write commits to an alternate location, and sends pull request by emails or by other ways. We call this workflow as a distributed workflow. It would be more convenient to work in a centralized workflow like what Gerrit provided for some cases. For example, a read-only user who cannot push to a branch directly can run the following `git push` command to push commits to a pseudo reference (has a prefix "refs/for/", not "refs/heads/") to create a code review. git push origin \ HEAD:refs/for/<branch-name>/<session> The `<branch-name>` in the above example can be as simple as "master", or a more complicated branch name like "foo/bar". The `<session>` in the above example command can be the local branch name of the client side, such as "my/topic". We cannot implement a centralized workflow elegantly by using "pre-receive" + "post-receive", because Git will call the internal function "execute_commands" to create references (even the special pseudo reference) between these two hooks. Even though we can delete the temporarily created pseudo reference via the "post-receive" hook, having a temporary reference is not safe for concurrent pushes. So, add a filter and a new handler to support this kind of workflow. The filter will check the prefix of the reference name, and if the command has a special reference name, the filter will turn a specific field (`run_proc_receive`) on for the command. Commands with this filed turned on will be executed by a new handler (a hook named "proc-receive") instead of the internal `execute_commands` function. We can use this "proc-receive" command to create pull requests or send emails for code review. Suggested by Junio, this "proc-receive" hook reads the commands, push-options (optional), and send result using a protocol in pkt-line format. In the following example, the letter "S" stands for "receive-pack" and letter "H" stands for the hook. # Version and features negotiation. S: PKT-LINE(version=1\0push-options atomic...) S: flush-pkt H: PKT-LINE(version=1\0push-options...) H: flush-pkt # Send commands from server to the hook. S: PKT-LINE(<old-oid> <new-oid> <ref>) S: ... ... S: flush-pkt # Send push-options only if the 'push-options' feature is enabled. S: PKT-LINE(push-option) S: ... ... S: flush-pkt # Receive result from the hook. # OK, run this command successfully. H: PKT-LINE(ok <ref>) # NO, I reject it. H: PKT-LINE(ng <ref> <reason>) # Fall through, let 'receive-pack' to execute it. H: PKT-LINE(ok <ref>) H: PKT-LINE(option fall-through) # OK, but has an alternate reference. The alternate reference name # and other status can be given in options H: PKT-LINE(ok <ref>) H: PKT-LINE(option refname <refname>) H: PKT-LINE(option old-oid <old-oid>) H: PKT-LINE(option new-oid <new-oid>) H: PKT-LINE(option forced-update) H: ... ... H: flush-pkt After receiving a command, the hook will execute the command, and may create/update different reference. For example, a command for a pseudo reference "refs/for/master/topic" may create/update different reference such as "refs/pull/123/head". The alternate reference name and other status are given in option lines. The list of commands returned from "proc-receive" will replace the relevant commands that are sent from user to "receive-pack", and "receive-pack" will continue to run the "execute_commands" function and other routines. Finally, the result of the execution of these commands will be reported to end user. The reporting function from "receive-pack" to "send-pack" will be extended in latter commit just like what the "proc-receive" hook reports to "receive-pack". Signed-off-by: Jiang Xin <zhiyou.jx@alibaba-inc.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-08-27 18:45:44 +03:00
struct ref_push_report {
const char *ref_name;
struct object_id *old_oid;
struct object_id *new_oid;
unsigned int forced_update:1;
struct ref_push_report *next;
};
struct ref {
struct ref *next;
struct object_id old_oid;
struct object_id new_oid;
struct object_id old_oid_expect; /* used by expect-old */
char *symref;
char *tracking_ref;
unsigned int
force:1,
forced_update:1,
expect_old_sha1:1,
exact_oid:1,
deletion:1,
/* Need to check if local reflog reaches the remote tip. */
check_reachable:1,
/*
* Store the result of the check enabled by "check_reachable";
* implies the local reflog does not reach the remote tip.
*/
unreachable:1;
enum {
REF_NOT_MATCHED = 0, /* initial value */
REF_MATCHED,
REF_UNADVERTISED_NOT_ALLOWED
} match_status;
/*
* Order is important here, as we write to FETCH_HEAD
* in numeric order. And the default NOT_FOR_MERGE
* should be 0, so that xcalloc'd structures get it
* by default.
*/
enum fetch_head_status {
FETCH_HEAD_MERGE = -1,
FETCH_HEAD_NOT_FOR_MERGE = 0,
FETCH_HEAD_IGNORE = 1
} fetch_head_status;
enum {
REF_STATUS_NONE = 0,
REF_STATUS_OK,
REF_STATUS_REJECT_NONFASTFORWARD,
REF_STATUS_REJECT_ALREADY_EXISTS,
REF_STATUS_REJECT_NODELETE,
REF_STATUS_REJECT_FETCH_FIRST,
REF_STATUS_REJECT_NEEDS_FORCE,
REF_STATUS_REJECT_STALE,
REF_STATUS_REJECT_SHALLOW,
REF_STATUS_REJECT_REMOTE_UPDATED,
REF_STATUS_UPTODATE,
REF_STATUS_REMOTE_REJECT,
REF_STATUS_EXPECTING_REPORT,
REF_STATUS_ATOMIC_PUSH_FAILED
} status;
char *remote_status;
2020-08-27 18:45:46 +03:00
struct ref_push_report *report;
struct ref *peer_ref; /* when renaming */
char name[FLEX_ARRAY]; /* more */
};
#define REF_NORMAL (1u << 0)
#define REF_HEADS (1u << 1)
#define REF_TAGS (1u << 2)
struct ref *find_ref_by_name(const struct ref *list, const char *name);
struct ref *alloc_ref(const char *name);
struct ref *copy_ref(const struct ref *ref);
struct ref *copy_ref_list(const struct ref *ref);
void sort_ref_list(struct ref **, int (*cmp)(const void *, const void *));
int count_refspec_match(const char *, struct ref *refs, struct ref **matched_ref);
int ref_compare_name(const void *, const void *);
int check_ref_type(const struct ref *ref, int flags);
/*
* Free a single ref and its peer, or an entire list of refs and their peers,
* respectively.
*/
void free_one_ref(struct ref *ref);
void free_refs(struct ref *ref);
struct oid_array;
struct packet_reader;
argv-array: rename to strvec The name "argv-array" isn't very good, because it describes what the data type can be used for (program argument arrays), not what it actually is (a dynamically-growing string array that maintains a NULL-terminator invariant). This leads to people being hesitant to use it for other cases where it would actually be a good fit. The existing name is also clunky to use. It's overly long, and the name often leads to saying things like "argv.argv" (i.e., the field names overlap with variable names, since they're describing the use, not the type). Let's give it a more neutral name. I settled on "strvec" because "vector" is the name for a dynamic array type in many programming languages. "strarray" would work, too, but it's longer and a bit more awkward to say (and don't we all say these things in our mind as we type them?). A more extreme direction would be a generic data structure which stores a NULL-terminated of _any_ type. That would be easy to do with void pointers, but we'd lose some type safety for the existing cases. Plus it raises questions about memory allocation and ownership. So I limited myself here to changing names only, and not semantics. If we do find a use for that more generic data type, we could perhaps implement it at a lower level and then provide type-safe wrappers around it for strings. But that can come later. This patch does the minimum to convert the struct and function names in the header and implementation, leaving a few things for follow-on patches: - files retain their original names for now - struct field names are retained for now - there's a preprocessor compat layer that lets most users remain the same for now. The exception is headers which made a manual forward declaration of the struct. I've converted them (and their dependent function declarations) here. Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-07-28 23:23:25 +03:00
struct strvec;
struct string_list;
struct ref **get_remote_heads(struct packet_reader *reader,
struct ref **list, unsigned int flags,
struct oid_array *extra_have,
struct oid_array *shallow_points);
/* Used for protocol v2 in order to retrieve refs from a remote */
struct ref **get_remote_refs(int fd_out, struct packet_reader *reader,
struct ref **list, int for_push,
struct transport_ls_refs_options *transport_options,
stateless-connect: send response end packet Currently, remote-curl acts as a proxy and blindly forwards packets between an HTTP server and fetch-pack. In the case of a stateless RPC connection where the connection is terminated before the transaction is complete, remote-curl will blindly forward the packets before waiting on more input from fetch-pack. Meanwhile, fetch-pack will read the transaction and continue reading, expecting more input to continue the transaction. This results in a deadlock between the two processes. This can be seen in the following command which does not terminate: $ git -c protocol.version=2 clone https://github.com/git/git.git --shallow-since=20151012 Cloning into 'git'... whereas the v1 version does terminate as expected: $ git -c protocol.version=1 clone https://github.com/git/git.git --shallow-since=20151012 Cloning into 'git'... fatal: the remote end hung up unexpectedly Instead of blindly forwarding packets, make remote-curl insert a response end packet after proxying the responses from the remote server when using stateless_connect(). On the RPC client side, ensure that each response ends as described. A separate control packet is chosen because we need to be able to differentiate between what the remote server sends and remote-curl's control packets. By ensuring in the remote-curl code that a server cannot send response end packets, we prevent a malicious server from being able to perform a denial of service attack in which they spoof a response end packet and cause the described deadlock to happen. Reported-by: Force Charlie <charlieio@outlook.com> Helped-by: Jeff King <peff@peff.net> Signed-off-by: Denton Liu <liu.denton@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-05-19 13:54:00 +03:00
const struct string_list *server_options,
int stateless_rpc);
int resolve_remote_symref(struct ref *ref, struct ref *list);
/*
* Remove and free all but the first of any entries in the input list
* that map the same remote reference to the same local reference. If
* there are two entries that map different remote references to the
* same local reference, emit an error message and die. Return a
* pointer to the head of the resulting list.
*/
struct ref *ref_remove_duplicates(struct ref *ref_map);
refspec: add support for negative refspecs Both fetch and push support pattern refspecs which allow fetching or pushing references that match a specific pattern. Because these patterns are globs, they have somewhat limited ability to express more complex situations. For example, suppose you wish to fetch all branches from a remote except for a specific one. To allow this, you must setup a set of refspecs which match only the branches you want. Because refspecs are either explicit name matches, or simple globs, many patterns cannot be expressed. Add support for a new type of refspec, referred to as "negative" refspecs. These are prefixed with a '^' and mean "exclude any ref matching this refspec". They can only have one "side" which always refers to the source. During a fetch, this refers to the name of the ref on the remote. During a push, this refers to the name of the ref on the local side. With negative refspecs, users can express more complex patterns. For example: git fetch origin refs/heads/*:refs/remotes/origin/* ^refs/heads/dontwant will fetch all branches on origin into remotes/origin, but will exclude fetching the branch named dontwant. Refspecs today are commutative, meaning that order doesn't expressly matter. Rather than forcing an implied order, negative refspecs will always be applied last. That is, in order to match, a ref must match at least one positive refspec, and match none of the negative refspecs. This is similar to how negative pathspecs work. Signed-off-by: Jacob Keller <jacob.keller@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-10-01 00:25:29 +03:00
/*
* Remove all entries in the input list which match any negative refspec in
* the refspec list.
*/
struct ref *apply_negative_refspecs(struct ref *ref_map, struct refspec *rs);
int query_refspecs(struct refspec *rs, struct refspec_item *query);
char *apply_refspecs(struct refspec *rs, const char *name);
int check_push_refs(struct ref *src, struct refspec *rs);
int match_push_refs(struct ref *src, struct ref **dst,
struct refspec *rs, int flags);
void set_ref_status_for_push(struct ref *remote_refs, int send_mirror,
int force_update);
/*
* Given a list of the remote refs and the specification of things to
* fetch, makes a (separate) list of the refs to fetch and the local
refspec: add support for negative refspecs Both fetch and push support pattern refspecs which allow fetching or pushing references that match a specific pattern. Because these patterns are globs, they have somewhat limited ability to express more complex situations. For example, suppose you wish to fetch all branches from a remote except for a specific one. To allow this, you must setup a set of refspecs which match only the branches you want. Because refspecs are either explicit name matches, or simple globs, many patterns cannot be expressed. Add support for a new type of refspec, referred to as "negative" refspecs. These are prefixed with a '^' and mean "exclude any ref matching this refspec". They can only have one "side" which always refers to the source. During a fetch, this refers to the name of the ref on the remote. During a push, this refers to the name of the ref on the local side. With negative refspecs, users can express more complex patterns. For example: git fetch origin refs/heads/*:refs/remotes/origin/* ^refs/heads/dontwant will fetch all branches on origin into remotes/origin, but will exclude fetching the branch named dontwant. Refspecs today are commutative, meaning that order doesn't expressly matter. Rather than forcing an implied order, negative refspecs will always be applied last. That is, in order to match, a ref must match at least one positive refspec, and match none of the negative refspecs. This is similar to how negative pathspecs work. Signed-off-by: Jacob Keller <jacob.keller@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-10-01 00:25:29 +03:00
* refs to store into. Note that negative refspecs are ignored here, and
* should be handled separately.
*
* *tail is the pointer to the tail pointer of the list of results
* beforehand, and will be set to the tail pointer of the list of
* results afterward.
*
* missing_ok is usually false, but when we are adding branch.$name.merge
* it is Ok if the branch is not at the remote anymore.
*/
int get_fetch_map(const struct ref *remote_refs, const struct refspec_item *refspec,
struct ref ***tail, int missing_ok);
struct ref *get_remote_ref(const struct ref *remote_refs, const char *name);
/*
* For the given remote, reads the refspec's src and sets the other fields.
*/
int remote_find_tracking(struct remote *remote, struct refspec_item *refspec);
/**
* struct branch holds the configuration for a branch. It can be looked up with
* branch_get(name) for "refs/heads/{name}", or with branch_get(NULL) for HEAD.
*/
struct branch {
struct hashmap_entry ent;
/* The short name of the branch. */
const char *name;
/* The full path for the branch ref. */
const char *refname;
/* The name of the remote listed in the configuration. */
const char *remote_name;
const char *pushremote_name;
/* An array of the "merge" lines in the configuration. */
const char **merge_name;
/**
* An array of the struct refspecs used for the merge lines. That is,
* merge[i]->dst is a local tracking ref which should be merged into this
* branch by default.
*/
struct refspec_item **merge;
/* The number of merge configurations */
int merge_nr;
int merge_alloc;
const char *push_tracking_ref;
};
struct branch *branch_get(const char *name);
const char *remote_for_branch(struct branch *branch, int *explicit);
const char *pushremote_for_branch(struct branch *branch, int *explicit);
const char *remote_ref_for_branch(struct branch *branch, int for_push);
/* returns true if the given branch has merge configuration given. */
int branch_has_merge_config(struct branch *branch);
Correct handling of branch.$name.merge in builtin-fetch My prior bug fix for git-push titled "Don't configure remote "." to fetch everything to itself" actually broke t5520 as we were unable to evaluate a branch configuration of: [branch "copy"] remote = . merge = refs/heads/master as remote "." did not have a "remote...fetch" configuration entry to offer up refs/heads/master as a possible candidate available to be fetched and merged. In shell script git-fetch and prior to the above mentioned commit this was hardcoded for a url of "." to be the set of local branches. Chasing down this bug led me to the conclusion that our prior behavior with regards to branch.$name.merge was incorrect. In the shell script based git-fetch implementation we only fetched and merged a branch if it appeared both in branch.$name.merge *and* in remote.$r.fetch, where $r = branch.$name.remote. In other words in the following config file: [remote "origin"] url = git://git.kernel.org/pub/scm/git/git.git fetch = refs/heads/master:refs/remotes/origin/master [branch "master"] remote = origin merge = refs/heads/master [branch "pu"] remote = origin merge = refs/heads/pu Attempting to run `git pull` while on branch "pu" would always give the user "Already up-to-date" as git-fetch did not fetch pu and thus did not mark it for merge in .git/FETCH_HEAD. The configured merge would always be ignored and the user would be left scratching her confused head wondering why merge did not work on "pu" but worked fine on "master". If we are using the "default fetch" specification for the current branch and the current branch has a branch.$name.merge configured we now union it with the list of refs in remote.$r.fetch. This way the above configuration does what the user expects it to do, which is to fetch only "master" by default but when on "pu" to fetch both "master" and "pu". This uncovered some breakage in the test suite where old-style Cogito branches (.git/branches/$r) did not fetch the branches listed in .git/config for merging and thus did not actually merge them if the user tried to use `git pull` on that branch. Junio and I discussed it on list and felt that the union approach here makes more sense to DWIM for the end-user than silently ignoring their configured request so the test vectors for t5515 have been updated to include for-merge lines in .git/FETCH_HEAD where they have been configured for-merge in .git/config. Since we are now performing a union of the fetch specification and the merge specification and we cannot allow a branch to be listed twice (otherwise it comes out twice in .git/FETCH_HEAD) we need to perform a double loop here over all of the branch.$name.merge lines and try to set their merge flag if we have already schedule that branch for fetching by remote.$r.fetch. If no match is found then we must add new specifications to fetch the branch but not store it as no local tracking branch has been designated. Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2007-09-18 12:54:53 +04:00
int branch_merge_matches(struct branch *, int n, const char *);
/**
* Return the fully-qualified refname of the tracking branch for `branch`.
* I.e., what "branch@{upstream}" would give you. Returns NULL if no
* upstream is defined.
*
* If `err` is not NULL and no upstream is defined, a more specific error
* message is recorded there (if the function does not return NULL, then
* `err` is not touched).
*/
const char *branch_get_upstream(struct branch *branch, struct strbuf *err);
/**
* Return the tracking branch that corresponds to the ref we would push to
* given a bare `git push` while `branch` is checked out.
*
* The return value and `err` conventions match those of `branch_get_upstream`.
*/
const char *branch_get_push(struct branch *branch, struct strbuf *err);
/* Flags to match_refs. */
enum match_refs_flags {
MATCH_REFS_NONE = 0,
MATCH_REFS_ALL = (1 << 0),
MATCH_REFS_MIRROR = (1 << 1),
MATCH_REFS_PRUNE = (1 << 2),
MATCH_REFS_FOLLOW_TAGS = (1 << 3)
};
/* Flags for --ahead-behind option. */
enum ahead_behind_flags {
AHEAD_BEHIND_UNSPECIFIED = -1,
AHEAD_BEHIND_QUICK = 0, /* just eq/neq reporting */
AHEAD_BEHIND_FULL = 1, /* traditional a/b reporting */
};
/* Reporting of tracking info */
int stat_tracking_info(struct branch *branch, int *num_ours, int *num_theirs,
const char **upstream_name, int for_push,
enum ahead_behind_flags abf);
int format_tracking_info(struct branch *branch, struct strbuf *sb,
enum ahead_behind_flags abf);
struct ref *get_local_heads(void);
/*
* Find refs from a list which are likely to be pointed to by the given HEAD
* ref. If 'all' is false, returns the most likely ref; otherwise, returns a
* list of all candidate refs. If no match is found (or 'head' is NULL),
* returns NULL. All returns are newly allocated and should be freed.
*/
struct ref *guess_remote_head(const struct ref *head,
const struct ref *refs,
int all);
/* Return refs which no longer exist on remote */
struct ref *get_stale_heads(struct refspec *rs, struct ref *fetch_map);
remote.c: add command line option parser for "--force-with-lease" Update "git push" and "git send-pack" to parse this commnd line option. The intended sematics is: * "--force-with-lease" alone, without specifying the details, will protect _all_ remote refs that are going to be updated by requiring their current value to be the same as some reasonable default, unless otherwise specified; * "--force-with-lease=refname", without specifying the expected value, will protect that refname, if it is going to be updated, by requiring its current value to be the same as some reasonable default. * "--force-with-lease=refname:value" will protect that refname, if it is going to be updated, by requiring its current value to be the same as the specified value; and * "--no-force-with-lease" will cancel all the previous --force-with-lease on the command line. For now, "some reasonable default" is tentatively defined as "the value of the remote-tracking branch we have for the ref of the remote being updated", and it is an error if we do not have such a remote-tracking branch. But this is known to be fragile, its use is not yet recommended, and hopefully we will find more reasonable default as we gain experience with this feature. The manual marks the feature as experimental unless the expected value is specified explicitly for this reason. Because the command line options are parsed _before_ we know which remote we are pushing to, there needs further processing to the parsed data after we instantiate the transport object to: * expand "refname" given by the user to a full refname to be matched with the list of "struct ref" used in match_push_refs() and set_ref_status_for_push(); and * learning the actual local ref that is the remote-tracking branch for the specified remote ref. Further, some processing need to be deferred until we find the set of remote refs and match_push_refs() returns in order to find the ones that need to be checked after explicit ones have been processed for "--force-with-lease" (no specific details). These post-processing will be the topic of the next patch. This option was originally called "cas" (for "compare and swap"), the name which nobody liked because it was too technical. The second attempt called it "lockref" (because it is conceptually like pushing after taking a lock) but the word "lock" was hated because it implied that it may reject push by others, which is not the way this option works. This round calls it "force-with-lease". You assume you took the lease on the ref when you fetched to decide what the rebased history should be, and you can push back only if the lease has not been broken. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-07-09 02:34:36 +04:00
/*
* Compare-and-swap
*/
#define CAS_OPT_NAME "force-with-lease"
struct push_cas_option {
unsigned use_tracking_for_rest:1;
unsigned use_force_if_includes:1;
remote.c: add command line option parser for "--force-with-lease" Update "git push" and "git send-pack" to parse this commnd line option. The intended sematics is: * "--force-with-lease" alone, without specifying the details, will protect _all_ remote refs that are going to be updated by requiring their current value to be the same as some reasonable default, unless otherwise specified; * "--force-with-lease=refname", without specifying the expected value, will protect that refname, if it is going to be updated, by requiring its current value to be the same as some reasonable default. * "--force-with-lease=refname:value" will protect that refname, if it is going to be updated, by requiring its current value to be the same as the specified value; and * "--no-force-with-lease" will cancel all the previous --force-with-lease on the command line. For now, "some reasonable default" is tentatively defined as "the value of the remote-tracking branch we have for the ref of the remote being updated", and it is an error if we do not have such a remote-tracking branch. But this is known to be fragile, its use is not yet recommended, and hopefully we will find more reasonable default as we gain experience with this feature. The manual marks the feature as experimental unless the expected value is specified explicitly for this reason. Because the command line options are parsed _before_ we know which remote we are pushing to, there needs further processing to the parsed data after we instantiate the transport object to: * expand "refname" given by the user to a full refname to be matched with the list of "struct ref" used in match_push_refs() and set_ref_status_for_push(); and * learning the actual local ref that is the remote-tracking branch for the specified remote ref. Further, some processing need to be deferred until we find the set of remote refs and match_push_refs() returns in order to find the ones that need to be checked after explicit ones have been processed for "--force-with-lease" (no specific details). These post-processing will be the topic of the next patch. This option was originally called "cas" (for "compare and swap"), the name which nobody liked because it was too technical. The second attempt called it "lockref" (because it is conceptually like pushing after taking a lock) but the word "lock" was hated because it implied that it may reject push by others, which is not the way this option works. This round calls it "force-with-lease". You assume you took the lease on the ref when you fetched to decide what the rebased history should be, and you can push back only if the lease has not been broken. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-07-09 02:34:36 +04:00
struct push_cas {
struct object_id expect;
remote.c: add command line option parser for "--force-with-lease" Update "git push" and "git send-pack" to parse this commnd line option. The intended sematics is: * "--force-with-lease" alone, without specifying the details, will protect _all_ remote refs that are going to be updated by requiring their current value to be the same as some reasonable default, unless otherwise specified; * "--force-with-lease=refname", without specifying the expected value, will protect that refname, if it is going to be updated, by requiring its current value to be the same as some reasonable default. * "--force-with-lease=refname:value" will protect that refname, if it is going to be updated, by requiring its current value to be the same as the specified value; and * "--no-force-with-lease" will cancel all the previous --force-with-lease on the command line. For now, "some reasonable default" is tentatively defined as "the value of the remote-tracking branch we have for the ref of the remote being updated", and it is an error if we do not have such a remote-tracking branch. But this is known to be fragile, its use is not yet recommended, and hopefully we will find more reasonable default as we gain experience with this feature. The manual marks the feature as experimental unless the expected value is specified explicitly for this reason. Because the command line options are parsed _before_ we know which remote we are pushing to, there needs further processing to the parsed data after we instantiate the transport object to: * expand "refname" given by the user to a full refname to be matched with the list of "struct ref" used in match_push_refs() and set_ref_status_for_push(); and * learning the actual local ref that is the remote-tracking branch for the specified remote ref. Further, some processing need to be deferred until we find the set of remote refs and match_push_refs() returns in order to find the ones that need to be checked after explicit ones have been processed for "--force-with-lease" (no specific details). These post-processing will be the topic of the next patch. This option was originally called "cas" (for "compare and swap"), the name which nobody liked because it was too technical. The second attempt called it "lockref" (because it is conceptually like pushing after taking a lock) but the word "lock" was hated because it implied that it may reject push by others, which is not the way this option works. This round calls it "force-with-lease". You assume you took the lease on the ref when you fetched to decide what the rebased history should be, and you can push back only if the lease has not been broken. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-07-09 02:34:36 +04:00
unsigned use_tracking:1;
char *refname;
} *entry;
int nr;
int alloc;
};
int parseopt_push_cas_option(const struct option *, const char *arg, int unset);
remote.c: add command line option parser for "--force-with-lease" Update "git push" and "git send-pack" to parse this commnd line option. The intended sematics is: * "--force-with-lease" alone, without specifying the details, will protect _all_ remote refs that are going to be updated by requiring their current value to be the same as some reasonable default, unless otherwise specified; * "--force-with-lease=refname", without specifying the expected value, will protect that refname, if it is going to be updated, by requiring its current value to be the same as some reasonable default. * "--force-with-lease=refname:value" will protect that refname, if it is going to be updated, by requiring its current value to be the same as the specified value; and * "--no-force-with-lease" will cancel all the previous --force-with-lease on the command line. For now, "some reasonable default" is tentatively defined as "the value of the remote-tracking branch we have for the ref of the remote being updated", and it is an error if we do not have such a remote-tracking branch. But this is known to be fragile, its use is not yet recommended, and hopefully we will find more reasonable default as we gain experience with this feature. The manual marks the feature as experimental unless the expected value is specified explicitly for this reason. Because the command line options are parsed _before_ we know which remote we are pushing to, there needs further processing to the parsed data after we instantiate the transport object to: * expand "refname" given by the user to a full refname to be matched with the list of "struct ref" used in match_push_refs() and set_ref_status_for_push(); and * learning the actual local ref that is the remote-tracking branch for the specified remote ref. Further, some processing need to be deferred until we find the set of remote refs and match_push_refs() returns in order to find the ones that need to be checked after explicit ones have been processed for "--force-with-lease" (no specific details). These post-processing will be the topic of the next patch. This option was originally called "cas" (for "compare and swap"), the name which nobody liked because it was too technical. The second attempt called it "lockref" (because it is conceptually like pushing after taking a lock) but the word "lock" was hated because it implied that it may reject push by others, which is not the way this option works. This round calls it "force-with-lease". You assume you took the lease on the ref when you fetched to decide what the rebased history should be, and you can push back only if the lease has not been broken. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-07-09 02:34:36 +04:00
int is_empty_cas(const struct push_cas_option *);
void apply_push_cas(struct push_cas_option *, struct remote *, struct ref *);
#endif