This makes the algorithm more honest about what it is doing.
We start from an already limited, topo-sorted list, and postprocess
it by simplifying the irrelevant merges away.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The --full-history traversal keeps all merges in addition to non-merge
commits that touch paths in the given pathspec. This is useful to view
both sides of a merge in a topology like this:
A---M---o
/ /
---O---B
even when A and B makes identical change to the given paths. The revision
traversal without --full-history aims to come up with the simplest history
to explain the final state of the tree, and one of the side branches can
be pruned away.
The behaviour to keep all merges however is inconvenient if neither A nor
B touches the paths we are interested in. --full-history reduces the
topology to:
---O---M---o
in such a case, without removing M.
This adds a post processing phase on top of --full-history traversal to
remove needless merges from the resulting history.
The idea is to compute, for each commit in the "full history" result set,
the commit that should replace it in the simplified history. The commit
to replace it in the final history is determined as follows:
* In any case, we first figure out the replacement commits of parents of
the commit we are looking at. The commit we are looking at is
rewritten as if the replacement commits of its original parents are its
parents. While doing so, we reduce the redundant parents from the
rewritten parent list by not just removing the identical ones, but also
removing a parent that is an ancestor of another parent.
* After the above parent simplification, if the commit is a root commit,
an UNINTERESTING commit, a merge commit, or modifies the paths we are
interested in, then the replacement commit of the commit is itself. In
other words, such a commit is not dropped from the final result.
The first point above essentially means that the history is rewritten in
the bottom up direction. We can rewrite the parent list of a commit only
after we know how all of its parents are rewritten. This means that the
processing needs to happen on the full history (i.e. after limit_list()).
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Commands which use parse_options() but also call setup_revisions()
must do their parsing in a two step process:
1. first, they parse all options. Anything unknown goes to
parse_revision_opt() (which calls handle_revision_opt), which
may claim the option or say "I don't recognize this"
2. the non-option remainder goes to setup_revisions() to
actually get turned into revisions
Some revision options are "non-options" in that they must be
parsed in order with their revision counterparts in
setup_revisions(). For example, "--all" functions as a
pseudo-option expanding to all refs, and "--no-walk" affects refs
after it on the command line, but not before. The revision option
parser in step 1 recognizes such options and sets them aside for
later parsing by setup_revisions().
However, the return value used from handle_revision_opt indicated
"I didn't recognize this", which was wrong. It did, and it took
appropriate action (even though that action was just deferring it
for later parsing). Thus it should return "yes, I recognized
this."
Previously, these pseudo-options generated an error when used with
parse_options parsers (currently just blame and shortlog). With
this patch, they should work fine, enabling things like "git
shortlog --all".
Signed-off-by: Jeff King <peff@peff.net>
Acked-By: Pierre Habouzit <madcoder@debian.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
* maint:
Start preparing 1.5.6.4 release notes
git fetch-pack: do not complain about "no common commits" in an empty repo
rebase-i: keep old parents when preserving merges
t7600-merge: Use test_expect_failure to test option parsing
Fix buffer overflow in prepare_attr_stack
Fix buffer overflow in git diff
Fix buffer overflow in git-grep
git-cvsserver: fix call to nonexistant cleanupWorkDir()
Documentation/git-cherry-pick.txt et al.: Fix misleading -n description
Conflicts:
RelNotes
If PATH_MAX on your system is smaller than a path stored, it may cause
buffer overflow and stack corruption in diff_addremove() and diff_change()
functions when running git-diff
Signed-off-by: Dmitry Potapov <dpotapov@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
* ph/parseopt-step-blame:
revisions: refactor handle_revision_opt into parse_revision_opt.
git-shortlog: migrate to parse-options partially.
git-blame: fix lapsus
git-blame: migrate to incremental parse-option [2/2]
git-blame: migrate to incremental parse-option [1/2]
revisions: split handle_revision_opt() from setup_revisions()
parse-opt: add PARSE_OPT_KEEP_ARGV0 parser option.
parse-opt: fake short strings for callers to believe in.
parse-opt: do not print errors on unknown options, return -2 intead.
parse-opt: create parse_options_step.
parse-opt: Export a non NORETURN usage dumper.
parse-opt: have parse_options_{start,end}.
git-blame --reverse
builtin-blame.c: allow more than 16 parents
builtin-blame.c: move prepare_final() into a separate function.
rev-list --children
revision traversal: --children option
Parent commits are usually older than their children. Thus,
on each iteration of the loop in rewrite_one, add_parents_to_list
traverses all commits previously processed by the loop.
It performs very poorly in case of very long rewrite chains.
Signed-off-by: Alexander Gavrilov <angavrilov@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
It seems we're using handle_revision_opt the same way each time, have a
wrapper around it that does the 9-liner we copy each time instead.
handle_revision_opt can be static in the module for now, it's always
possible to make it public again if needed.
Signed-off-by: Pierre Habouzit <madcoder@debian.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Add two fields to struct rev_info:
- .def to store --default argument; and
- .show_merge 1-bit field.
handle_revision_opt() is able to deal with any revision option, and
consumes them, and leaves revision arguments or pseudo arguments
(like --all, --not, ...) in place.
For now setup_revisions() does a pass of handle_revision_opt() again
so that code not using it in a parse-opt parser still work the same.
Signed-off-by: Pierre Habouzit <madcoder@debian.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Reading rev-list parameters from the command line can be reused by
commands other than rev-list. Move this function to more "library-ish"
place to promote code reuse.
Signed-off-by: Adam Brewster <asb@bu.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Now get_revision() sorts the boundary commits when topo_order is set.
Since sort_in_topological_order() takes a struct commit_list, it first
places the boundary commits into revs->commits.
Signed-off-by: Adam Simpkins <adam@adamsimpkins.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Previously the graphing API wasn't aware of the revs->boundary flag, and
it always assumed that commits marked UNINTERESTING would not be
displayed. As a result, the boundary commits were printed at the end of
the log output, but they didn't have any branch lines connecting them to
their children in the graph.
There was also another bug in the get_revision() code that caused
graph_update() to be called twice on the first boundary commit. This
caused the graph API to think that a commit had been skipped, and print
a "..." line in the output.
Signed-off-by: Adam Simpkins <adam@adamsimpkins.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
With the --graph option, the graph already outputs 'o' instead of '*'
for boundary commits. Make it emit '<' or '>' when --left-right is
specified.
(This change also disables the '^' prefix for UNINTERESTING commits.
The graph code currently doesn't print anything special for these
commits, since it assumes no UNINTERESTING, non-BOUNDARY commits are
displayed. This is potentially a bug if UNINTERESTING non-BOUNDARY
commits can actually be displayed via some code path.)
[jc: squashed the left-right change from Dscho and Adam's fixup into one]
Signed-off-by: Adam Simpkins <adam@adamsimpkins.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In add_parents_to_list, if any parent of a revision had already been
SEEN, the current code would continue with the next parent, skipping
the test for --first-parent. This patch inverts the test for SEEN so
that the test for --first-parent is always performed.
Signed-off-by: Lars Hjemli <hjemli@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This new option causes a text-based representation of the history to be
printed to the left of the normal output.
Signed-off-by: Adam Simpkins <adam@adamsimpkins.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This change allows parent rewriting to be performed without causing
the log and rev-list commands to print the parents.
Signed-off-by: Adam Simpkins <adam@adamsimpkins.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The purpose of --first-parent is to view the tree without looking at
side branche. This is accomplished by pretending there are no other
parents than the first parent when encountering a merge.
The current code marks the other parents as seen, which means that the tree
traversal will behave differently depending on the order merges are handled.
When a fast forward is artificially recorded as a merge,
-----
/ \
D---E---F---G master
the current first-parent code considers E to be seen and stops the
traversal after showing G and F.
Signed-off-by: Stephen R. van den Berg <srb@cuci.nl>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This adds a new --children option to the revision machinery. In addition
to the list of parents, child commits of each commit are computed and
stored as a decoration to each commit.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
* maint-1.5.4:
bisect: fix bad rev checking in "git bisect good"
revision.c: make --date-order overriddable
Fix section about backdating tags in the git-tag docs
Document option --only of git commit
Documentation/git-request-pull: Fixed a typo ("send" -> "end")
Jan Engelhardt noticed that while --topo-order can be overridden by a
subsequent --date-order, the reverse was not possible. That's because
setup_revisions() failed to set revs->lifo properly.
Signed-off-by: Michele Ballabio <barra_cuda@katamail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This attached patch introduces a single bit "use_terminator" in "struct
rev_info", which is normally false (i.e. most formats use separator
semantics) but by flipping it to true, you can ask for terminator
semantics just like oneline format does.
The function get_commit_format(), which is what parses "--pretty=" option,
now takes a pointer to "struct rev_info" and updates its commit_format and
use_terminator fields. It used to return the value of type "enum
cmit_fmt", but all the callers assigned it to rev->commit_format.
There are only two cases the code turns use_terminator on. Obviously, the
traditional oneline format (--pretty=oneline) is one of them, and the new
case is --pretty=tformat:... that acts like --pretty=format:... but flips
the bit on.
With this, "--pretty=tformat:%H %s" acts like --pretty=oneline.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The revision limiter uses the commit date to decide when it has seen
enough commits to finalize the revision list, but that can get confused
if there are incorrect dates far in the past on some commits.
This makes the logic a bit more robust by
- we always walk an extra SLOP commits from the source list even if we
decide that the source list is probably all done (unless the source is
entirely empty, of course, because then we really can't do anything at
all)
- we keep track of the date of the last commit we added to the
destination list (this will *generally* be the oldest entry we've seen
so far)
- we compare that with the youngest entry (the first one) of the source
list, and if the destination is older than the source, we know we want
to look at the source.
which causes occasional date mishaps to be handled cleanly.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
These flags are already known to rev-parse and have the same meaning.
This patch allows to run gitk as follows:
gitk --branches --not --remotes
to show only your local work.
Signed-off-by: Uwe Kleine-König <Uwe.Kleine-Koenig@digi.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Add support for -F | --fixed-strings option to "git log --grep"
and friends: "git log --author", "git log --committer=<pattern>".
Code is based on implementation of this option in "git grep".
Signed-off-by: Jakub Narebski <jnareb@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
As these functions are directly called with the result
from lookup_tree/blob, they must handle NULL.
Signed-off-by: Martin Koegler <mkoegler@auto.tuwien.ac.at>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
It's really not very easy to visualize the commit walker, because - on
purpose - it obvously doesn't show the uninteresting commits!
This adds a "--show-all" flag to the revision walker, which will make
it show uninteresting commits too, and they'll have a '^' in front of
them (it also fixes a logic error for !verbose_header for boundary
commits - we should show the '-' even if left_right isn't shown).
A separate patch to gitk to teach it the new '^' was sent
to paulus. With the change in place, it actually is interesting
even for the cases that git doesn't have any problems with, ie
for the kernel you can do:
gitk -d --show-all v2.6.24..
and you see just how far down it has to parse things to see it all. The
use of "-d" is a good idea, since the date-ordered toposort is much better
at showing why it goes deep down (ie the date of some of those commits
after 2.6.24 is much older, because they were merged from trees that
weren't rebased).
So I think this is a useful feature even for non-debugging - just to
visualize what git does internally more.
When it actually breaks out due to the "everybody_uninteresting()"
case, it adds the uninteresting commits (both the one it's looking at
now, and the list of pending ones) to the list
This way, we really list *all* the commits we've looked at.
Because we now end up listing commits we may not even have been parsed
at all "show_log" and "show_commit" need to protect against commits
that don't have a commit buffer entry.
That second part is debatable just how it should work. Maybe we shouldn't
show such entries at all (with this patch those entries do get shown, they
just don't get any message shown with them). But I think this is a useful
case.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This allows the --relative option to say which subdirectory to
pretend to be in, so that in a bare repository, you can say:
$ git log --relative=drivers/ v2.6.20..v2.6.22 -- drivers/scsi/
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This adds --relative option to the diff family. When you start
from a subdirectory:
$ git diff --relative
shows only the diff that is inside your current subdirectory,
and without $prefix part. People who usually live in
subdirectories may like it.
There are a few things I should also mention about the change:
- This works not just with diff but also works with the log
family of commands, but the history pruning is not affected.
In other words, if you go to a subdirectory, you can say:
$ git log --relative -p
but it will show the log message even for commits that do not
touch the current directory. You can limit it by giving
pathspec yourself:
$ git log --relative -p .
This originally was not a conscious design choice, but we
have a way to affect diff pathspec and pruning pathspec
independently. IOW "git log --full-diff -p ." tells it to
prune history to commits that affect the current subdirectory
but show the changes with full context. I think it makes
more sense to leave pruning independent from --relative than
the obvious alternative of always pruning with the current
subdirectory, which would break the symmetry.
- Because this works also with the log family, you could
format-patch a single change, limiting the effect to your
subdirectory, like so:
$ cd gitk-git
$ git format-patch -1 --relative 911f1eb
But because that is a special purpose usage, this option will
never become the default, with or without repository or user
preference configuration. The risk of producing a partial
patch and sending it out by mistake is too great if we did
so.
- This is inherently incompatible with --no-index, which is a
bolted-on hack that does not have much to do with git
itself. I didn't bother checking and erroring out on the
combined use of the options, but probably I should.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In commit b7bb760d5e (Fix revision
log diff setup, avoid unnecessary diff generation) an optimization was
made to avoid unnecessary diff generation. This was partly fixed in
99516e35d0 (Fix embarrassing "git log
--follow" bug). The '--diff-filter' option also needs the diff machinery
in action.
Signed-off-by: Arjen Laarhoven <arjen@yaph.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Instead of warning the user that it is expecting git log output from
the standard input (and waiting for the user to type the log from
the keyboard, which is a silly thing to do), default to traverse from
HEAD when there is no rev parameter given and the standard input is
a tty.
This factors out a useful helper "add_head()" from builtin-diff.c to a
more appropriate place revision.c while renaming it to more descriptive
name add_head_to_pending(), as that is what the function is about.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
* lt/rev-list-interactive:
Fix parent rewriting in --early-output
revision walker: mini clean-up
Enhance --early-output format
Add "--early-output" log flag for interactive GUI use
Simplify topo-sort logic
We cannot tell a node that has been checked and found not to be
interesting (which does not have the TREECHANGE flag) from a
node that hasn't been checked if it is interesting or not,
without relying on something else, such as object->parsed.
But an object can get the "parsed" flag for other reasons.
Which means that "TREECHANGE" has the wrong polarity.
This changes the way how the path pruning logic marks an
uninteresting commits. From now on, we consider a commit
interesting by default, and explicitly mark the ones we decided
to prune. The flag is renamed to "TREESAME".
Then, this fixes the logic to show the early output with
incomplete pruning. It basically says "a commit that has
TREESAME set is kind-of-UNINTERESTING", but obviously in a
different way than an outright UNINTERESTING commit. Until we
parse and examine enough parents to determine if a commit
becomes surely "kind-of-UNINTERESTING", we avoid rewriting
the ancestry so that later rounds can fix things up.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The function mark_tree_uninteresting() assumed that the tree entries
are blob when they are not trees. This is not so. Since we do
not traverse into submodules (yet), the gitlinks should be ignored.
In general, we should try to start moving away from using the
"S_ISLNK()" like things for internal git state. It was a mistake to
just assume the numbers all were same across all systems in the first
place. This implementation converts to the "object_type", and then
uses a case statement.
Noticed by Ilari on IRC.
Test script taken from an earlier version by Dscho.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
reverse_diff was a bit-value in disguise, it's merged in the flags now.
Signed-off-by: Pierre Habouzit <madcoder@debian.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This removes the unnecessary indirection of "revs->prune_fn",
since that function is always the same one (or NULL), and there
is in fact not even an abstraction reason to make it a function
(i.e. its not called from some other file and doesn't allow us
to keep the function itself static or anything like that).
It then just replaces it with a bit that says "prune or not",
and if not pruning, every commit gets TREECHANGE.
That in turn means that
- if (!revs->prune_fn || (flags & TREECHANGE))
- if (revs->prune_fn && !(flags & TREECHANGE))
just become
- if (flags & TREECHANGE)
- if (!(flags & TREECHANGE))
respectively.
Together with adding the "single_parent()" helper function, the "complex"
conditional now becomes
if (!(flags & TREECHANGE) && rev->dense && single_parent(commit))
continue;
Also indirection of "revs->dense" checking is thrown away the
same way, because TREECHANGE bit is set appropriately now.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This makes --early-output a bit more advanced, and actually makes it
generate multiple "Final output:" headers as it updates things
asynchronously. I realize that the "Final output:" line is now illogical,
since it's not really final until it also says "done", but
It now _always_ generates a "Final output:" header in front of any commit
list, and that output header gives you a *guess* at the maximum number of
commits available. However, it should be noted that the guess can be
completely off: I do a reasonable job estimating it, but it is not meant
to be exact.
So what happens is that you may get output like this:
- at 0.1 seconds:
Final output: 2 incomplete
.. 2 commits listed ..
- half a second later:
Final output: 33 incomplete
.. 33 commits listed ..
- another half a second after that:
Final output: 71 incomplete
.. 71 commits listed ..
- another half second later:
Final output: 136 incomplete
.. 100 commits listed: we hit the --early-output limit, and
.. will only output 100 commits, and after this you'll not
.. see an "incomplete" report any more since you got as much
.. early output as you asked for!
- .. and then finally:
Final output: 73106 done
.. all the commits ..
The above is a real-life scenario on my current kernel tree after having
flushed all the caches.
Tested with the experimental gitk patch that Paul sent out, and by looking
at the actual log output (and verifying that my commit count guesses
actually match real life fairly well).
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This adds support for "--early-output[=n]" as a flag to the "git log"
family of commands. This allows GUI programs to state that they want to
get some output early, in order to be able to show at least something
quickly, even if the full output may take longer to generate.
If no count is specified, a default count of a hundred commits will be
used, although the actual numbr of commits output may be smaller
depending on how many commits were actually found in the first tenth of
a second (or if *everything* was found before that, in which case no
early output will be provided, and only the final list is made
available).
When the full list is generated, there will be a "Final output:" string
prepended to it, regardless of whether any early commits were shown or
not, so that the consumer can always know the difference between early
output and the final list.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Junio C Hamano
Pierre Habouzit
Dmitry Potapov
Alexander N. Gavrilov
Adam Brewster
Adam Simpkins
Lars Hjemli
Stephen R. van den Berg
Michele Ballabio
Linus Torvalds
Uwe Kleine-König
Jakub Narebski
Martin Koegler
Arjen Laarhoven