Code

git.git
14 years agoWhy is "git tag --contains" so slow?
Jeff King [Mon, 5 Jul 2010 12:39:23 +0000 (08:39 -0400)]
Why is "git tag --contains" so slow?

On Mon, Jul 05, 2010 at 08:27:23AM -0400, Jeff King wrote:

> As you probably guessed from the specificity of the number, I wrote a
> short program to actually traverse and find the worst skew. It takes
> about 5 seconds to run (unsurprisingly, since it is doing the same full
> traversal that we end up doing in the above numbers). So we could
> "autoskew" by setting up the configuration on clone, and then
> periodically updating it as part of "git gc".

This patch doesn't implement auto-detection of skew, but is the program
I used to calculate, and would provide the basis for such
auto-detection. It would be interesting to see average skew numbers for
popular repositories. You can run it as "git skew --all".

Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agodefault core.clockskew variable to one day
Jeff King [Mon, 5 Jul 2010 12:35:34 +0000 (08:35 -0400)]
default core.clockskew variable to one day

This is the slop value used by name-rev, so presumably is a
reasonable default.

Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agolimit "contains" traversals based on commit timestamp
Jeff King [Mon, 5 Jul 2010 12:34:19 +0000 (08:34 -0400)]
limit "contains" traversals based on commit timestamp

When looking for commits that contain other commits (e.g.,
via "git tag --contains"), we can end up traversing useless
portions of the graph. For example, if I am looking for a
tag that contains a commit made last week, there is not much
point in traversing portions of the history graph made five
years ago.

This optimization can provide massive speedups. For example,
doing "git tag --contains HEAD~200" in the linux-2.6
repository goes from:

  real    0m5.302s
  user    0m5.116s
  sys     0m0.184s

to:

  real    0m0.030s
  user    0m0.020s
  sys     0m0.008s

The downside is that we will no longer find some answers in
the face of extreme clock skew, as we will stop the
traversal early when seeing commits skewed too far into the
past.

Name-rev already implements a similar optimization, using a
"slop" of one day to allow for a certain amount of clock
skew in commit timestamps. This patch introduces a
"core.clockskew" variable, which allows specifying the
allowable amount of clock skew in seconds.  For safety, it
defaults to "none", causing a full traversal (i.e., no
change in behavior from previous versions).

Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agotag: speed up --contains calculation
Jeff King [Mon, 5 Jul 2010 12:33:36 +0000 (08:33 -0400)]
tag: speed up --contains calculation

When we want to know if commit A contains commit B (or any
one of a set of commits, B through Z), we generally
calculate the merge bases and see if B is a merge base of A
(or for a set, if any of the commits B through Z have that
property).

When we are going to check a series of commits A1 through An
to see whether each contains B (e.g., because we are
deciding which tags to show with "git tag --contains"), we
do a series of merge base calculations. This can be very
expensive, as we repeat a lot of traversal work.

Instead, let's leverage the fact that we are going to use
the same --contains list for each tag, and mark areas of the
commit graph is definitely containing those commits, or
definitely not containing those commits. Later tags can then
stop traversing as soon as they see a previously calculated
answer.

This sped up "git tag --contains HEAD~200" in the linux-2.6
repository from:

  real    0m15.417s
  user    0m15.197s
  sys     0m0.220s

to:

  real    0m5.329s
  user    0m5.144s
  sys     0m0.184s

Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoGit 1.7.2-rc1 v1.7.2-rc1
Junio C Hamano [Wed, 30 Jun 2010 19:22:13 +0000 (12:22 -0700)]
Git 1.7.2-rc1

Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agogit.spec.in: Add gitweb subpackage
Junio C Hamano [Wed, 30 Jun 2010 22:49:07 +0000 (15:49 -0700)]
git.spec.in: Add gitweb subpackage

Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoMerge branch 'ar/decorate-color'
Junio C Hamano [Wed, 30 Jun 2010 18:55:40 +0000 (11:55 -0700)]
Merge branch 'ar/decorate-color'

* ar/decorate-color:
  Add test for correct coloring of git log --decoration
  Allow customizable commit decorations colors
  log --decorate: Colorize commit decorations
  log-tree.c: Use struct name_decoration's type for classifying decoration
  commit.h: add 'type' to struct name_decoration

14 years agoMerge branch 'mg/doc-rev-parse-treepath-syntax'
Junio C Hamano [Wed, 30 Jun 2010 18:55:40 +0000 (11:55 -0700)]
Merge branch 'mg/doc-rev-parse-treepath-syntax'

* mg/doc-rev-parse-treepath-syntax:
  git-rev-parse.txt: Add more examples for caret and colon
  git-rev-parse.txt: Document ":path" specifier

14 years agoMerge branch 'cc/cherry-pick-stdin'
Junio C Hamano [Wed, 30 Jun 2010 18:55:39 +0000 (11:55 -0700)]
Merge branch 'cc/cherry-pick-stdin'

* cc/cherry-pick-stdin:
  revert: do not rebuild argv on heap
  revert: accept arbitrary rev-list options
  t3508 (cherry-pick): futureproof against unmerged files

14 years agoMerge branch 'jl/status-ignore-submodules'
Junio C Hamano [Wed, 30 Jun 2010 18:55:39 +0000 (11:55 -0700)]
Merge branch 'jl/status-ignore-submodules'

* jl/status-ignore-submodules:
  Add the option "--ignore-submodules" to "git status"
  git submodule: ignore dirty submodules for summary and status

Conflicts:
builtin/commit.c
t/t7508-status.sh
wt-status.c
wt-status.h

14 years agoMerge branch 'jk/url-decode'
Junio C Hamano [Wed, 30 Jun 2010 18:55:38 +0000 (11:55 -0700)]
Merge branch 'jk/url-decode'

* jk/url-decode:
  url_decode: URL scheme ends with a colon and does not require a slash

14 years agoMerge branch 'jn/grep-open'
Junio C Hamano [Wed, 30 Jun 2010 18:55:38 +0000 (11:55 -0700)]
Merge branch 'jn/grep-open'

* jn/grep-open:
  t/t7811-grep-open.sh: remove broken/redundant creation of fake "less" script
  t/t7811-grep-open.sh: ensure fake "less" is made executable
  t/lib-pager.sh: remove unnecessary '^' from 'expr' regular expression
  grep -O: allow optional argument specifying the pager (or editor)
  grep: Add the option '--open-files-in-pager'
  Unify code paths of threaded greps
  grep: refactor grep_objects loop into its own function

Conflicts:
t/t7006-pager.sh

14 years agoMerge branch 'jp/string-list-api-cleanup'
Junio C Hamano [Wed, 30 Jun 2010 18:55:38 +0000 (11:55 -0700)]
Merge branch 'jp/string-list-api-cleanup'

* jp/string-list-api-cleanup:
  string_list: Fix argument order for string_list_append
  string_list: Fix argument order for string_list_lookup
  string_list: Fix argument order for string_list_insert_at_index
  string_list: Fix argument order for string_list_insert
  string_list: Fix argument order for for_each_string_list
  string_list: Fix argument order for print_string_list

14 years agoMerge branch 'tr/rev-list-count'
Junio C Hamano [Wed, 30 Jun 2010 18:55:38 +0000 (11:55 -0700)]
Merge branch 'tr/rev-list-count'

* tr/rev-list-count:
  bash completion: Support "divergence from upstream" messages in __git_ps1
  rev-list: introduce --count option

Conflicts:
contrib/completion/git-completion.bash

14 years agoMerge branch 'as/maint-completion-set-u-fix'
Junio C Hamano [Wed, 30 Jun 2010 18:55:37 +0000 (11:55 -0700)]
Merge branch 'as/maint-completion-set-u-fix'

* as/maint-completion-set-u-fix:
  bash-completion: Fix __git_ps1 to work with "set -u"

14 years agoMerge branch 'mg/rev-parse-tests'
Junio C Hamano [Wed, 30 Jun 2010 18:55:37 +0000 (11:55 -0700)]
Merge branch 'mg/rev-parse-tests'

* mg/rev-parse-tests:
  t6018: make sure all tested symbolic names are different revs
  t6018: add tests for rev-list's --branches and --tags

14 years agoMerge branch 'jl/maint-diff-ignore-submodules'
Junio C Hamano [Wed, 30 Jun 2010 18:55:37 +0000 (11:55 -0700)]
Merge branch 'jl/maint-diff-ignore-submodules'

* jl/maint-diff-ignore-submodules:
  t4027,4041: Use test -s to test for an empty file
  Add optional parameters to the diff option "--ignore-submodules"
  git diff: rename test that had a conflicting name

14 years agoAdd test for correct coloring of git log --decoration
Nazri Ramliy [Tue, 29 Jun 2010 07:47:08 +0000 (15:47 +0800)]
Add test for correct coloring of git log --decoration

Signed-off-by: Nazri Ramliy <ayiehere@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoSync with 1.7.1.1
Junio C Hamano [Tue, 29 Jun 2010 18:24:36 +0000 (11:24 -0700)]
Sync with 1.7.1.1

Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agorevert: do not rebuild argv on heap
Jonathan Nieder [Mon, 14 Jun 2010 05:32:09 +0000 (00:32 -0500)]
revert: do not rebuild argv on heap

Set options in struct rev_info directly so we can reuse the
arguments collected from parse_options without modification.

This is just a cleanup; no noticeable change is intended.

Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoMerge commit 'v1.7.2-rc0~6^2' into cc/cherry-pick-stdin
Junio C Hamano [Tue, 29 Jun 2010 17:20:53 +0000 (10:20 -0700)]
Merge commit 'v1.7.2-rc0~6^2' into cc/cherry-pick-stdin

* commit 'v1.7.2-rc0~6^2':
  DWIM 'git show -5' to 'git show --do-walk -5'
  Documentation/SubmittingPatches: Fix typo in GMail section
  Documentation/config: describe status.submodulesummary

This commit fixes one test in t3508 by making "cherry-pick -<num>"
walk the history.

A test update from Elijah Newren is squashed as an evil merge.

14 years agoGit 1.7.1.1 v1.7.1.1
Junio C Hamano [Tue, 29 Jun 2010 16:59:56 +0000 (09:59 -0700)]
Git 1.7.1.1

Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agot/t9001: use egrep when regular expressions are involved
Brandon Casey [Mon, 28 Jun 2010 22:46:22 +0000 (17:46 -0500)]
t/t9001: use egrep when regular expressions are involved

Supplying backslashed, extended regular expressions to grep is not
portable.  Use egrep instead.

Signed-off-by: Brandon Casey <casey@nrlssc.navy.mil>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agogit-rev-parse.txt: Add more examples for caret and colon
Michael J Gruber [Mon, 28 Jun 2010 19:01:07 +0000 (21:01 +0200)]
git-rev-parse.txt: Add more examples for caret and colon

Several items in the caret, colon and friends section contain examples
already. Make sure they all come with examples, and that examples come
early so that they serve as a visual guide, as well.

Signed-off-by: Michael J Gruber <git@drmicha.warpmail.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoMerge branch 'maint'
Junio C Hamano [Tue, 29 Jun 2010 00:42:26 +0000 (17:42 -0700)]
Merge branch 'maint'

* maint:
  Update draft release notes to 1.7.1.1
  notes: Initialise variable to appease gcc
  notes: check number of parameters to "git notes copy"

14 years agoUpdate draft release notes to 1.7.1.1
Junio C Hamano [Tue, 29 Jun 2010 00:42:18 +0000 (17:42 -0700)]
Update draft release notes to 1.7.1.1

Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoMerge branch 'tr/send-email-8bit' into maint
Junio C Hamano [Mon, 28 Jun 2010 23:19:03 +0000 (16:19 -0700)]
Merge branch 'tr/send-email-8bit' into maint

* tr/send-email-8bit:
  send-email: ask about and declare 8bit mails

14 years agoMerge branch 'pb/maint-perl-errmsg-no-dir' into maint
Junio C Hamano [Mon, 28 Jun 2010 23:18:58 +0000 (16:18 -0700)]
Merge branch 'pb/maint-perl-errmsg-no-dir' into maint

* pb/maint-perl-errmsg-no-dir:
  Git.pm: better error message

14 years agoMerge branch 'js/maint-am-rebase-invalid-author' into maint
Junio C Hamano [Mon, 28 Jun 2010 23:18:43 +0000 (16:18 -0700)]
Merge branch 'js/maint-am-rebase-invalid-author' into maint

* js/maint-am-rebase-invalid-author:
  am: use get_author_ident_from_commit instead of mailinfo when rebasing

14 years agoMerge branch 'jc/maint-simpler-common-prefix' into maint
Junio C Hamano [Mon, 28 Jun 2010 23:18:15 +0000 (16:18 -0700)]
Merge branch 'jc/maint-simpler-common-prefix' into maint

* jc/maint-simpler-common-prefix:
  common_prefix: simplify and fix scanning for prefixes

14 years agoMerge branch 'bd/maint-unpack-trees-parawalk-fix' into maint
Junio C Hamano [Mon, 28 Jun 2010 23:18:02 +0000 (16:18 -0700)]
Merge branch 'bd/maint-unpack-trees-parawalk-fix' into maint

* bd/maint-unpack-trees-parawalk-fix:
  unpack-trees: Make index lookahead less pessimal

14 years agonotes: Initialise variable to appease gcc
Ramsay Jones [Mon, 21 Jun 2010 18:52:29 +0000 (19:52 +0100)]
notes: Initialise variable to appease gcc

gcc version 3.4.4 thinks that the 'cmp' variable could be used
while uninitialised and complains thus:

    notes.c: In function `write_each_non_note_until':
    notes.c:719: warning: 'cmp' might be used uninitialized in \
        this function

Note that gcc versions 4.1.2 and 4.4.0 do not issue this warning.

Signed-off-by: Ramsay Jones <ramsay@ramsay1.demon.co.uk>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agonotes: check number of parameters to "git notes copy"
Jeff King [Mon, 28 Jun 2010 08:59:07 +0000 (04:59 -0400)]
notes: check number of parameters to "git notes copy"

Otherwise we may segfault with too few parameters.

Signed-off-by: Jeff King <peff@peff.net>
Tested-by: Bert Wesarg <Bert.Wesarg@googlemail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agomsvc: Select the "fast" definition of the {get,put}_be32() macros
Ramsay Jones [Wed, 23 Jun 2010 19:47:02 +0000 (20:47 +0100)]
msvc: Select the "fast" definition of the {get,put}_be32() macros

On Intel machines, the msvc compiler defines the CPU architecture
macros _M_IX86 and _M_X64 (equivalent to __i386__ and __x86_64__
respectively). Use these macros in the pre-processor expression
to select the "fast" definition of the {get,put}_be32() macros.

Signed-off-by: Ramsay Jones <ramsay@ramsay1.demon.co.uk>
Acked-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agogit-rev-parse.txt: Document ":path" specifier
Michael J Gruber [Sun, 27 Jun 2010 13:15:22 +0000 (15:15 +0200)]
git-rev-parse.txt: Document ":path" specifier

The empty treeish in ":path" means "index". This is actually a special
case of the ":stage:path" syntax where it is documented, but mentioning
it also together with "treeish:path" is helpful, so do it.

Signed-off-by: Michael J Gruber <git@drmicha.warpmail.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoGit 1.7.2-rc0 v1.7.2-rc0
Junio C Hamano [Sun, 27 Jun 2010 19:01:12 +0000 (12:01 -0700)]
Git 1.7.2-rc0

Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoMerge branch 'cp/textconv-cat-file'
Junio C Hamano [Sun, 27 Jun 2010 19:07:55 +0000 (12:07 -0700)]
Merge branch 'cp/textconv-cat-file'

* cp/textconv-cat-file:
  git-cat-file.txt: Document --textconv
  t/t8007: test textconv support for cat-file
  textconv: support for cat_file
  sha1_name: add get_sha1_with_context()

14 years agoMerge branch 'pb/maint-perl-errmsg-no-dir'
Junio C Hamano [Sun, 27 Jun 2010 19:07:45 +0000 (12:07 -0700)]
Merge branch 'pb/maint-perl-errmsg-no-dir'

* pb/maint-perl-errmsg-no-dir:
  Git.pm: better error message

14 years agoMerge branch 'tr/send-email-8bit'
Junio C Hamano [Sun, 27 Jun 2010 19:07:45 +0000 (12:07 -0700)]
Merge branch 'tr/send-email-8bit'

* tr/send-email-8bit:
  send-email: ask about and declare 8bit mails

14 years agoMerge branch 'js/maint-am-rebase-invalid-author'
Junio C Hamano [Sun, 27 Jun 2010 19:07:44 +0000 (12:07 -0700)]
Merge branch 'js/maint-am-rebase-invalid-author'

* js/maint-am-rebase-invalid-author:
  am: use get_author_ident_from_commit instead of mailinfo when rebasing

14 years agoMerge branch 'ab/blame-textconv'
Junio C Hamano [Sun, 27 Jun 2010 19:07:44 +0000 (12:07 -0700)]
Merge branch 'ab/blame-textconv'

* ab/blame-textconv:
  t/t8006: test textconv support for blame
  textconv: support for blame
  textconv: make the API public

Conflicts:
diff.h

14 years agoMerge branch 'jn/show-num-walks'
Junio C Hamano [Sun, 27 Jun 2010 19:07:44 +0000 (12:07 -0700)]
Merge branch 'jn/show-num-walks'

* jn/show-num-walks:
  DWIM 'git show -5' to 'git show --do-walk -5'

14 years agot6018: make sure all tested symbolic names are different revs
Michael J Gruber [Fri, 14 May 2010 18:26:51 +0000 (20:26 +0200)]
t6018: make sure all tested symbolic names are different revs

Signed-off-by: Michael J Gruber <git@drmicha.warpmail.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agot6018: add tests for rev-list's --branches and --tags
Michael J Gruber [Fri, 14 May 2010 18:26:50 +0000 (20:26 +0200)]
t6018: add tests for rev-list's --branches and --tags

so that we know when they break.

Signed-off-by: Michael J Gruber <git@drmicha.warpmail.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoMerge branch 'jp/string-list-api-cleanup' into jn/grep-open
Julian Phillips [Fri, 25 Jun 2010 23:41:39 +0000 (00:41 +0100)]
Merge branch 'jp/string-list-api-cleanup' into jn/grep-open

An evil merge to adjust the series to cleaned-up API.

  From: Julian Phillips <julian@quantumfyre.co.uk>
  Subject: [PATCH v2 7/7] grep: fix string_list_append calls
  Date: Sat, 26 Jun 2010 00:41:39 +0100
  Message-ID: <20100625234140.18927.35025.julian@quantumfyre.co.uk>

* jp/string-list-api-cleanup:
  string_list: Fix argument order for string_list_append
  string_list: Fix argument order for string_list_lookup
  string_list: Fix argument order for string_list_insert_at_index
  string_list: Fix argument order for string_list_insert
  string_list: Fix argument order for for_each_string_list
  string_list: Fix argument order for print_string_list

Signed-off-by: Julian Phillips <julian@quantumfyre.co.uk>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agostring_list: Fix argument order for string_list_append
Julian Phillips [Fri, 25 Jun 2010 23:41:38 +0000 (00:41 +0100)]
string_list: Fix argument order for string_list_append

Update the definition and callers of string_list_append to use the
string_list as the first argument.  This helps make the string_list
API easier to use by being more consistent.

Signed-off-by: Julian Phillips <julian@quantumfyre.co.uk>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agostring_list: Fix argument order for string_list_lookup
Julian Phillips [Fri, 25 Jun 2010 23:41:37 +0000 (00:41 +0100)]
string_list: Fix argument order for string_list_lookup

Update the definition and callers of string_list_lookup to use the
string_list as the first argument.  This helps make the string_list
API easier to use by being more consistent.

Signed-off-by: Julian Phillips <julian@quantumfyre.co.uk>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agostring_list: Fix argument order for string_list_insert_at_index
Julian Phillips [Fri, 25 Jun 2010 23:41:36 +0000 (00:41 +0100)]
string_list: Fix argument order for string_list_insert_at_index

Update the definition and callers of string_list_insert_at_index to
use the string_list as the first argument.  This helps make the
string_list API easier to use by being more consistent.

Signed-off-by: Julian Phillips <julian@quantumfyre.co.uk>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agostring_list: Fix argument order for string_list_insert
Julian Phillips [Fri, 25 Jun 2010 23:41:35 +0000 (00:41 +0100)]
string_list: Fix argument order for string_list_insert

Update the definition and callers of string_list_insert to use the
string_list as the first argument.  This helps make the string_list
API easier to use by being more consistent.

Signed-off-by: Julian Phillips <julian@quantumfyre.co.uk>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agostring_list: Fix argument order for for_each_string_list
Julian Phillips [Fri, 25 Jun 2010 23:41:34 +0000 (00:41 +0100)]
string_list: Fix argument order for for_each_string_list

Update the definition and callers of for_each_string_list to use the
string_list as the first argument.  This helps make the string_list
API easier to use by being more consistent.

Signed-off-by: Julian Phillips <julian@quantumfyre.co.uk>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agostring_list: Fix argument order for print_string_list
Julian Phillips [Fri, 25 Jun 2010 23:41:33 +0000 (00:41 +0100)]
string_list: Fix argument order for print_string_list

Update the definition and callers of print_string_list to use the
string_list as the first argument.  This helps make the API easier to
use by being more consistent.

Signed-off-by: Julian Phillips <julian@quantumfyre.co.uk>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoMerge branch 'maint'
Junio C Hamano [Fri, 25 Jun 2010 18:45:27 +0000 (11:45 -0700)]
Merge branch 'maint'

* maint:
  msvc: Fix some compiler warnings
  Documentation: grep: fix asciidoc problem with --
  msvc: Fix some "expr evaluates to function" compiler warnings

14 years agoAdd the option "--ignore-submodules" to "git status"
Jens Lehmann [Fri, 25 Jun 2010 14:56:47 +0000 (16:56 +0200)]
Add the option "--ignore-submodules" to "git status"

In some use cases it is not desirable that "git status" considers
submodules that only contain untracked content as dirty. This may happen
e.g. when the submodule is not under the developers control and not all
build generated files have been added to .gitignore by the upstream
developers. Using the "untracked" parameter for the "--ignore-submodules"
option disables checking for untracked content and lets git diff report
them as changed only when they have new commits or modified content.

Sometimes it is not wanted to have submodules show up as changed when they
just contain changes to their work tree (this was the behavior before
1.7.0). An example for that are scripts which just want to check for
submodule commits while ignoring any changes to the work tree. Also users
having large submodules known not to change might want to use this option,
as the - sometimes substantial - time it takes to scan the submodule work
tree(s) is saved when using the "dirty" parameter.

And if you want to ignore any changes to submodules, you can now do that
by using this option without parameters or with "all" (when the config
option status.submodulesummary is set, using "all" will also suppress the
output of the submodule summary).

A new function handle_ignore_submodules_arg() is introduced to parse this
option new to "git status" in a single location, as "git diff" already
knew it.

Signed-off-by: Jens Lehmann <Jens.Lehmann@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agogit submodule: ignore dirty submodules for summary and status
Jens Lehmann [Fri, 25 Jun 2010 14:56:02 +0000 (16:56 +0200)]
git submodule: ignore dirty submodules for summary and status

The summary and status commands only care about submodule commits, so it is
rather pointless that they check for dirty work trees. This saves the time
needed to scan the submodules work tree. Even "git status" profits from these
savings when the status.submodulesummary config option is set, as this lead to
traversing the submodule work trees twice, once for status and once again for
the submodule summary. And if the submodule was just dirty, submodule summary
produced rather meaningless output anyway:

 * sub 1234567...1234567 (0):

Signed-off-by: Jens Lehmann <Jens.Lehmann@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agomsvc: Fix some compiler warnings
Ramsay Jones [Wed, 23 Jun 2010 19:47:50 +0000 (20:47 +0100)]
msvc: Fix some compiler warnings

In particular, using the normal (or production) compiler
warning level (-W3), msvc complains as follows:

.../sha1.c(244) : warning C4018: '<' : signed/unsigned mismatch
.../sha1.c(270) : warning C4244: 'function' : conversion from \
   'unsigned __int64' to 'unsigned long', possible loss of data
.../sha1.c(271) : warning C4244: 'function' : conversion from \
   'unsigned __int64' to 'unsigned long', possible loss of data

Note that gcc issues a similar complaint about line 244 when
compiling with -Wextra.

Signed-off-by: Ramsay Jones <ramsay@ramsay1.demon.co.uk>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agot4027,4041: Use test -s to test for an empty file
Brian Gernhardt [Fri, 25 Jun 2010 17:20:48 +0000 (13:20 -0400)]
t4027,4041: Use test -s to test for an empty file

The tests used a mixture of 'echo -n' (which is non-portable) and either
test_cmp or diff to check if a file is empty.  The much easier and portable
method to check for an empty file is '! test -s'

While we're in t4027, there was an excess test_done.  Remove it.

Signed-off-by: Brian Gernhardt <brian@gernhardtsoftware.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoDocumentation: grep: fix asciidoc problem with --
Christian Couder [Fri, 25 Jun 2010 02:16:24 +0000 (04:16 +0200)]
Documentation: grep: fix asciidoc problem with --

Asciidoc interprets two dashes separated by spaces as a single big
dash. So let's escape the first dash, so that "\--" will properly
appear as "--".

Signed-off-by: Christian Couder <chriscool@tuxfamily.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agorevert: accept arbitrary rev-list options
Christian Couder [Mon, 14 Jun 2010 05:29:38 +0000 (00:29 -0500)]
revert: accept arbitrary rev-list options

This can be useful to do something like:

git rev-list --reverse master -- README | git cherry-pick -n --stdin

without using xargs.

Signed-off-by: Christian Couder <chriscool@tuxfamily.org>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agourl_decode: URL scheme ends with a colon and does not require a slash
Junio C Hamano [Thu, 24 Jun 2010 20:36:30 +0000 (13:36 -0700)]
url_decode: URL scheme ends with a colon and does not require a slash

Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agogit-cat-file.txt: Document --textconv
Michael J Gruber [Thu, 24 Jun 2010 12:56:55 +0000 (14:56 +0200)]
git-cat-file.txt: Document --textconv

Signed-off-by: Michael J Gruber <git@drmicha.warpmail.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoAllow customizable commit decorations colors
Nazri Ramliy [Thu, 24 Jun 2010 00:21:16 +0000 (08:21 +0800)]
Allow customizable commit decorations colors

Signed-off-by: Nazri Ramliy <ayiehere@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agobash completion: Support "divergence from upstream" messages in __git_ps1
Andrew Sayers [Thu, 17 Jun 2010 21:32:02 +0000 (22:32 +0100)]
bash completion: Support "divergence from upstream" messages in __git_ps1

Add a notification in the command prompt specifying whether (and optionally how
far) your branch has diverged from its upstream.  This is especially helpful in
small teams that very frequently (forget to) push to each other.

Support git-svn upstream detection as a special case, as migrators from
centralised version control systems are especially likely to forget to push.

Support for other types of upstream than SVN should be easy to add if anyone is
so inclined.

Signed-off-by: Andrew Sayers <andrew-git@pileofstuff.org>
Acked-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agomsvc: Fix some "expr evaluates to function" compiler warnings
Ramsay Jones [Wed, 23 Jun 2010 19:40:19 +0000 (20:40 +0100)]
msvc: Fix some "expr evaluates to function" compiler warnings

In particular, the following warning is issued while compiling
notes.c:

    notes.c(927) : warning C4550: expression evaluates to a \
function which is missing an argument list

along with identical warnings on lines 928, 1016 and 1017.

In order to suppress the warning, we change the definition of
combine_notes_fn, so that the symbol type is an (explicit)
"pointer to function ...".  As a result, several other
declarations need some minor fix-up to take account of the
new typedef.

Signed-off-by: Ramsay Jones <ramsay@ramsay1.demon.co.uk>
Acked-by: Johan Herland <johan@herland.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoMerge branch 'jk/url-decode'
Junio C Hamano [Wed, 23 Jun 2010 17:43:28 +0000 (10:43 -0700)]
Merge branch 'jk/url-decode'

* jk/url-decode:
  url.c: "<scheme>://" part at the beginning should not be URL decoded

14 years agourl.c: "<scheme>://" part at the beginning should not be URL decoded
Junio C Hamano [Wed, 23 Jun 2010 17:27:39 +0000 (10:27 -0700)]
url.c: "<scheme>://" part at the beginning should not be URL decoded

When using the protocol git+ssh:// for example we do not want to
decode the '+' as a space. The url decoding must take place only
for the server name and parameters.

This fixes a regression introduced in 9d2e942.

Initial-fix-by: Pascal Obry <pascal.obry@gmail.com>
Acked-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoUpdate draft release notes to 1.7.2
Junio C Hamano [Tue, 22 Jun 2010 17:03:04 +0000 (10:03 -0700)]
Update draft release notes to 1.7.2

Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoMerge branch 'jc/maint-simpler-common-prefix'
Junio C Hamano [Tue, 22 Jun 2010 16:45:23 +0000 (09:45 -0700)]
Merge branch 'jc/maint-simpler-common-prefix'

* jc/maint-simpler-common-prefix:
  common_prefix: simplify and fix scanning for prefixes

14 years agoMerge branch 'sb/format-patch-signature'
Junio C Hamano [Tue, 22 Jun 2010 16:45:22 +0000 (09:45 -0700)]
Merge branch 'sb/format-patch-signature'

* sb/format-patch-signature:
  completion: Add --signature and format.signature
  format-patch: Add a signature option (--signature)

14 years agoMerge branch 'mg/pretty-magic-space'
Junio C Hamano [Tue, 22 Jun 2010 16:45:22 +0000 (09:45 -0700)]
Merge branch 'mg/pretty-magic-space'

* mg/pretty-magic-space:
  pretty: Introduce ' ' modifier to add space if non-empty

Conflicts:
pretty.c

14 years agoMerge branch 'jn/gitweb-return-or-exit-cleanup'
Junio C Hamano [Tue, 22 Jun 2010 16:45:22 +0000 (09:45 -0700)]
Merge branch 'jn/gitweb-return-or-exit-cleanup'

* jn/gitweb-return-or-exit-cleanup:
  gitweb: Return or exit after done serving request

Conflicts:
gitweb/gitweb.perl

14 years agoMerge branch 'bd/maint-unpack-trees-parawalk-fix'
Junio C Hamano [Tue, 22 Jun 2010 16:45:22 +0000 (09:45 -0700)]
Merge branch 'bd/maint-unpack-trees-parawalk-fix'

* bd/maint-unpack-trees-parawalk-fix:
  unpack-trees: Make index lookahead less pessimal

14 years agoMerge branch 'cc/cherry-pick-series'
Junio C Hamano [Tue, 22 Jun 2010 16:45:21 +0000 (09:45 -0700)]
Merge branch 'cc/cherry-pick-series'

* cc/cherry-pick-series:
  Documentation/revert: describe passing more than one commit
  Documentation/cherry-pick: describe passing more than one commit
  revert: add tests to check cherry-picking many commits
  revert: allow cherry-picking more than one commit
  revert: change help_msg() to take no argument
  revert: refactor code into a do_pick_commit() function
  revert: use run_command_v_opt() instead of execv_git_cmd()
  revert: cleanup code for -x option

14 years agoMerge branch 'jc/rev-list-ancestry-path'
Junio C Hamano [Tue, 22 Jun 2010 16:45:21 +0000 (09:45 -0700)]
Merge branch 'jc/rev-list-ancestry-path'

* jc/rev-list-ancestry-path:
  revision: Turn off history simplification in --ancestry-path mode
  revision: Fix typo in --ancestry-path error message
  Documentation/rev-list-options.txt: Explain --ancestry-path
  Documentation/rev-list-options.txt: Fix missing line in example history graph
  revision: --ancestry-path

14 years agoMerge branch 'lt/extended-sha1-match-commit-with-regexp'
Junio C Hamano [Tue, 22 Jun 2010 16:45:21 +0000 (09:45 -0700)]
Merge branch 'lt/extended-sha1-match-commit-with-regexp'

* lt/extended-sha1-match-commit-with-regexp:
  Make :/ accept a regex rather than a fixed pattern

14 years agoMerge branch 'maint'
Junio C Hamano [Tue, 22 Jun 2010 16:35:36 +0000 (09:35 -0700)]
Merge branch 'maint'

* maint:
  Update draft release notes to 1.7.1.1
  tests: remove unnecessary '^' from 'expr' regular expression

Conflicts:
diff.c

14 years agoUpdate draft release notes to 1.7.1.1
Junio C Hamano [Tue, 22 Jun 2010 16:18:55 +0000 (09:18 -0700)]
Update draft release notes to 1.7.1.1

Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoMerge branch 'ic/maint-rebase-i-abort' into maint
Junio C Hamano [Tue, 22 Jun 2010 16:31:48 +0000 (09:31 -0700)]
Merge branch 'ic/maint-rebase-i-abort' into maint

* ic/maint-rebase-i-abort:
  rebase -i: Abort cleanly if new base cannot be checked out

14 years agoMerge branch 'cc/maint-commit-reflog-msg' into maint
Junio C Hamano [Tue, 22 Jun 2010 16:31:48 +0000 (09:31 -0700)]
Merge branch 'cc/maint-commit-reflog-msg' into maint

* cc/maint-commit-reflog-msg:
  commit: use value of GIT_REFLOG_ACTION env variable as reflog message

14 years agoMerge branch 'jk/maint-advice-empty-amend' into maint
Junio C Hamano [Tue, 22 Jun 2010 16:31:48 +0000 (09:31 -0700)]
Merge branch 'jk/maint-advice-empty-amend' into maint

* jk/maint-advice-empty-amend:
  commit: give advice on empty amend

14 years agoMerge branch 'tc/commit-abbrev-fix' into maint
Junio C Hamano [Tue, 22 Jun 2010 16:31:47 +0000 (09:31 -0700)]
Merge branch 'tc/commit-abbrev-fix' into maint

* tc/commit-abbrev-fix:
  commit::print_summary(): don't use format_commit_message()
  t7502-commit: add summary output tests for empty and merge commits
  t7502-commit: add tests for summary output

14 years agoMerge branch 'jn/document-rebase-i-p-limitation' into maint
Junio C Hamano [Tue, 22 Jun 2010 16:31:47 +0000 (09:31 -0700)]
Merge branch 'jn/document-rebase-i-p-limitation' into maint

* jn/document-rebase-i-p-limitation:
  rebase -i -p: document shortcomings

14 years agoMerge branch 'jn/checkout-doc' into maint
Junio C Hamano [Tue, 22 Jun 2010 16:31:47 +0000 (09:31 -0700)]
Merge branch 'jn/checkout-doc' into maint

* jn/checkout-doc:
  Documentation/checkout: clarify description
  Documentation/checkout: clarify description

14 years agoMerge branch 'cc/maint-diff-CC-binary' into maint
Junio C Hamano [Tue, 22 Jun 2010 16:04:14 +0000 (09:04 -0700)]
Merge branch 'cc/maint-diff-CC-binary' into maint

* cc/maint-diff-CC-binary:
  diff: fix "git show -C -C" output when renaming a binary file

Conflicts:
diff.c

14 years agoMerge branch 'jc/t9129-any-utf8' into maint
Junio C Hamano [Tue, 22 Jun 2010 15:31:53 +0000 (08:31 -0700)]
Merge branch 'jc/t9129-any-utf8' into maint

* jc/t9129-any-utf8:
  t9129: fix UTF-8 locale detection

14 years agoMerge branch 'cb/ls-files-cdup' into maint
Junio C Hamano [Tue, 22 Jun 2010 15:31:46 +0000 (08:31 -0700)]
Merge branch 'cb/ls-files-cdup' into maint

* cb/ls-files-cdup:
  ls-files: allow relative pathspec
  quote.c: separate quoting and relative path generation

14 years agoMerge branch 'tc/merge-m-log' into maint
Junio C Hamano [Tue, 22 Jun 2010 15:31:25 +0000 (08:31 -0700)]
Merge branch 'tc/merge-m-log' into maint

* tc/merge-m-log:
  merge: --log appends shortlog to message if specified
  fmt-merge-msg: add function to append shortlog only
  fmt-merge-msg: refactor merge title formatting
  fmt-merge-msg: minor refactor of fmt_merge_msg()
  merge: rename variable
  merge: update comment
  t7604-merge-custom-message: show that --log doesn't append to -m
  t7604-merge-custom-message: shift expected output creation

14 years agoMerge branch 'ph/clone-message-reword' into maint
Junio C Hamano [Tue, 22 Jun 2010 15:31:20 +0000 (08:31 -0700)]
Merge branch 'ph/clone-message-reword' into maint

* ph/clone-message-reword:
  clone: reword messages to match the end-user perception

14 years agoMerge branch 'jn/maint-amend-missing-name' into maint
Junio C Hamano [Tue, 22 Jun 2010 15:30:44 +0000 (08:30 -0700)]
Merge branch 'jn/maint-amend-missing-name' into maint

* jn/maint-amend-missing-name:
  commit --amend: cope with missing display name

14 years agoMerge branch 'pc/remove-warn' into maint
Junio C Hamano [Tue, 22 Jun 2010 15:30:38 +0000 (08:30 -0700)]
Merge branch 'pc/remove-warn' into maint

* pc/remove-warn:
  Remove a redundant errno test in a usage of remove_path
  Introduce remove_or_warn function
  Implement the rmdir_or_warn function
  Generalise the unlink_or_warn function

14 years agotests: remove unnecessary '^' from 'expr' regular expression
Junio C Hamano [Mon, 21 Jun 2010 18:18:54 +0000 (11:18 -0700)]
tests: remove unnecessary '^' from 'expr' regular expression

As Brandon noticed, a regular expression match given to 'expr' is already
anchored at the beginning.  Some versions of expr even complain about this.

Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agot/t7811-grep-open.sh: remove broken/redundant creation of fake "less" script
Brandon Casey [Mon, 21 Jun 2010 17:37:15 +0000 (12:37 -0500)]
t/t7811-grep-open.sh: remove broken/redundant creation of fake "less" script

The fake "less" script was already created in a previous test titled
'setup: fake "less"', so it is redundant.  Additionally, it is broken since
the redirection of 'cat' is to a file named 'less', but the chmod operates
on the file named by the $less variable which may not contain the value
'less'.

So, just remove this code, and rely on the creation of the fake "less"
script performed earlier within the test script.

Signed-off-by: Brandon Casey <casey@nrlssc.navy.mil>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agot/t7811-grep-open.sh: ensure fake "less" is made executable
Brandon Casey [Mon, 21 Jun 2010 17:37:14 +0000 (12:37 -0500)]
t/t7811-grep-open.sh: ensure fake "less" is made executable

The fake "less" script was not being made executable.  This can cause the
tests that follow to fail.  This failure is not apparent on platforms which
have DEFAULT_PAGER set to the string "less", since lib-pager.sh will have
set the $less variable to "less" and the SIMPLEPAGER prerequisite will have
been set, and so the "less" script will have already been created properly
and made executable in test 2 'git grep -O'.  On platforms which set
DEFAULT_PAGER to something like "more", no such script will have been
previously created, and tests 7 and 8 will fail.

So, add a call to chmod to make the fake "less" script executable.

Signed-off-by: Brandon Casey <casey@nrlssc.navy.mil>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agot/lib-pager.sh: remove unnecessary '^' from 'expr' regular expression
Brandon Casey [Mon, 21 Jun 2010 17:37:13 +0000 (12:37 -0500)]
t/lib-pager.sh: remove unnecessary '^' from 'expr' regular expression

Regular expressions matched by 'expr' have an implicit '^' at the beginning
of them and so are anchored to the beginning of the string.  Using the '^'
character to mean "match at the beginning", is redundant and could produce
the wrong result if 'expr' implementations interpret the '^' as a literal
'^'.  Additionally, GNU expr 5.97 complains like this:

   expr: warning: unportable BRE: `^[a-z][a-z]*$': using `^' as the first character of the basic regular expression is not portable; it is being ignored

Signed-off-by: Brandon Casey <casey@nrlssc.navy.mil>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
14 years agoMerge branch 'js/maint-receive-pack-symref-alias'
Junio C Hamano [Mon, 21 Jun 2010 13:02:50 +0000 (06:02 -0700)]
Merge branch 'js/maint-receive-pack-symref-alias'

* js/maint-receive-pack-symref-alias:

14 years agoMerge branch 'cc/maint-commit-reflog-msg'
Junio C Hamano [Mon, 21 Jun 2010 13:02:50 +0000 (06:02 -0700)]
Merge branch 'cc/maint-commit-reflog-msg'

* cc/maint-commit-reflog-msg:
  commit: use value of GIT_REFLOG_ACTION env variable as reflog message

14 years agoMerge branch 'ic/maint-rebase-i-abort'
Junio C Hamano [Mon, 21 Jun 2010 13:02:50 +0000 (06:02 -0700)]
Merge branch 'ic/maint-rebase-i-abort'

* ic/maint-rebase-i-abort:
  rebase -i: Abort cleanly if new base cannot be checked out

14 years agoMerge branch 'jk/maint-advice-empty-amend'
Junio C Hamano [Mon, 21 Jun 2010 13:02:49 +0000 (06:02 -0700)]
Merge branch 'jk/maint-advice-empty-amend'

* jk/maint-advice-empty-amend:
  commit: give advice on empty amend

14 years agoMerge branch 'eb/core-eol'
Junio C Hamano [Mon, 21 Jun 2010 13:02:49 +0000 (06:02 -0700)]
Merge branch 'eb/core-eol'

* eb/core-eol:
  Add "core.eol" config variable
  Rename the "crlf" attribute "text"
  Add per-repository eol normalization
  Add tests for per-repository eol normalization

Conflicts:
Documentation/config.txt
Makefile

14 years agoMerge branch 'fg/autocrlf'
Junio C Hamano [Mon, 21 Jun 2010 13:02:47 +0000 (06:02 -0700)]
Merge branch 'fg/autocrlf'

* fg/autocrlf:
  autocrlf: Make it work also for un-normalized repositories

14 years agoMerge branch 'sm/branch-broken-ref'
Junio C Hamano [Mon, 21 Jun 2010 13:02:47 +0000 (06:02 -0700)]
Merge branch 'sm/branch-broken-ref'

* sm/branch-broken-ref:
  branch: don't fail listing branches if one of the commits wasn't found
  branch: exit status now reflects if branch listing finds an error