"git help $cmd" unnecessarily enumerated potential command names
from the filesystem, even when $cmd is known to be a built-in.
Ideas for further optimization, primarily by killing the use of
is_in_cmdlist(), were suggested in the discussion, but they can
come as follow-ups on top of this series.
* ss/builtin-cleanup:
builtin/help.c: speed up is_git_command() by checking for builtin commands first
builtin/help.c: call load_command_list() only when it is needed
git.c: consistently use the term "builtin" instead of "internal command"
Update the way the user-manual is formatted via AsciiDoc to save
trees.
* ta/format-user-manual-as-an-article:
user-manual: improve html and pdf formatting
Two-level configuration variable names in "branch.*" and "remote.*"
hierarchies whose variables are predominantly three-level where not
completed by hitting a <TAB> in bash and zsh completions.
* rr/completion-branch-config:
completion: fix remote.pushdefault
completion: fix branch.autosetup(merge|rebase)
completion: introduce __gitcomp_nl_append ()
zsh completion: find matching custom bash completion
There is no reason to have a hardcoded upper limit of the number of
parents for an octopus merge, created via the graft mechanism.
* js/lift-parent-count-limit:
Remove the line length limit for graft files
The basic test used to leave unnecessary trash directories in the
t/ directory.
* jk/test-framework-updates:
t0000: drop "known breakage" test
t0000: simplify HARNESS_ACTIVE hack
t0000: set TEST_OUTPUT_DIRECTORY for sub-tests
"git merge-base --octopus" used to leave cleaning up suboptimal
result to the caller, but now it does the clean-up itself.
* bm/merge-base-octopus-dedup:
merge-base --octopus: reduce the result from get_octopus_merge_bases()
merge-base: separate "--independent" codepath into its own helper
A "gc" process running as a different user should be able to stop a
new "gc" process from starting.
* km/gc-eperm:
gc: notice gc processes run by other users
Using the same username and password during the tests would not
catch a potential breakage of sending one when we should be sending
the other.
* jk/http-auth-tests-robustify:
use distinct username/password for http auth tests
Teach "cat-file --batch" to show delta-base object name for a
packed object that is represented as a delta.
* jk/oi-delta-base:
cat-file: provide %(deltabase) batch format
sha1_object_info_extended: provide delta base sha1s
"git add -A" (no other arguments) in a totally empty working tree
used to emit an error.
* nd/add-empty-fix:
add: don't complain when adding empty project root
Fetching 'frotz' branch with "git fetch", while having
'frotz/nitfol' remote-tracking branch from an earlier fetch, would
error out, primarily because the command has not been told to
remove anything on our side. In such a case, "git fetch --prune"
can be used to remove 'frotz/nitfol' to make room to fetch and
store 'frotz' remote-tracking branch.
* tm/fetch-prune:
fetch --prune: Run prune before fetching
fetch --prune: always print header url
Allow "git diff -O<file>" to be configured with a new configuration
variable.
* sb/diff-orderfile-config:
diff: add diff.orderfile configuration variable
diff: let "git diff -O" read orderfile from any file and fail properly
t4056: add new tests for "git diff -O"
When we figure out how many file descriptors to allocate for
keeping packfiles open, a system with non-working getrlimit() could
cause us to die(), but because we make this call only to get a
rough estimate of how many is available and we do not even attempt
to use up all file descriptors available ourselves, it is nicer to
fall back to a reasonable low value rather than dying.
* jh/rlimit-nofile-fallback:
get_max_fd_limit(): fall back to OPEN_MAX upon getrlimit/sysconf failure
A few places where we relied on a fixed length buffer to hold
pathnames in these two programs have been converted to use strbuf.
* mh/path-max:
builtin/prune.c: use strbuf to avoid having to worry about PATH_MAX
prune-packed: use strbuf to avoid having to worry about PATH_MAX
read_sha1_file() that is the workhorse to read the contents given
an object name honoured object replacements, but there is no
corresponding mechanism to sha1_object_info() that is used to
obtain the metainfo (e.g. type & size) about the object, leading
callers to weird inconsistencies.
* cc/replace-object-info:
replace info: rename 'full' to 'long' and clarify in-code symbols
Documentation/git-replace: describe --format option
builtin/replace: unset read_replace_refs
t6050: add tests for listing with --format
builtin/replace: teach listing using short, medium or full formats
sha1_file: perform object replacement in sha1_object_info_extended()
t6050: show that git cat-file --batch fails with replace objects
sha1_object_info_extended(): add an "unsigned flags" parameter
sha1_file.c: add lookup_replace_object_extended() to pass flags
replace_object: don't check read_replace_refs twice
rename READ_SHA1_FILE_REPLACE flag to LOOKUP_REPLACE_OBJECT
Introduce "negative pathspec" magic, to allow "git log -- . ':!dir'" to
tell us "I am interested in everything but 'dir' directory".
* nd/negative-pathspec:
pathspec.c: support adding prefix magic to a pathspec with mnemonic magic
Support pathspec magic :(exclude) and its short form :!
glossary-content.txt: rephrase magic signature part
When no arguments are specified, $switch_to is empty so we end up
passing the empty string to "git merge-base --fork-point", which causes
an error. git-rebase carries on at this point, but in fact we have
failed to apply the fork-point operation.
It turns out that the test in t3400 that was meant to test this didn't
actually need the fork-point behaviour, so enhance it to make sure that
the fork-point is applied correctly. The modified test fails without
the change to git-rebase.sh in this patch.
Reported-by: Andreas Krey <a.krey@gmx.de>
Signed-off-by: John Keeping <john@keeping.me.uk>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
As long as we're being pathologically stingy with mallocs, we might as
well do the math right and save 6 (!) bytes.
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
To replace "%.*s" with "%s", all we have to do is use snprintf()
to interpolate "%s" into the pattern.
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When filling the scanf_fmts array, use a separate variable to keep
track of the offset to avoid clobbering total_len (which we will need
in the next commit).
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Commit 48d25ca adds a new commit "7" to the repo that the next test case
in commit 1609488 clones from. But the next test case does not expect
this commit. For these tests, it's the bottom that's important, not
the top. Fix the expected commit list.
While at it, fix the default http port number to 5537. Otherwise when
t5536 learns to test httpd, running test in parallel may fail.
References:
48d25ca fetch: add --update-shallow to accept... - 2013-12-05
1609488 smart-http: support shallow fetch/clone - 2013-12-05
Noticed-by: Jeff King <peff@peff.net>
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The "Submodules" section of the "git rm" documentation mentions what will
happen when a submodule with a gitfile gets removed with newer git. But it
doesn't talk about what happens when the user changes between commits
before and after the removal, which does not remove the submodule from the
work tree like using the rm command did the first time.
Explain what happens and what the user has to do manually to fix that in
the new BUGS section. Also document this behavior in a new test.
Signed-off-by: Jens Lehmann <Jens.Lehmann@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The "Submodules" section of the "git mv" documentation mentions what will
happen when a submodule with a gitfile gets moved with newer git. But it
doesn't talk about what happens when the user changes between commits
before and after the move, which does not update the work tree like using
the mv command did the first time.
Explain what happens and what the user has to do manually to fix that in
the new BUGS section. Also document this behavior in a new test.
Reported-by: George Papanikolaou <g3orge.app@gmail.com>
Signed-off-by: Jens Lehmann <Jens.Lehmann@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We should always have been freeing our strbuf, but doing so
consistently was annoying until the refactoring in the
previous patch.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This just pulls the return value for the function out of the
inner loop, so we can break out of the loop rather than do
an early return. This will make it easier to put any cleanup
for the function in one place.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When trying to pop/apply a stash specified with an argument
containing IFS whitespace, git-stash will throw an error:
$ git stash pop 'stash@{two hours ago}'
Too many revisions specified: stash@{two hours ago}
This happens because word splitting is used to count non-option
arguments. Make use of rev-parse's --sq option to quote the arguments
for us to ensure a correct count. Add quotes where necessary.
Also add a test that verifies correct behaviour.
Helped-by: Thomas Rast <tr@thomasrast.ch>
Signed-off-by: Øystein Walle <oystwa@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When seeing a full 40-hex object name, get_sha1_basic()
unconditionally checks if the string can also be interpreted as a
refname, but the result will not be used unless warn_ambiguous_refs
is in effect.
Omitting this unnecessary ref resolution provides a substantial
performance improvement, especially when passing many hashes to a
command (like "git rev-list --stdin") and core.warnambiguousrefs is
set to false. The check incurs 6 stat()s for every hash supplied,
which can be costly over NFS.
Signed-off-by: Brodie Rao <brodie@sf.io>
Acked-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Acked-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
On systems with lv configured as the preferred pager (i.e.,
DEFAULT_PAGER=lv at build time, or PAGER=lv exported in the
environment) git commands that use color show control codes instead of
color in the pager:
$ git diff
^[[1mdiff --git a/.mailfilter b/.mailfilter^[[m
^[[1mindex aa4f0b2..17e113e 100644^[[m
^[[1m--- a/.mailfilter^[[m
^[[1m+++ b/.mailfilter^[[m
^[[36m@@ -1,11 +1,58 @@^[[m
"less" avoids this problem because git uses the LESS environment
variable to pass the -R option ('output ANSI color escapes in raw
form') by default. Use the LV environment variable to pass 'lv' the
-c option ('allow ANSI escape sequences for text decoration / color')
to fix it for lv, too.
Noticed when the default value for color.ui flipped to 'auto' in
v1.8.4-rc0~36^2~1 (2013-06-10).
Reported-by: Olaf Meeuwissen <olaf.meeuwissen@avasys.jp>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
'checkout' is documented as one of the valid values for the
'submodule.<name>.update' variable, and in a repository with the
variable set to 'checkout', "git submodule update" command does
update using the 'checkout' mode.
However, it has been an accident that the implementation works this
way; any unknown value would trigger the same codepath and update
using the 'checkout' mode.
Explicitly list 'checkout' as one of the known update modes, and
error out when an unknown update mode is used.
Teach the codepath that initializes the configuration variable from
an in-tree .gitmodules that 'checkout' is one of the valid values.
The code since ac1fbbda (submodule: do not copy unknown update mode
from .gitmodules, 2013-12-02) used to treat the value 'checkout' as
unknown and mapped it to 'none', which made little sense. With this
change, 'checkout' specified in .gitmodules will stay to be 'checkout'.
Signed-off-by: Francesco Pretto <ceztko@gmail.com>
Signed-off-by: Signed-off-by: Junio C Hamano <gitster@pobox.com>
Use asciidoc style 'article' instead of 'book' and change asciidoc
title level. This removes blank first page and superfluous "Part I"
page (there is no "Part II") in pdf output. Also pdf size is
decreased by this from 77 to 67 pages. In html output this removes
unnecessary sub-tocs and chapter numbering.
Signed-off-by: Thomas Ackermann <th.acker@arcor.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Since 2dce956 is_git_command() is a bit slow as it does file I/O in
the call to list_commands_in_dir(). Avoid the file I/O by adding an
early check for the builtin commands.
Signed-off-by: Sebastian Schuberth <sschuberth@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This avoids list_commands_in_dir() being called when not needed which is
quite slow due to file I/O in order to list matching files in a directory.
Signed-off-by: Sebastian Schuberth <sschuberth@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Add a new possible error result that can be returned by
safe_create_leading_directories() and
safe_create_leading_directories_const(): SCLD_VANISHED. This value
indicates that a file or directory on the path existed at one point
(either it already existed or the function created it), but then it
disappeared. This probably indicates that another process deleted the
directory while we were working. If SCLD_VANISHED is returned, the
caller might want to retry the function call, as there is a chance
that a new attempt will succeed.
Why doesn't safe_create_leading_directories() do the retrying
internally? Because an empty directory isn't really ever safe until
it holds a file. So even if safe_create_leading_directories() were
absolutely sure that the directory existed before it returned, there
would be no guarantee that the directory still existed when the caller
tried to write something in it.
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
safe_create_leading_directories_const() returns a non-zero value on
error. The old code at this calling site recognized a couple of
particular error values, and treated all other return values as
success. Instead, be more conservative: recognize the errors we are
interested in, but treat any other nonzero values as failures. This
is more robust in case somebody adds another possible return value
without telling us.
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Instead of returning magic integer values (which a couple of callers
go to the trouble of distinguishing), return values from an enum. Add
a docstring.
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Always restore the slash that we scribbled over at the end of the
loop, rather than also fixing it up at each premature exit from the
loop. This makes it harder to forget to do the cleanup as new paths
are added to the code.
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
If the input path has multiple slashes between path components (e.g.,
"foo//bar"), then the old code was breaking the path at the last
slash, not the first one. So in the above example, the second slash
was overwritten with NUL, resulting in the parent directory being
sought as "foo/".
When stat() is called on "foo/", it fails with ENOTDIR if "foo" exists
but is not a directory. This caused the wrong path to be taken in the
subsequent logic.
So instead, split path components at the first intercomponent slash
rather than the last one.
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Rename "pos" to "next_component", because now it always points at the
next component of the path name that has to be processed.
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
