Junio C Hamano [Mon, 7 May 2012 20:29:16 +0000 (13:29 -0700)]
Merge branch 'nh/empty-rebase'
By Neil Horman
* nh/empty-rebase:
git cherry-pick: do not dereference a potential NULL pointer
Junio C Hamano [Mon, 7 May 2012 20:29:08 +0000 (13:29 -0700)]
Merge branch 'zj/diff-empty-chmod'
"git diff --stat" used to fully count a binary file with modified
execution bits whose contents is unmodified, which was not right.
By Zbigniew Jędrzejewski-Szmek (4) and Johannes Sixt (1)
* zj/diff-empty-chmod:
t4006: Windows do not have /dev/zero
diff --stat: do not run diff on indentical files
diff --stat: report mode-only changes for binary files like text files
tests: check --[short]stat output after chmod
test: modernize style of t4006
Conflicts:
diff.c
Junio C Hamano [Mon, 7 May 2012 20:29:01 +0000 (13:29 -0700)]
Merge branch 'jk/maint-tformat-with-z'
"log -z --pretty=tformat:..." does not terminate each record with NUL
and this is a beginning of an attempt to fix it. It still is not right
but the patch does not make externally observable behaviour worse.
By Jan Krüger (1) and Junio C Hamano (1)
* jk/maint-tformat-with-z:
log-tree: the previous one is still not quite right
log-tree: use custom line terminator in line termination mode
Junio C Hamano [Mon, 7 May 2012 20:28:44 +0000 (13:28 -0700)]
Merge branch 'th/difftool-diffall'
Rolls the two-directory-diff logic from diffall script (in contrib/) into
"git difftool" framework.
By Tim Henigan
* th/difftool-diffall:
difftool: print list of valid tools with '--tool-help'
difftool: teach difftool to handle directory diffs
difftool: eliminate setup_environment function
difftool: stop appending '.exe' to git
difftool: remove explicit change of PATH
difftool: exit(0) when usage is printed
difftool: add '--no-gui' option
difftool: parse options using Getopt::Long
Junio C Hamano [Mon, 7 May 2012 20:28:34 +0000 (13:28 -0700)]
Sync with maint
Junio C Hamano [Mon, 7 May 2012 20:24:13 +0000 (13:24 -0700)]
Start preparing for 1.7.10.2
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Junio C Hamano [Mon, 7 May 2012 20:17:31 +0000 (13:17 -0700)]
Merge branch 'jk/maint-gitweb-test-use-sane-perl' into maint
When using a Perl script on a system where "perl" found on user's $PATH
could be ancient or otherwise broken, we allow builders to specify the
path to a good copy of Perl with $PERL_PATH. The gitweb test forgot to
use that Perl when running its test.
By Jeff King (1) and Zbigniew Jędrzejewski-Szmek (1)
* jk/maint-gitweb-test-use-sane-perl:
Consistently use perl from /usr/bin/ for scripts
t/gitweb-lib: use $PERL_PATH to run gitweb
Junio C Hamano [Mon, 7 May 2012 20:17:17 +0000 (13:17 -0700)]
Merge branch 'js/daemon-test-race-fix' into maint
The test scaffolding for git-daemon was flaky.
By Johannes Sixt
* js/daemon-test-race-fix:
t5570: fix forwarding of git-daemon messages via cat
Junio C Hamano [Mon, 7 May 2012 20:14:08 +0000 (13:14 -0700)]
Merge branch 'jk/maint-config-bogus-section' into maint
"git config --rename-section" to rename an existing section into a bogus
one did not check the new name.
By Jeff King
* jk/maint-config-bogus-section:
config: reject bogus section names for --rename-section
Junio C Hamano [Mon, 7 May 2012 20:13:55 +0000 (13:13 -0700)]
Merge branch 'pw/t5800-import-race-fix' into maint
The test scaffolding for fast-import was flaky.
By Pete Wyckoff
* pw/t5800-import-race-fix:
git-remote-testgit: fix race when spawning fast-import
Junio C Hamano [Mon, 7 May 2012 20:13:43 +0000 (13:13 -0700)]
Merge branch 'rt/cherry-revert-conflict-summary' into maint
In the older days, the header "Conflicts:" in "cherry-pick" and "merge"
was separated by a blank line from the list of paths that follow for
readability, but when "merge" was rewritten in C, we lost it by
mistake. Remove the newline from "cherry-pick" to make them match again.
By Ralf Thielow
* rt/cherry-revert-conflict-summary:
sequencer: remove additional blank line
Junio C Hamano [Mon, 7 May 2012 20:13:03 +0000 (13:13 -0700)]
Merge branch 'cb/maint-report-mount-point-correctly-in-setup' into maint
The filesystem boundary was not correctly reported when .git directory
discovery stopped at a mount point.
By Clemens Buchacher
* cb/maint-report-mount-point-correctly-in-setup:
properly keep track of current working directory
Junio C Hamano [Mon, 7 May 2012 19:35:36 +0000 (12:35 -0700)]
Merge branch 'ef/maint-1.7.6-clone-progress-fix' into ef/maint-clone-progress-fix
By Erik Faye-Lund
* ef/maint-1.7.6-clone-progress-fix:
clone: fix progress-regression
Erik Faye-Lund [Mon, 7 May 2012 19:23:13 +0000 (21:23 +0200)]
clone: fix progress-regression
In
5bd631b3 ("clone: support multiple levels of verbosity"), the
default behavior to show progress of the implicit checkout in
the clone-command regressed so that progress was only shown if
the verbose-option was specified.
Fix this by making option_verbosity == 0 output progress as well.
Signed-off-by: Erik Faye-Lund <kusmabite@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Angus Hammond [Sun, 6 May 2012 18:17:15 +0000 (19:17 +0100)]
grep.c: remove redundant line of code
Signed-off-by: Angus Hammond <angusgh@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Pete Wyckoff [Sun, 6 May 2012 15:58:38 +0000 (11:58 -0400)]
git p4 doc: fix formatting
Attach example sections to previous level of indenting.
Fix a trailing ::
Signed-off-by: Pete Wyckoff <pw@padd.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Felipe Contreras [Tue, 1 May 2012 18:42:56 +0000 (20:42 +0200)]
completion: simplify __git_complete_revlist_file
Use new __gitcomp_nl; this is the last place that uses COMPREPLY and
compgen directly outside __gitcomp* functions.
Signed-off-by: Felipe Contreras <felipe.contreras@gmail.com>
Acked-by: SZEDER Gábor <szeder@ira.uka.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Johannes Sixt [Fri, 4 May 2012 18:14:48 +0000 (20:14 +0200)]
checkout (detached): truncate list of orphaned commits at the new HEAD
When git checkout switches from a detached HEAD to any other commit, then
all orphaned commits were listed in a warning:
Warning: you are leaving 2 commits behind...:
a5e5396 another fixup
6aa1af6 fixup foo
But if the new commit is actually one from this list (
6aa1af6 in this
example), then the list in the warning can be truncated at the new HEAD,
because history beginning at HEAD is not "left behind". This makes it so.
Signed-off-by: Johannes Sixt <j6t@kdbg.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Johannes Sixt [Fri, 4 May 2012 18:13:14 +0000 (20:13 +0200)]
t2020-checkout-detach: check for the number of orphaned commits
Change the test that orphans commits to leave 2 commits behind. Add a test
that leaves only one of these behind.
The next patch will truncate the list of orphaned commits earlier. With
this preliminary update, its effect will become more obvious.
Signed-off-by: Johannes Sixt <j6t@kdbg.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Junio C Hamano [Wed, 2 May 2012 22:12:10 +0000 (15:12 -0700)]
Makefile: NO_INSTALL_HARDLINKS
Your filesystem may support hardlinks, but you may choose not to use them
when installing git-foo builtins and favor symblic links or copies for
whatever reason.
The installation procedure of git-gui/ directory is not touched with this
patch and git-citool still ends up being a hardlink to git-gui, but it
needs to be addressed separately.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Jeff King [Fri, 4 May 2012 05:27:25 +0000 (01:27 -0400)]
reflog-walk: always make HEAD@{0} show indexed selectors
When we are showing reflog selectors during a walk, we infer
from context whether the user wanted to see the index in
each selector, or the reflog date. The current rules are:
1. if the user asked for an explicit date format in the
output, show the date
2. if the user asked for ref@{now}, show the date
3. if neither is true, show the index
However, if we see "ref@{0}", that should be a strong clue
that the user wants to see the counted version. In fact, it
should be much stronger than the date format in (1). The
user may have been setting the date format to use in another
part of the output (e.g., in --format="%gd (%ad)", they may
have wanted to influence the author date).
This patch flips the rules to:
1. if the user asked for ref@{0}, always show the index
2. if the user asked for ref@{now}, always show the date
3. otherwise, we have just "ref"; show them counted by
default, but respect the presence of "--date" as a clue
that the user wanted them date-based
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Jeff King [Fri, 4 May 2012 05:26:26 +0000 (01:26 -0400)]
reflog-walk: clean up "flag" field of commit_reflog struct
When we prepare to walk a reflog, we parse the specification
and pull some information from it, such as which reflog to
look in (e.g., HEAD), and where to start (e.g., HEAD@{10} or
HEAD@{yesterday}). The resulting struct has a "recno" field
to show where in the reflog we are starting. It also has a
"flag" field; if true, it means the recno field came from
parsing a date like HEAD@{yesterday}.
There are two problems with this:
1. "flag" is an absolutely terrible name, as it conveys
nothing about the meaning
2. you can tell "HEAD" from "HEAD@{yesterday}", but you
can't differentiate "HEAD" from "HEAD{0}"
This patch converts the flag into a tri-state (and gives it
a better name!).
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Jeff King [Fri, 4 May 2012 05:25:18 +0000 (01:25 -0400)]
log: respect date_mode_explicit with --format:%gd
When we show a reflog selector (e.g., via "git log -g"), we
perform some DWIM magic: while we normally show the entry's
index (e.g., HEAD@{1}), if the user has given us a date
with "--date", then we show a date-based select (e.g.,
HEAD@{yesterday}).
However, we don't want to trigger this magic if the
alternate date format we got was from the "log.date"
configuration; that is not sufficiently strong context for
us to invoke this particular magic. To fix this, commit
f4ea32f (improve reflog date/number heuristic, 2009-09-24)
introduced a "date_mode_explicit" flag in rev_info. This
flag is set only when we see a "--date" option on the
command line, and we a vanilla date to the reflog code if
the date was not explicit.
Later, commit
8f8f547 (Introduce new pretty formats %g[sdD]
for reflog information, 2009-10-19) added another way to
show selectors, and it did not respect the date_mode_explicit
flag from
f4ea32f.
This patch propagates the date_mode_explicit flag to the
pretty-print code, which can then use it to pass the
appropriate date field to the reflog code. This brings the
behavior of "%gd" in line with the other formats, and means
that its output is independent of any user configuration.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Jeff King [Fri, 4 May 2012 05:23:14 +0000 (01:23 -0400)]
t1411: add more selector index/date tests
We already check that @{now} and "--date" cause the
displayed selector to use the date for both the multiline
and oneline formats. However, we miss several cases:
1. The --format=%gd selector is not tested at all.
2. We do not check how the log.date config interacts with the
"--date" magic (according to
f4ea32f, it should not
impact the output).
Doing so reveals that the combination of both (log.date
combined with the %gd format) does not behave as expected.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Nguyễn Thái Ngọc Duy [Fri, 4 May 2012 13:52:36 +0000 (20:52 +0700)]
Makefile: keep many variable list sorted
We tend to keep long lists sorted (extensions are not taken into
account), which helps spot a name easily by eye. Rearrange a few
items so these lists remain sorted.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Junio C Hamano [Thu, 3 May 2012 22:12:54 +0000 (15:12 -0700)]
refs: fix find_containing_dir() regression
The function used to return NULL when asked to find the containing
directory for a ref that does not exist, allowing the caller to
omit iteration altogether. But a misconversion in an earlier change
"refs.c: extract function search_for_subdir()" started returning the
top-level directory entry, forcing callers to walk everything.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Nguyễn Thái Ngọc Duy [Thu, 3 May 2012 13:12:00 +0000 (20:12 +0700)]
branch: remove lego in i18n tracking info strings
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Neil Horman [Thu, 3 May 2012 12:10:22 +0000 (08:10 -0400)]
git cherry-pick: do not dereference a potential NULL pointer
In the case the pointer could be NULL, the function that gave the caller
the NULL pointer would already have issued an error message, so simply
returning early with an error status without issuing a new message is
sufficient. The same for parse_commit() that will show necessary error
message when the argument is not NULL, and will return error silently
when the argument is NULL.
Noticed-by: Michael Mueller
Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Junio C Hamano [Thu, 3 May 2012 22:37:28 +0000 (15:37 -0700)]
Merge branch 'maint'
* maint:
Junio C Hamano [Thu, 3 May 2012 22:36:15 +0000 (15:36 -0700)]
Merge branch 'jc/merge-reduce-parents-early' into maint
Octopus merge strategy did not reduce heads that are recorded in the final
commit correctly.
By Junio C Hamano (4) and Michał Kiedrowicz (1)
* jc/merge-reduce-parents-early:
fmt-merge-msg: discard needless merge parents
builtin/merge.c: reduce parents early
builtin/merge.c: collect other parents early
builtin/merge.c: remove "remoteheads" global variable
merge tests: octopus with redundant parents
Junio C Hamano [Thu, 3 May 2012 22:35:19 +0000 (15:35 -0700)]
Merge branch 'cb/cherry-pick-rev-path-confusion' into maint
The command line parser choked "git cherry-pick $name" when $name can be
both revision name and a pathname, even though $name can never be a path
in the context of the command.
By Clemens Buchacher
* cb/cherry-pick-rev-path-confusion:
cherry-pick: do not expect file arguments
Junio C Hamano [Thu, 3 May 2012 22:34:51 +0000 (15:34 -0700)]
Merge branch 'cb/http-multi-curl-auth' into maint
HTTP transport that requires authentication did not work correctly when
multiple connections are used simultaneously.
By Jeff King (3) and Clemens Buchacher (1)
* cb/http-multi-curl-auth:
http: use newer curl options for setting credentials
http: clean up leak in init_curl_http_auth
fix http auth with multiple curl handles
http auth fails with multiple curl handles
Junio C Hamano [Thu, 3 May 2012 22:29:25 +0000 (15:29 -0700)]
Merge branch 'mb/fetch-call-a-non-branch-a-ref' into maint
The report from "git fetch" said "new branch" even for a non branch ref.
By Marc Branchaud
* mb/fetch-call-a-non-branch-a-ref:
fetch: describe new refs based on where it came from
fetch: Give remote_ref to update_local_ref() as well
Junio C Hamano [Thu, 3 May 2012 22:24:22 +0000 (15:24 -0700)]
Merge branch 'rl/maint-stash-i18n-save-error' into maint
i18n marking for an error message for "git stash --notavalidoption"
was incorrect.
By Ross Lagerwall
* rl/maint-stash-i18n-save-error:
stash: use eval_gettextln correctly
Junio C Hamano [Thu, 3 May 2012 22:23:17 +0000 (15:23 -0700)]
Merge branch 'jm/maint-strncpy-diff-no-index' into maint
"diff --no-index" codepath had a few places that used fixed-size
buffer and truncated paths that are too long.
By Jim Meyering (1) and Junio C Hamano (1)
* jm/maint-strncpy-diff-no-index:
diff --no-index: use strbuf for temporary pathnames
diff: avoid stack-buffer-read-overrun for very long name
Junio C Hamano [Thu, 3 May 2012 22:13:55 +0000 (15:13 -0700)]
Merge branch 'jk/maint-push-progress'
"git push" over smart-http lost progress output and this resurrects it.
By Jeff King
* jk/maint-push-progress:
t5541: test more combinations of --progress
teach send-pack about --[no-]progress
send-pack: show progress when isatty(2)
Junio C Hamano [Thu, 3 May 2012 22:13:31 +0000 (15:13 -0700)]
Merge branch 'nd/columns'
A couple of commands learn --column option to produce columnar output.
By Nguyễn Thái Ngọc Duy (9) and Zbigniew Jędrzejewski-Szmek (1)
* nd/columns:
tag: add --column
column: support piping stdout to external git-column process
status: add --column
branch: add --column
help: reuse print_columns() for help -a
column: add dense layout support
t9002: work around shells that are unable to set COLUMNS to 1
column: add columnar layout
Stop starting pager recursively
Add column layout skeleton and git-column
Michael Haggerty [Thu, 26 Apr 2012 22:27:07 +0000 (00:27 +0200)]
refs: read loose references lazily
Instead of reading the whole directory of loose references the first
time any are needed, only read them on demand, one directory at a
time.
Use a new ref_entry flag bit REF_INCOMPLETE to indicate that the entry
represents a REF_DIR that hasn't been read yet. Whenever any entries
from such a directory are needed, read all of the loose references
from that directory.
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Michael Haggerty [Thu, 26 Apr 2012 22:27:06 +0000 (00:27 +0200)]
read_loose_refs(): eliminate ref_cache argument
The ref_cache can now be read from the ref_dir.
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Michael Haggerty [Thu, 26 Apr 2012 22:27:05 +0000 (00:27 +0200)]
struct ref_dir: store a reference to the enclosing ref_cache
This means that a directory ref_entry contains all of the information
needed by read_loose_refs().
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Michael Haggerty [Thu, 26 Apr 2012 22:27:04 +0000 (00:27 +0200)]
search_for_subdir(): return (ref_dir *) instead of (ref_entry *)
That is what all the callers want, so give it to them.
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Michael Haggerty [Thu, 26 Apr 2012 22:27:03 +0000 (00:27 +0200)]
get_ref_dir(): add function for getting a ref_dir from a ref_entry
Convert all accesses of a ref_dir within a ref_entry to use this
function. This function will later be responsible for reading loose
references from disk on demand.
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Michael Haggerty [Thu, 26 Apr 2012 22:27:02 +0000 (00:27 +0200)]
read_loose_refs(): rename function from get_ref_dir()
The new name better describes the function's purpose, and also makes
the old name available for a more suitable purpose.
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Michael Haggerty [Thu, 26 Apr 2012 22:27:01 +0000 (00:27 +0200)]
refs: wrap top-level ref_dirs in ref_entries
Make it turtles all the way down. This affects the loose and packed
fields of ref_cache instances.
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Michael Haggerty [Thu, 26 Apr 2012 22:27:00 +0000 (00:27 +0200)]
find_containing_dir(): use strbuf in implementation of this function
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Michael Haggerty [Thu, 26 Apr 2012 22:26:59 +0000 (00:26 +0200)]
bisect: copy filename string obtained from git_path()
Prevent the string from being overwritten by other callers of
git_path() and friends before we are done using it.
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Michael Haggerty [Tue, 24 Apr 2012 22:45:14 +0000 (00:45 +0200)]
do_for_each_reflog(): use a strbuf to hold logfile name
This simplifies the bookkeeping and allows an (artificial) restriction
on refname component length to be removed.
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Michael Haggerty [Tue, 24 Apr 2012 22:45:13 +0000 (00:45 +0200)]
do_for_each_reflog(): return early on error
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Michael Haggerty [Tue, 24 Apr 2012 22:45:12 +0000 (00:45 +0200)]
get_ref_dir(): take the containing directory as argument
Previously, the "dir" argument to get_ref_dir() was a pointer to the
top-level ref_dir. Change the function to expect a pointer to the
ref_dir corresponding to dirname. This allows entries to be added
directly to dir, without having to recurse through the reference trie
each time (i.e., we can use add_entry_to_dir() instead of add_ref()).
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Michael Haggerty [Tue, 24 Apr 2012 22:45:11 +0000 (00:45 +0200)]
refs.c: extract function search_for_subdir()
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Michael Haggerty [Tue, 24 Apr 2012 22:45:10 +0000 (00:45 +0200)]
get_ref_dir(): require that the dirname argument ends in '/'
This removes some conditional code and makes it consistent with the
way that direntry names are stored. Please note that this function is
never used on the top-level .git directory; it is always called for
directories at level .git/refs or deeper.
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Michael Haggerty [Tue, 24 Apr 2012 22:45:09 +0000 (00:45 +0200)]
get_ref_dir(): rename "base" parameter to "dirname"
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Michael Haggerty [Tue, 24 Apr 2012 22:45:08 +0000 (00:45 +0200)]
get_ref_dir(): use a strbuf to hold refname
This simplifies the bookkeeping and allows an (artificial) restriction
on refname component length to be removed.
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
René Scharfe [Thu, 3 May 2012 08:52:16 +0000 (10:52 +0200)]
t5000: rationalize unzip tests
Factor out a function for checking the contents of ZIP archives. It
extracts their contents and compares them to the original files. This
removes some duplicate code. Tests that just create archives can lose
their UNZIP prerequisite.
Signed-off-by: Rene Scharfe <rene.scharfe@lsrfire.ath.cx>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
René Scharfe [Thu, 3 May 2012 01:51:08 +0000 (08:51 +0700)]
archive-zip: streaming for deflated files
After an entry has been streamed out, its CRC and sizes are written as
part of a data descriptor.
For simplicity, we make the buffer for the compressed chunks twice as
big as for the uncompressed ones, to be sure the result fit in even
if deflate makes them bigger.
t5000 verifies output. t1050 makes sure the command always respects
core.bigfilethreshold
Signed-off-by: Rene Scharfe <rene.scharfe@lsrfire.ath.cx>
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
René Scharfe [Thu, 3 May 2012 01:51:07 +0000 (08:51 +0700)]
archive-zip: streaming for stored files
Write a data descriptor containing the CRC of the entry and its sizes
after streaming it out. For simplicity, do that only if we're storing
files (option -0) for now.
t5000 verifies output. t1050 makes sure the command always respects
core.bigfilethreshold
Signed-off-by: Rene Scharfe <rene.scharfe@lsrfire.ath.cx>
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
René Scharfe [Thu, 3 May 2012 01:51:06 +0000 (08:51 +0700)]
archive-zip: factor out helpers for writing sizes and CRC
We're going to reuse them soon for streaming. Also, update the ZIP
directory only at the very end, which will also make streaming easier.
Signed-off-by: Rene Scharfe <rene.scharfe@lsrfire.ath.cx>
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
René Scharfe [Thu, 3 May 2012 01:51:05 +0000 (08:51 +0700)]
archive-zip: remove uncompressed_size
We only need size and compressed_size.
Signed-off-by: Rene Scharfe <rene.scharfe@lsrfire.ath.cx>
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Nguyễn Thái Ngọc Duy [Thu, 3 May 2012 01:51:04 +0000 (08:51 +0700)]
archive-tar: stream large blobs to tar file
t5000 verifies output while t1050 makes sure the command always
respects core.bigfilethreshold
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Nguyễn Thái Ngọc Duy [Thu, 3 May 2012 01:51:03 +0000 (08:51 +0700)]
archive: delegate blob reading to backend
archive-tar.c and archive-zip.c now perform conversion check, with
help of sha1_file_to_archive() from archive.c
This gives backends more freedom in dealing with (streaming) large
blobs.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Nguyễn Thái Ngọc Duy [Thu, 3 May 2012 01:51:02 +0000 (08:51 +0700)]
archive-tar: unindent write_tar_entry by one level
It's used to be
if (!sha1) {
...
} else if (!path) {
...
} else {
...
}
Now that the first two blocks are no-op. We can remove the if/else
skeleton and put the else block back by one indent level.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Nguyễn Thái Ngọc Duy [Thu, 3 May 2012 01:51:01 +0000 (08:51 +0700)]
archive-tar: turn write_tar_entry into blob-writing only
Before this patch write_tar_entry() can:
- write global header
by write_global_extended_header() calling write_tar_entry with
with both sha1 and path == NULL
- write extended header for symlinks, by write_tar_entry() calling
itself with sha1 != NULL and path == NULL
- write a normal blob. In this case both sha1 and path are valid.
After this patch, the first two call sites are modified to write the
header without calling write_tar_entry(). The function is now for
writing blobs only. This simplifies handling when write_tar_entry()
learns about large blobs.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
René Scharfe [Thu, 3 May 2012 01:51:00 +0000 (08:51 +0700)]
streaming: void pointer instead of char pointer
Allow any kind of buffer to be fed to read_istream() without an explicit
cast by making it's buf argument a void pointer. It's about arbitrary
data, not only characters.
Signed-off-by: Rene Scharfe <rene.scharfe@lsrfire.ath.cx>
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Jeff King [Wed, 2 May 2012 23:43:52 +0000 (19:43 -0400)]
doc/config: fix inline literals
Since commit
6cf378f, asciidoc backticks are now inline
literals; therefore quoting {tilde} inside them is wrong
(this instance was missed in
6cf378f because it happened on
a parallel line of development).
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Junio C Hamano [Wed, 2 May 2012 21:11:36 +0000 (14:11 -0700)]
The eighth batch of topics graduated to 'master'
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Junio C Hamano [Wed, 2 May 2012 21:38:29 +0000 (14:38 -0700)]
Merge branch 'master' of git://git.bogomips.org/git-svn
via Eric Wong
* 'master' of git://git.bogomips.org/git-svn:
git-svn: introduce SVN version comparison function
Junio C Hamano [Wed, 2 May 2012 20:54:58 +0000 (13:54 -0700)]
Merge branch 'tr/xdiff-fast-hash'
Use word-at-a-time comparison to find end of line or NUL (end of buffer),
borrowed from the linux-kernel discussion.
By Thomas Rast
* tr/xdiff-fast-hash:
xdiff: choose XDL_FAST_HASH code on sizeof(long) instead of __WORDSIZE
xdiff: load full words in the inner loop of xdl_hash_record
Junio C Hamano [Wed, 2 May 2012 20:53:40 +0000 (13:53 -0700)]
Merge branch 'jk/maint-gitweb-test-use-sane-perl'
When using a Perl script on a system where "perl" found on user's $PATH
could be ancient or otherwise broken, we allow builders to specify the
path to a good copy of Perl with $PERL_PATH. The gitweb test forgot to
use that Perl when running its test.
By Jeff King (1) and Zbigniew Jędrzejewski-Szmek (1)
* jk/maint-gitweb-test-use-sane-perl:
Consistently use perl from /usr/bin/ for scripts
t/gitweb-lib: use $PERL_PATH to run gitweb
Junio C Hamano [Wed, 2 May 2012 20:53:35 +0000 (13:53 -0700)]
Merge branch 'pw/message-cleanup'
Many error/warning messages had extra trailing newlines that are
unnecessary.
By Pete Wyckoff
* pw/message-cleanup:
remove blank filename in error message
remove superfluous newlines in error messages
Junio C Hamano [Wed, 2 May 2012 20:53:28 +0000 (13:53 -0700)]
Merge branch 'zj/diff-stat-smaller-num-columns'
Spend only minimum number of columns necessary to show the number of lines
in the output from "diff --stat", instead of always allocating 4 columns
even when showing changes that are much smaller than 1000 lines.
By Zbigniew Jędrzejewski-Szmek
* zj/diff-stat-smaller-num-columns:
diff --stat: use less columns for change counts
Junio C Hamano [Wed, 2 May 2012 20:53:23 +0000 (13:53 -0700)]
Merge branch 'pw/p4-various'
Miscellaneous updates to "git p4".
By Pete Wyckoff
* pw/p4-various:
git p4: submit files with wildcards
git p4: fix writable file after rename or copy
git p4: test submit
git p4: bring back files in deleted client directory
Junio C Hamano [Wed, 2 May 2012 20:53:15 +0000 (13:53 -0700)]
Merge branch 'jc/rerere-train'
A script written long time ago proved to be useful this week for me ;-)
with a minor tweak.
* jc/rerere-train:
contrib/rerere-train: use installed git-sh-setup
Junio C Hamano [Wed, 2 May 2012 20:53:02 +0000 (13:53 -0700)]
Merge branch 'ab/i18n'
Fix some constructs that build messages meant for i18n by concatenating
pieces of strings.
By Ævar Arnfjörð Bjarmason
* ab/i18n:
git-commit: remove lego in i18n messages
git-commit: remove lego in i18n messages
git-branch: remove lego in i18n messages
Junio C Hamano [Wed, 2 May 2012 20:51:59 +0000 (13:51 -0700)]
Merge branch 'lp/diffstat-with-graph'
"log --graph" was not very friendly with "--stat" option and its output
had line breaks at wrong places.
By Lucian Poston (5) and Zbigniew Jędrzejewski-Szmek (2)
* lp/diffstat-with-graph:
t4052: work around shells unable to set COLUMNS to 1
Prevent graph_width of stat width from falling below min
t4052: Test diff-stat output with minimum columns
t4052: Adjust --graph --stat output for prefixes
Adjust stat width calculations to take --graph output into account
Add output_prefix_length to diff_options
t4052: test --stat output with --graph
Junio C Hamano [Wed, 2 May 2012 20:51:53 +0000 (13:51 -0700)]
Merge branch 'zj/mksh-columns-breakage'
A broken shell may not let us set an environment value to an arbitrary
value, interfering with some of the tests. Introduce a test prerequisite
so that we can skip some tests on such a platform.
By Zbigniew Jędrzejewski-Szmek
* zj/mksh-columns-breakage:
test-lib: skip test with COLUMNS=1 under mksh
Junio C Hamano [Wed, 2 May 2012 20:51:45 +0000 (13:51 -0700)]
Merge branch 'jk/doc-asciidoc-inline-literal'
Our documentation was written for an ancient version of AsciiDoc,
making the source not very readable.
By Jeff King
* jk/doc-asciidoc-inline-literal:
docs: stop using asciidoc no-inline-literal
Junio C Hamano [Wed, 2 May 2012 20:51:35 +0000 (13:51 -0700)]
Merge branch 'nd/i18n'
More message strings marked for i18n.
By Nguyễn Thái Ngọc Duy (10) and Jonathan Nieder (1)
* nd/i18n:
help: replace underlining "help -a" headers using hyphens with a blank line
i18n: bundle: mark strings for translation
i18n: index-pack: mark strings for translation
i18n: apply: update say_patch_name to give translators complete sentence
i18n: apply: mark strings for translation
i18n: remote: mark strings for translation
i18n: make warn_dangling_symref() automatically append \n
i18n: help: mark strings for translation
i18n: mark relative dates for translation
strbuf: convenience format functions with \n automatically appended
Makefile: feed all header files to xgettext
Junio C Hamano [Wed, 2 May 2012 20:51:24 +0000 (13:51 -0700)]
Merge branch 'mm/simple-push'
New users tend to work on one branch at a time and push the result
out. The current and upstream modes of push is a more suitable default
mode than matching mode for these people, but neither is surprise-free
depending on how the project is set up. Introduce a "simple" mode that
is a subset of "upstream" but only works when the branch is named the same
between the remote and local repositories.
The plan is to make it the new default when push.default is not
configured.
By Matthieu Moy (5) and others
* mm/simple-push:
push.default doc: explain simple after upstream
push: document the future default change for push.default (matching -> simple)
t5570: use explicit push refspec
push: introduce new push.default mode "simple"
t5528-push-default.sh: add helper functions
Undocument deprecated alias 'push.default=tracking'
Documentation: explain push.default option a bit more
Junio C Hamano [Wed, 2 May 2012 20:51:13 +0000 (13:51 -0700)]
Merge branch 'jc/index-v4'
Trivially shrinks the on-disk size of the index file to save both I/O and
checksum overhead.
The topic should give a solid base to build on further updates, with the
code refactoring in its earlier parts, and the backward compatibility
mechanism in its later parts.
* jc/index-v4:
index-v4: document the entry format
unpack-trees: preserve the index file version of original
update-index: upgrade/downgrade on-disk index version
read-cache.c: write prefix-compressed names in the index
read-cache.c: read prefix-compressed names in index on-disk version v4
read-cache.c: move code to copy incore to ondisk cache to a helper function
read-cache.c: move code to copy ondisk to incore cache to a helper function
read-cache.c: report the header version we do not understand
read-cache.c: make create_from_disk() report number of bytes it consumed
read-cache.c: allow unaligned mapping of the index file
cache.h: hide on-disk index details
varint: make it available outside the context of pack
Junio C Hamano [Wed, 2 May 2012 19:53:50 +0000 (19:53 +0000)]
git-svn: introduce SVN version comparison function
With double-digit version components in SVN::Core::VERSION,
a naive string comparison is no longer sufficient and
a numeric comparison is required for correct results.
This fixes a regression introduced in
commit
082afee621aeb2d3746c8ae290af98823f981f34
("git-svn: use platform specific auth providers")
where SVN version "1.6.6" was incorrectly assumed to
be newer than the required "1.6.12" version.
[mk: fix namespace references]
[ew: commit message]
Tested-by: Matthijs Kooijman <matthijs@stdin.nl>
Acked-by: Eric Wong <normalperson@yhbt.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Johannes Sixt [Wed, 2 May 2012 07:36:44 +0000 (09:36 +0200)]
t4006: Windows do not have /dev/zero
We only need to have a file with _some_ binary contents; be nice to
our Windows friends and avoid using /dev/zero
Signed-off-by: Johannes Sixt <j.sixt@viscovery.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Zbigniew Jędrzejewski-Szmek [Tue, 1 May 2012 17:10:15 +0000 (19:10 +0200)]
diff --stat: do not run diff on indentical files
If two objects are known to be equal, there is no point running the diff.
Signed-off-by: Zbigniew Jędrzejewski-Szmek <zbyszek@in.waw.pl>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Zbigniew Jędrzejewski-Szmek [Tue, 1 May 2012 17:10:14 +0000 (19:10 +0200)]
diff --stat: report mode-only changes for binary files like text files
Mode-only changes to binary files without content change were reported as
if they were rewritten, but text files in the same situation were reported
as "unchanged". Let's treat binary files like text files here, and simply
say that they are unchanged.
Output of --shortstat is modified in the same way.
Reported-by: Martin Mareš <mj@ucw.cz>
Signed-off-by: Zbigniew Jędrzejewski-Szmek <zbyszek@in.waw.pl>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Junio C Hamano [Wed, 2 May 2012 04:22:35 +0000 (21:22 -0700)]
Merge branch 'master' of https://github.com/git-l10n/git-po
By Jiang Xin (2) and Ralf Thielow (1)
via Jiang Xin
* 'master' of https://github.com/git-l10n/git-po:
l10n: Update German translation
l10n: Update Simplified Chinese translation
l10n: Update git.pot (33 new, 24 deleted messages)
Junio C Hamano [Wed, 2 May 2012 04:20:46 +0000 (21:20 -0700)]
Sync with v1.7.10.1
Junio C Hamano [Wed, 2 May 2012 04:18:44 +0000 (21:18 -0700)]
Git 1.7.10.1
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Junio C Hamano [Wed, 2 May 2012 04:12:46 +0000 (21:12 -0700)]
Merge branch 'pw/fast-import-dataref-parsing' into maint
The parser in "fast-import" did not diagnose ":9" style references that is
not followed by required SP/LF as an error.
By Pete Wyckoff
* pw/fast-import-dataref-parsing:
fast-import: tighten parsing of datarefs
Junio C Hamano [Wed, 2 May 2012 04:12:36 +0000 (21:12 -0700)]
Merge branch 'it/fetch-pack-many-refs' into maint
When "git fetch" encounters repositories with too many references, the
command line of "fetch-pack" that is run by a helper e.g. remote-curl, may
fail to hold all of them. Now such an internal invocation can feed the
references through the standard input of "fetch-pack".
By Ivan Todoroski
* it/fetch-pack-many-refs:
remote-curl: main test case for the OS command line overflow
fetch-pack: test cases for the new --stdin option
remote-curl: send the refs to fetch-pack on stdin
fetch-pack: new --stdin option to read refs from stdin
Conflicts:
t/t5500-fetch-pack.sh
Junio C Hamano [Wed, 2 May 2012 04:12:25 +0000 (21:12 -0700)]
Merge branch 'jl/maint-submodule-recurse-fetch' into maint
"git fetch" that recurses into submodules on demand did not check if it
needs to go into submodules when non branches (most notably, tags) are
fetched.
By Jens Lehmann
* jl/maint-submodule-recurse-fetch:
submodules: recursive fetch also checks new tags for submodule commits
Junio C Hamano [Wed, 2 May 2012 04:11:49 +0000 (21:11 -0700)]
Merge branch 'jc/maint-blame-minimal' into maint
"git blame" started missing quite a few changes from the origin since we
stopped using the diff minimalization by default in v1.7.2 era.
Teach "--minimal" option to "git blame" to work around this regression.
* jc/maint-blame-minimal:
blame: accept --need-minimal
Junio C Hamano [Wed, 2 May 2012 04:11:40 +0000 (21:11 -0700)]
Merge branch 'lp/maint-diff-three-dash-with-graph' into maint
"log -p --graph" used with "--stat" had a few formatting error.
By Lucian Poston
* lp/maint-diff-three-dash-with-graph:
t4202: add test for "log --graph --stat -p" separator lines
log --graph: fix break in graph lines
log --graph --stat: three-dash separator should come after graph lines
Junio C Hamano [Wed, 2 May 2012 04:11:26 +0000 (21:11 -0700)]
Merge branch 'jk/rebase-i-submodule-conflict-only' into maint
Giving "--continue" to a conflicted "rebase -i" session skipped a
commit that only results in changes to submodules.
By John Keeping
* jk/rebase-i-submodule-conflict-only:
rebase -i continue: don't skip commits that only change submodules
Junio C Hamano [Wed, 2 May 2012 04:09:46 +0000 (21:09 -0700)]
Merge branch 'maint' of https://github.com/git-l10n/git-po
By Byrial Jensen (2) and others
via Jiang Xin (1) and Ralf Thielow (1)
* 'maint' of https://github.com/git-l10n/git-po:
l10n: Initial German translation
l10n: Update Simplified Chinese translation
l10n: Update git.pot (2 new messages)
l10n: Add the German translation team and initialize de.po
l10n: Add Danish team (da) to list of teams
l10n: New da.po file with 0 translations
l10n: Updated pt_PT language
Ralf Thielow [Sat, 28 Apr 2012 08:28:03 +0000 (10:28 +0200)]
l10n: Update German translation
Translate new messages for git master branch.
Signed-off-by: Ralf Thielow <ralf.thielow@googlemail.com>
Jiang Xin [Tue, 1 May 2012 23:30:14 +0000 (07:30 +0800)]
Merge l10n updates from branch 'maint' into master
By Ralf Thielow
* maint:
l10n: Initial German translation
Zbigniew Jędrzejewski-Szmek [Tue, 1 May 2012 20:18:18 +0000 (22:18 +0200)]
Consistently use perl from /usr/bin/ for scripts
While the majority of scripts use '#!/usr/bin/perl', some use
'#!/usr/bin/env perl'. In the end there is no difference, because the
Makefile rewrites "#!.*perl" with "#!$PERL_PATH" in scripted
Porcelains before installing. Nevertheless, the second form can be
misleading, because it suggests that perl found first in $PATH will be
used.
Suggested-by: Junio C Hamano <gitster@pobox.com>
Signed-off-by: Zbigniew Jędrzejewski-Szmek <zbyszek@in.waw.pl>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Thomas Rast [Tue, 1 May 2012 10:23:20 +0000 (12:23 +0200)]
xdiff: choose XDL_FAST_HASH code on sizeof(long) instead of __WORDSIZE
Darwin does not define __WORDSIZE, and compiles the 32-bit code path
on 64-bit systems, resulting in a totally broken git.
I could not find an alternative -- other than the platform symbols
(__x86_64__ etc.) -- that does the test in the preprocessor. However,
we can also just test for the size of a 'long', which is what really
matters here. Any compiler worth its salt will leave only the branch
relevant for its platform, and indeed on Linux/GCC the numbers don't
change:
Test tr/darwin-xdl-fast-hash origin/next origin/master
------------------------------------------------------------------------------------------------------------------
4000.1: log -3000 (baseline) 0.09(0.07+0.01) 0.09(0.07+0.01) -5.5%* 0.09(0.07+0.01) -4.1%
4000.2: log --raw -3000 (tree-only) 0.47(0.41+0.05) 0.47(0.40+0.05) -0.5% 0.45(0.38+0.06) -3.5%.
4000.3: log -p -3000 (Myers) 1.81(1.67+0.12) 1.81(1.67+0.13) +0.3% 1.99(1.84+0.12) +10.2%***
4000.4: log -p -3000 --histogram 1.79(1.66+0.11) 1.80(1.67+0.11) +0.4% 1.96(1.82+0.10) +9.2%***
4000.5: log -p -3000 --patience 2.17(2.02+0.13) 2.20(2.04+0.13) +1.3%. 2.33(2.18+0.13) +7.4%***
------------------------------------------------------------------------------------------------------------------
Significance hints: '.' 0.1 '*' 0.05 '**' 0.01 '***' 0.001
Noticed-by: Brian Gernhardt <brian@gernhardtsoftware.com>
Signed-off-by: Thomas Rast <trast@student.ethz.ch>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Jeff King [Tue, 1 May 2012 17:55:00 +0000 (13:55 -0400)]
t/gitweb-lib: use $PERL_PATH to run gitweb
The current code runs "perl gitweb.cgi" to test gitweb. This
will use whatever version of perl happens to be first in the
PATH. We are better off using the specific perl that the
user specified via PERL_PATH, which matches what gets put on
the #!-line of the built gitweb.cgi.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Zbigniew Jędrzejewski-Szmek [Tue, 1 May 2012 17:10:13 +0000 (19:10 +0200)]
tests: check --[short]stat output after chmod
Add a test to check 'diff --stat' output with a text file after chmod,
and the same for a binary file. This demonstrates that text and binary
files are treated differently, which can be misleading.
While at it, add tests to check --shortstat output, too.
Reported-by: Martin Mareš <mj@ucw.cz>
Signed-off-by: Zbigniew Jędrzejewski-Szmek <zbyszek@in.waw.pl>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Zbigniew Jędrzejewski-Szmek [Tue, 1 May 2012 17:10:12 +0000 (19:10 +0200)]
test: modernize style of t4006
Signed-off-by: Zbigniew Jędrzejewski-Szmek <zbyszek@in.waw.pl>
Signed-off-by: Junio C Hamano <gitster@pobox.com>