git
18 years agoMerge part of 'js/fmt-patch' for RFC2822 dates into 'sp/reflog'
Junio C Hamano [Fri, 19 May 2006 22:25:57 +0000 (15:25 -0700)] 
Merge part of 'js/fmt-patch' for RFC2822 dates into 'sp/reflog'

An earlier patch from Shawn Pearce dependes on a change that is
only in "next".  I do not want to make this series hostage to
the yet-to-graduate js/fmt-patch branch, but let's try fixing it
by merging the early parts of the branch to see what happens.

Right now, 'sp/reflog' will not be in "next" for now, so I won't
have to regret this -- if this merge causes problem down the road
merging I can always rebuild the topic branch ;-).

18 years agoEnable ref log creation in git checkout -b.
Shawn Pearce [Fri, 19 May 2006 09:17:16 +0000 (05:17 -0400)] 
Enable ref log creation in git checkout -b.

Switch git checkout -b to use git-update-ref rather than echo and
a shell I/O redirection.  This is more in line with typical GIT
commands and allows -b to be logged according to the normal ref
logging rules.

Added -l option to allow users to create the ref log at the same
time as creating a branch.

Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoCreate/delete branch ref logs.
Shawn Pearce [Fri, 19 May 2006 09:16:58 +0000 (05:16 -0400)] 
Create/delete branch ref logs.

When crating a new branch offer '-l' as a way for the user to
quickly enable ref logging for the new branch.

When deleting a branch also delete its ref log.

Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoInclude ref log detail in commit, reset, etc.
Shawn Pearce [Fri, 19 May 2006 09:16:18 +0000 (05:16 -0400)] 
Include ref log detail in commit, reset, etc.

When updating a ref at the direction of the user include a reason why
head was changed as part of the ref log (assuming it was enabled).

Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoChange order of -m option to update-ref.
Shawn Pearce [Fri, 19 May 2006 09:15:50 +0000 (05:15 -0400)] 
Change order of -m option to update-ref.

The actual position doesn't matter but most people prefer to see
options appear before the arguments.

Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoCorrect force_write bug in refs.c
Shawn Pearce [Fri, 19 May 2006 09:15:28 +0000 (05:15 -0400)] 
Correct force_write bug in refs.c

My earlier attempt at forcing a write for non-existant refs worked;
it forced a write for pretty much all refs.  This corrects the
condition to only force a write for refs which don't exist yet.

Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoChange 'master@noon' syntax to 'master@{noon}'.
Shawn Pearce [Fri, 19 May 2006 07:29:43 +0000 (03:29 -0400)] 
Change 'master@noon' syntax to 'master@{noon}'.

Its ambiguous to parse "master@2006-05-17 18:30:foo" when foo is
meant as a file name and ":30" is meant as 30 minutes past 6 pm.
Therefore all date specifications in a sha1 expression must now
appear within brackets and the ':' splitter used for the path name
in a sha1 expression ignores ':' appearing within brackets.

Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoLog ref updates made by fetch.
Shawn Pearce [Fri, 19 May 2006 07:29:26 +0000 (03:29 -0400)] 
Log ref updates made by fetch.

If a ref is changed by http-fetch, local-fetch or ssh-fetch
record the change and the remote URL/name in the log for the ref.
This requires loading the config file to check logAllRefUpdates.

Also fixed a bug in the ref lock generation; the log file name was
not being produced right due to a bad prefix length.

Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoForce writing ref if it doesn't exist.
Shawn Pearce [Fri, 19 May 2006 07:29:05 +0000 (03:29 -0400)] 
Force writing ref if it doesn't exist.

Normally we try to skip writing a ref if its value hasn't changed
but in the special case that the ref doesn't exist but the new
value is going to be 0{40} then force writing the ref anyway.

Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoAdded logs/ directory to repository layout.
Shawn Pearce [Fri, 19 May 2006 07:28:46 +0000 (03:28 -0400)] 
Added logs/ directory to repository layout.

Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoGeneral ref log reading improvements.
Shawn Pearce [Fri, 19 May 2006 07:28:19 +0000 (03:28 -0400)] 
General ref log reading improvements.

Corrected the log starting time displayed in the error message
(as it was always showing the epoch due to a bad input to strtoul).

Improved the log parser so we only scan backwards towards the
'\n' from the end of the prior log; during this scan the last '>'
is remembered to improve performance (rather than scanning forward
to it).

If the log record matched is the last log record in the file only
use its new sha1 value if the date matches exactly; otherwise we
leave the passed in sha1 alone as it already contains the current
value of the ref.  This way lookups of dates later than the log
end to stick with the current ref value in case the ref was updated
without logging.

If it looks like someone changed the ref without logging it and we
are going to return the sha1 which should have been valid during
the missing period then warn the user that there might be log data
missing and thus their query result may not be accurate.  The check
isn't perfect as its just based on comparing the old and new sha1
values between the two log records but its better than not checking
at all.

Implemented test cases based on git-rev-parse for most of the
boundary conditions.

Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoFix ref log parsing so it works properly.
Shawn Pearce [Wed, 17 May 2006 22:34:48 +0000 (18:34 -0400)] 
Fix ref log parsing so it works properly.

The log parser was only ever matching the last log record due to
calling strtoul on "> 1136091609" rather than " 1136091609".  Also
once a match for '@' has been found after the name of the ref there
is no point in looking for another '@' within the remaining text.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoSupport 'master@2 hours ago' syntax
Shawn Pearce [Wed, 17 May 2006 09:56:09 +0000 (05:56 -0400)] 
Support 'master@2 hours ago' syntax

Extended sha1 expressions may now include date specifications
which indicate a point in time within the local repository's
history.  If the ref indicated to the left of '@' has a log in
$GIT_DIR/logs/<ref> then the value of the ref at the time indicated
by the specification is obtained from the ref's log.

Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoLog ref updates to logs/refs/<ref>
Shawn Pearce [Wed, 17 May 2006 09:55:40 +0000 (05:55 -0400)] 
Log ref updates to logs/refs/<ref>

If config parameter core.logAllRefUpdates is true or the log
file already exists then append a line to ".git/logs/refs/<ref>"
whenever git-update-ref <ref> is executed.  Each log line contains
the following information:

  oldsha1 <SP> newsha1 <SP> committer <LF>

where committer is the current user, date, time and timezone in
the standard GIT ident format.  If the caller is unable to append
to the log file then git-update-ref will fail without updating <ref>.

An optional message may be included in the log line with the -m flag.

Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoConvert update-ref to use ref_lock API.
Shawn Pearce [Wed, 17 May 2006 09:55:19 +0000 (05:55 -0400)] 
Convert update-ref to use ref_lock API.

This conversion also adds the '-m' switch to update-ref allowing
the caller to record why the ref is changing.  At present this is
merely copied down into the ref_lock API.

Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoImprove abstraction of ref lock/write.
Shawn Pearce [Wed, 17 May 2006 09:55:02 +0000 (05:55 -0400)] 
Improve abstraction of ref lock/write.

Created 'struct ref_lock' to contain the data necessary to perform
a ref update.  This change improves writing a ref as the file names
are generated only once (rather than twice) and supports following
symrefs (up to the maximum depth).  Further the ref_lock structure
provides room to extend the update API with ref logging.

Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoRemove unnecessary local in get_ref_sha1.
Shawn Pearce [Wed, 17 May 2006 09:54:46 +0000 (05:54 -0400)] 
Remove unnecessary local in get_ref_sha1.

Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agocommit: allow --pretty= args to be abbreviated
Eric Wong [Mon, 15 May 2006 00:20:46 +0000 (17:20 -0700)] 
commit: allow --pretty= args to be abbreviated

Signed-off-by: Eric Wong <normalperson@yhbt.net>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoRemove old "git-grep.sh" remnants
Linus Torvalds [Tue, 16 May 2006 23:46:31 +0000 (16:46 -0700)] 
Remove old "git-grep.sh" remnants

It's built-in now.

Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoMerge branch 'maint'
Junio C Hamano [Wed, 17 May 2006 00:21:02 +0000 (17:21 -0700)] 
Merge branch 'maint'

* maint:
  merge-base: Clarify the comments on post processing.
  Update the documentation for git-merge-base

18 years agoMerge branch 'np/pack'
Junio C Hamano [Wed, 17 May 2006 00:20:24 +0000 (17:20 -0700)] 
Merge branch 'np/pack'

* np/pack:
  improve depth heuristic for maximum delta size
  pack-object: slightly more efficient
  simple euristic for further free packing improvements

18 years agomerge-base: Clarify the comments on post processing.
Junio C Hamano [Wed, 17 May 2006 00:18:22 +0000 (17:18 -0700)] 
merge-base: Clarify the comments on post processing.

The comment fooled myself believing that we still had an
unsolved horizon effect.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoimprove depth heuristic for maximum delta size
Nicolas Pitre [Tue, 16 May 2006 20:29:14 +0000 (16:29 -0400)] 
improve depth heuristic for maximum delta size

This provides a linear decrement on the penalty related to delta depth
instead of being an 1/x function.  With this another 5% reduction is
observed on packs for both the GIT repo and the Linux kernel repo, as
well as fixing a pack size regression in another sample repo I have.

Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoMerge branch 'se/tag'
Junio C Hamano [Tue, 16 May 2006 06:43:27 +0000 (23:43 -0700)] 
Merge branch 'se/tag'

* se/tag:
  Strip useless "tags/" prefix from git-tag -l output

18 years agoMerge branch 'se/rev-parse'
Junio C Hamano [Tue, 16 May 2006 06:43:23 +0000 (23:43 -0700)] 
Merge branch 'se/rev-parse'

* se/rev-parse:
  Add "--branches", "--tags" and "--remotes" options to git-rev-parse.

18 years agoMerge branch 'se/diff'
Junio C Hamano [Tue, 16 May 2006 06:42:37 +0000 (23:42 -0700)] 
Merge branch 'se/diff'

* se/diff:
  Convert some "apply --summary" users to "diff --summary".
  Add "--summary" option to git diff.

18 years agoMerge branch 'se/rebase'
Junio C Hamano [Tue, 16 May 2006 06:35:24 +0000 (23:35 -0700)] 
Merge branch 'se/rebase'

* se/rebase:
  Make git rebase interactive help match documentation.

18 years agoUpdate the documentation for git-merge-base
Fredrik Kuivinen [Tue, 16 May 2006 05:58:15 +0000 (07:58 +0200)] 
Update the documentation for git-merge-base

Signed-off-by: Fredrik Kuivinen <freku045@student.liu.se>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agobuiltin-diff: fix comparison between two blobs.
Junio C Hamano [Tue, 16 May 2006 02:05:50 +0000 (19:05 -0700)] 
builtin-diff: fix comparison between two blobs.

The code forgot that setup_revisions() leaves parsed object
names in reverse in the list.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoMerge branch 'lt/oneway'
Junio C Hamano [Tue, 16 May 2006 01:15:26 +0000 (18:15 -0700)] 
Merge branch 'lt/oneway'

* lt/oneway:
  read-tree --reset -u fix.
  read-tree -u one-way merge fix to check out locally modified paths.
  Simplify "git reset --hard"
  Allow one-way tree merge to remove old files

18 years agoMerge branch 'ew/send-email'
Junio C Hamano [Tue, 16 May 2006 01:15:03 +0000 (18:15 -0700)] 
Merge branch 'ew/send-email'

* ew/send-email:
  send-email: quiet some warnings, reject invalid addresses
  send-email: allow sendmail binary to be used instead of SMTP

18 years agoMerge branch 'lt/config'
Junio C Hamano [Tue, 16 May 2006 01:12:57 +0000 (18:12 -0700)] 
Merge branch 'lt/config'

* lt/config:

18 years agoMerge branch 'jc/grep'
Junio C Hamano [Tue, 16 May 2006 01:12:06 +0000 (18:12 -0700)] 
Merge branch 'jc/grep'

* jc/grep: (22 commits)
  Fix silly typo in new builtin grep
  builtin-grep: unparse more command line options.
  builtin-grep: use external grep when we can take advantage of it
  builtin-grep: -F (--fixed-strings)
  builtin-grep: -w fix
  builtin-grep: typofix
  builtin-grep: tighten argument parsing.
  builtin-grep: documentation
  Teach -f <file> option to builtin-grep.
  builtin-grep: -L (--files-without-match).
  builtin-grep: binary files -a and -I
  builtin-grep: terminate correctly at EOF
  builtin-grep: tighten path wildcard vs tree traversal.
  builtin-grep: support -w (--word-regexp).
  builtin-grep: support -c (--count).
  builtin-grep: allow more than one patterns.
  builtin-grep: allow -<n> and -[ABC]<n> notation for context lines.
  builtin-grep: printf %.*s length is int, not ptrdiff_t.
  builtin-grep: do not use setup_revisions()
  builtin-grep: support '-l' option.
  ...

18 years agoMerge branch 'lt/diff'
Junio C Hamano [Tue, 16 May 2006 01:09:15 +0000 (18:09 -0700)] 
Merge branch 'lt/diff'

* lt/diff:
  git diff: support "-U" and "--unified" options properly

18 years agoFix silly typo in new builtin grep
Linus Torvalds [Tue, 16 May 2006 00:54:01 +0000 (17:54 -0700)] 
Fix silly typo in new builtin grep

The "-F" flag apparently got mis-translated due to some over-eager
copy-paste work into a duplicate "-H" when using the external grep.

Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoMerge branch 'fix'
Junio C Hamano [Mon, 15 May 2006 20:48:22 +0000 (13:48 -0700)] 
Merge branch 'fix'

* fix:
  Fix pack-index issue on 64-bit platforms a bit more portably.
  Install git-send-email by default
  Fix compilation on newer NetBSD systems
  git config syntax updates
  Another config file parsing fix.
  checkout: use --aggressive when running a 3-way merge (-m).

18 years agobuiltin-grep: unparse more command line options.
Junio C Hamano [Mon, 15 May 2006 20:28:01 +0000 (13:28 -0700)] 
builtin-grep: unparse more command line options.

The earlier one to use external grep missed some often used options.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoFix pack-index issue on 64-bit platforms a bit more portably. v1.3.3
Junio C Hamano [Mon, 15 May 2006 19:52:00 +0000 (12:52 -0700)] 
Fix pack-index issue on 64-bit platforms a bit more portably.

Apparently <stdint.h> is not enough for uint32_t on OpenBSD; use
"unsigned int" -- hopefully that would stay 32-bit on every
platform we care about, at least until we update the pack-index
file format.

Our sha1 routines optimized for architectures use uint32_t and
expects '#include <stdint.h>' to be enough, so OpenBSD on arm or
ppc might have similar issues down the road, I dunno.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoInstall git-send-email by default
Eric Wong [Mon, 15 May 2006 02:26:56 +0000 (19:26 -0700)] 
Install git-send-email by default

After 567ffeb7722eefab3991cb894c96548b92b57cc2 and
4bc87a28be020a6bf7387161c65ea3d8e4a0228b, git-send-email no
longer requires any non-standard Perl modules, so there's no
reason to special-case it.

Signed-off-by: Eric Wong <normalperson@yhbt.net>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoFix compilation on newer NetBSD systems
Dennis Stosberg [Thu, 11 May 2006 17:35:31 +0000 (19:35 +0200)] 
Fix compilation on newer NetBSD systems

NetBSD >=2.0 has iconv() in libc.  A libiconv is not required and
does not exist.

See: http://netbsd.gw.com/cgi-bin/man-cgi?iconv+3+NetBSD-2.0

[jc: with a bit of simplification later discussed on the list.]

Signed-off-by: Dennis Stosberg <dennis@stosberg.net>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agopack-object: slightly more efficient
Nicolas Pitre [Mon, 15 May 2006 17:47:16 +0000 (13:47 -0400)] 
pack-object: slightly more efficient

Avoid creating a delta index for objects with maximum depth since they
are not going to be used as delta base anyway.  This also reduce peak
memory usage slightly as the current object's delta index is not useful
until the next object in the loop is considered for deltification. This
saves a bit more than 1% on CPU usage.

Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agosimple euristic for further free packing improvements
Nicolas Pitre [Mon, 15 May 2006 15:40:05 +0000 (11:40 -0400)] 
simple euristic for further free packing improvements

Given that the early eviction of objects with maximum delta depth
may exhibit bad packing on its own, why not considering a bias against
deep base objects in try_delta() to mitigate that bad behavior.

This patch adjust the MAX_size allowed for a delta based on the depth of
the base object as well as enabling the early eviction of max depth
objects from the object window.  When used separately, those two things
produce slightly better and much worse results respectively.  But their
combined effect is a surprising significant packing improvement.

With this really simple patch the GIT repo gets nearly 15% smaller, and
the Linux kernel repo about 5% smaller, with no significantly measurable
CPU usage difference.

Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoread-tree --reset -u fix.
Linus Torvalds [Mon, 15 May 2006 15:09:31 +0000 (08:09 -0700)] 
read-tree --reset -u fix.

The previous commit makes -u to mean "I do want to remove the
local changes, just update it from the read tree" only for
one-way merge.  It makes sense to have it depend on the
"--reset" flag instead.

Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agosend-email: quiet some warnings, reject invalid addresses
Eric Wong [Mon, 15 May 2006 09:41:01 +0000 (02:41 -0700)] 
send-email: quiet some warnings, reject invalid addresses

I'm not sure why we never actually rejected invalid addresses in
the first place.  We just seemed to be using our email validity
checkers to kill duplicates.

Now we just drop invalid email addresses completely and warn
the user about it.

Since we support local sendmail, we'll also accept username-only
addresses.

Signed-off-by: Eric Wong <normalperson@yhbt.net>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agosend-email: allow sendmail binary to be used instead of SMTP
Eric Wong [Mon, 15 May 2006 09:34:44 +0000 (02:34 -0700)] 
send-email: allow sendmail binary to be used instead of SMTP

This should make local mailing possible for machines without
a connection to an SMTP server.

It'll default to using /usr/sbin/sendmail or /usr/lib/sendmail
if no SMTP server is specified (the default).  If it can't find
either of those paths, it'll fall back to connecting to an SMTP
server on localhost.

Signed-off-by: Eric Wong <normalperson@yhbt.net>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoStrip useless "tags/" prefix from git-tag -l output
Sean [Mon, 15 May 2006 00:07:39 +0000 (20:07 -0400)] 
Strip useless "tags/" prefix from git-tag -l output

18 years agoMerge branch 'fix'
Junio C Hamano [Mon, 15 May 2006 07:51:37 +0000 (00:51 -0700)] 
Merge branch 'fix'

* fix:
  Ensure author & committer before asking for commit message.

18 years agoMerge branch 'fix' into maint
Junio C Hamano [Mon, 15 May 2006 07:49:25 +0000 (00:49 -0700)] 
Merge branch 'fix' into maint

* fix:
  Ensure author & committer before asking for commit message.

18 years agoread-tree -u one-way merge fix to check out locally modified paths.
Junio C Hamano [Mon, 15 May 2006 07:46:05 +0000 (00:46 -0700)] 
read-tree -u one-way merge fix to check out locally modified paths.

The "-u" flag means "update the working tree files", but to
other types of merges, it also implies "I want to keep my local
changes" -- because they prevent local changes from getting lost
by using verify_uptodate.  The one-way merge is different from
other merges in that its purpose is opposite of doing something
else while keeping unrelated local changes.  The point of
one-way merge is to nuke local changes.  So while it feels
somewhat wrong that this actively loses local changes, it is the
right thing to do.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoInstall git-send-email by default
Eric Wong [Mon, 15 May 2006 02:26:56 +0000 (19:26 -0700)] 
Install git-send-email by default

After 567ffeb7722eefab3991cb894c96548b92b57cc2 and
4bc87a28be020a6bf7387161c65ea3d8e4a0228b, git-send-email no
longer requires any non-standard Perl modules, so there's no
reason to special-case it.

Signed-off-by: Eric Wong <normalperson@yhbt.net>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agosend-email: address expansion for common mailers
Eric Wong [Mon, 15 May 2006 02:13:44 +0000 (19:13 -0700)] 
send-email: address expansion for common mailers

mutt, gnus, pine, mailrc formats should be supported.

Testing and feedback for correctness and completeness of all formats
and support for additional formats would be good.

Nested expansions are also supported.

More than one alias file to be used.

All alias file formats must still of be the same type, though.

Two git repo-config keys are required for this
(as suggested by Ryan Anderson):

    sendemail.aliasesfile = <filename of aliases file>
    sendemail.aliasfiletype = (mutt|gnus|pine|mailrc)

Signed-off-by: Eric Wong <normalperson@yhbt.net>
Acked-by: Ryan Anderson <ryan@michonline.com>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agobuiltin-grep: use external grep when we can take advantage of it
Linus Torvalds [Mon, 15 May 2006 03:49:15 +0000 (20:49 -0700)] 
builtin-grep: use external grep when we can take advantage of it

It's not perfect, but it gets the "git grep some-random-string" down to
the good old half-a-second range for the kernel.

It should convert more of the argument flags for "grep", that should be
trivial to expand (I did a few just as an example). It should also bother
to try to return the right "hit" value (which it doesn't, right now - the
code is kind of there, but I didn't actually bother to do it _right_).

Also, right now it _just_ limits by number of arguments, but it should
also strictly speaking limit by total argument size (ie add up the length
of the filenames, and do the "exec_grep()" flush call if it's bigger than
some random value like 32kB).

But I think that it's _conceptually_ doing all the right things, and it
seems to work. So maybe somebody else can do some of the final polish.

Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agodiffstat rename squashing fix.
Junio C Hamano [Mon, 15 May 2006 05:07:28 +0000 (22:07 -0700)] 
diffstat rename squashing fix.

When renaming leading/a/filename to leading/b/filename (and
"filename" is sufficiently long), we tried to squash the rename
to "leading/{a => b}/filename".  However, when "/a" or "/b" part
is empty, we underflowed and tried to print a substring of
length -1.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoSimplify "git reset --hard"
Linus Torvalds [Sun, 14 May 2006 18:20:37 +0000 (11:20 -0700)] 
Simplify "git reset --hard"

Now that the one-way merge strategy does the right thing wrt files that do
not exist in the result, just remove all the random crud we did in "git
reset" to do this all by hand.

Instead, just pass in "-u" to git-read-tree when we do a hard reset, and
depend on git-read-tree to update the working tree appropriately.

This basically means that git reset turns into

# Always update the HEAD ref
git update-ref HEAD "$rev"

case "--soft"
# do nothing to index/working tree
case "--hard"
# read index _and_ update working tree
git-read-tree --reset -u "$rev"
case "--mixed"
# update just index, report on working tree differences
git-read-tree --reset "$rev"
git-update-index --refresh

which is what it was always semantically doing, it just did it in a
rather strange way because it was written to not expect git-read-tree to
do anything to the working tree.

Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoAllow one-way tree merge to remove old files
Linus Torvalds [Sun, 14 May 2006 17:43:50 +0000 (10:43 -0700)] 
Allow one-way tree merge to remove old files

For some random reason (probably just because nobody noticed), the one-way
merge strategy didn't mark deleted files as deleted, so if you used

git-read-tree -m -u <newtree>

it would update the files that got changed in the index, but it would not
delete the files that got deleted.

This should fix it, and I can't imagine that anybody depends on the old
strange "update only existing files" behaviour.

Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoConvert some "apply --summary" users to "diff --summary".
Sean [Sun, 14 May 2006 12:16:06 +0000 (08:16 -0400)] 
Convert some "apply --summary" users to "diff --summary".

Signed-off-by: Sean Estabrooks <seanlkml@sympatico.ca>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoAdd "--summary" option to git diff.
Sean [Sun, 14 May 2006 12:13:49 +0000 (08:13 -0400)] 
Add "--summary" option to git diff.

Remove the need to pipe git diff through git apply to
get the extended headers summary.

Signed-off-by: Sean Estabrooks <seanlkml@sympatico.ca>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoMake git rebase interactive help match documentation.
Sean [Sun, 14 May 2006 03:34:08 +0000 (23:34 -0400)] 
Make git rebase interactive help match documentation.

Signed-off-by: Sean Estabrooks <seanlkml@sympatico.ca>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoEnsure author & committer before asking for commit message.
Sean [Sun, 14 May 2006 03:09:32 +0000 (23:09 -0400)] 
Ensure author & committer before asking for commit message.

It's better to find out you need to fix your author and
committer information before you enter a long commit message.

Signed-off-by: Sean Estabrooks <seanlkml@sympatico.ca>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agogit diff: support "-U" and "--unified" options properly
Linus Torvalds [Sat, 13 May 2006 20:23:48 +0000 (13:23 -0700)] 
git diff: support "-U" and "--unified" options properly

We used to parse "-U" and "--unified" as part of the GIT_DIFF_OPTS
environment variable, but strangely enough we would _not_ parse them as
part of the normal diff command line (where we only accepted "-u").

This adds parsing of -U and --unified, both with an optional numeric
argument. So now you can just say

git diff --unified=5

to get a unified diff with a five-line context, instead of having to do
something silly like

GIT_DIFF_OPTS="--unified=5" git diff -u

(that silly format does continue to still work, of course).

Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoAdd "--branches", "--tags" and "--remotes" options to git-rev-parse.
Sean [Sun, 14 May 2006 01:43:00 +0000 (21:43 -0400)] 
Add "--branches", "--tags" and "--remotes" options to git-rev-parse.

"git branch" uses "rev-parse --all" and becomes much too slow when
there are many tags (it scans all refs).  Use the new "--branches"
option of rev-parse to speed things up.

Signed-off-by: Sean Estabrooks <seanlkml@sympatico.ca>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoMerge branch 'fix'
Junio C Hamano [Sun, 14 May 2006 23:20:15 +0000 (16:20 -0700)] 
Merge branch 'fix'

* fix:
  include header to define uint32_t, necessary on Mac OS X

18 years agoMerge branch 'fix' into maint
Junio C Hamano [Sun, 14 May 2006 23:20:09 +0000 (16:20 -0700)] 
Merge branch 'fix' into maint

* fix:
  include header to define uint32_t, necessary on Mac OS X

18 years agoinclude header to define uint32_t, necessary on Mac OS X
Ben Clifford [Sun, 14 May 2006 20:34:56 +0000 (21:34 +0100)] 
include header to define uint32_t, necessary on Mac OS X

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoMerge branch 'ml/cvs'
Junio C Hamano [Sun, 14 May 2006 07:42:16 +0000 (00:42 -0700)] 
Merge branch 'ml/cvs'

* ml/cvs:
  Change to allow subdir updates from Eclipse
  Many fixes for most operations in Eclipse.
  Added logged warnings for CVS error returns
  cvsserver: use git-rev-list instead of git-log
  git-cvsexportcommit: Add -f(orce) and -m(essage prefix) flags, small cleanups.

18 years agoMerge branch 'fix'
Junio C Hamano [Sun, 14 May 2006 05:24:18 +0000 (22:24 -0700)] 
Merge branch 'fix'

* fix:
  Fix git-pack-objects for 64-bit platforms

18 years agoMerge branch 'lt/fix-config' into lt/config
Junio C Hamano [Sat, 13 May 2006 21:10:48 +0000 (14:10 -0700)] 
Merge branch 'lt/fix-config' into lt/config

* lt/fix-config:
  git config syntax updates
  Another config file parsing fix.
  checkout: use --aggressive when running a 3-way merge (-m).
  Fix git-pack-objects for 64-bit platforms

with manual adjustment of t/t1300 for "git repo-config --list" option.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agogit config syntax updates
Linus Torvalds [Tue, 9 May 2006 19:24:02 +0000 (12:24 -0700)] 
git config syntax updates

This updates the hierarchical section name syntax to

[section<space>+"<randomstring>"]

where the only rule for "randomstring" is that it can't contain a newline,
and if you really want to insert a double-quote, you do it with \".

It turns that into the section name "secion.randomstring".  The
"section" part is still case insensitive, but the "randomstring"
part is case sensitive.

So you could use this for things like

[email "torvalds@osdl.org"]
name = Linus Torvalds

if you wanted to do the "email->name" conversion as part of the config
file format (I'm not claiming that is sensible, I'm just giving it as an
insane example). That would show up as the association

email.torvalds@osdl.org.name -> Linus Torvalds

which is easy to parse (the "." in the email _looks_ ambiguous, but it
isn't: you know that there will always be a single key-name, so you find
the key name with "strrchr(name, '.')" and things are entirely
unambiguous).

Repo-config is updated to be able to parse the new format, and also
write things out in the new format.

[jc: rolled two patches from Linus and one fix-up from Sean into one,
 with additional adjustments for t/t1300 test to check the case
 insensitiveness of section base and variable and case sensitiveness
 of the extended section part.  Then stripped some part off to make
 the result applicable to the stale 1.3.X series that does not have
 recent enhancements. ]

Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Sean Estabrooks <seanlkml@sympatico.ca>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoAnother config file parsing fix.
sean [Sat, 6 May 2006 18:14:02 +0000 (14:14 -0400)] 
Another config file parsing fix.

If the variable we need to store should go into a section
that currently only has a single variable (not matching
the one we're trying to insert), we will already be into
the next section before we notice we've bypassed the correct
location to insert the variable.

To handle this case we store the current location as soon
as we find a variable matching the section of our new
variable.

This breakage was brought up by Linus.

Signed-off-by: Sean Estabrooks <seanlkml@sympatico.ca>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agocheckout: use --aggressive when running a 3-way merge (-m).
Junio C Hamano [Wed, 10 May 2006 02:23:23 +0000 (19:23 -0700)] 
checkout: use --aggressive when running a 3-way merge (-m).

After doing an in-index 3-way merge, we always do the stock
"merge-index merge-one-file" without doing anything fancy;
use of --aggressive helps performance quite a bit.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoFix git-pack-objects for 64-bit platforms
Dennis Stosberg [Thu, 11 May 2006 17:36:32 +0000 (19:36 +0200)] 
Fix git-pack-objects for 64-bit platforms

The offset of an object in the pack is recorded as a 4-byte integer
in the index file.  When reading the offset from the mmap'ed index
in prepare_pack_revindex(), the address is dereferenced as a long*.
This works fine as long as the long type is four bytes wide.  On
NetBSD/sparc64, however, a long is 8 bytes wide and so dereferencing
the offset produces garbage.

[jc: taking suggestion by Linus to use uint32_t]

Signed-off-by: Dennis Stosberg <dennis@stosberg.net>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agofix diff-delta bad memory access
Nicolas Pitre [Wed, 10 May 2006 16:26:08 +0000 (12:26 -0400)] 
fix diff-delta bad memory access

It cannot be assumed that the given buffer will never be moved when
shrinking the allocated memory size with realloc().  So let's ignore
that optimization for now.

This patch makes Electric Fence happy on Linux.

Signed-off-by: Nicolas Pitre <nico@cam.org>
Acked-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoMerge branch 'fix'
Junio C Hamano [Wed, 10 May 2006 02:24:16 +0000 (19:24 -0700)] 
Merge branch 'fix'

* fix:
  checkout: use --aggressive when running a 3-way merge (-m).
  revert/cherry-pick: use aggressive merge.

18 years agocheckout: use --aggressive when running a 3-way merge (-m).
Junio C Hamano [Wed, 10 May 2006 02:23:23 +0000 (19:23 -0700)] 
checkout: use --aggressive when running a 3-way merge (-m).

After doing an in-index 3-way merge, we always do the stock
"merge-index merge-one-file" without doing anything fancy;
use of --aggressive helps performance quite a bit.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agorevert/cherry-pick: use aggressive merge.
Linus Torvalds [Wed, 10 May 2006 02:22:25 +0000 (19:22 -0700)] 
revert/cherry-pick: use aggressive merge.

After doing an in-index 3-way merge, we always do the stock
"merge-index merge-one-file" without doing anything fancy;
use of --aggressive helps performance quite a bit.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agobuiltin-grep: -F (--fixed-strings)
Junio C Hamano [Wed, 10 May 2006 01:28:41 +0000 (18:28 -0700)] 
builtin-grep: -F (--fixed-strings)

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agobuiltin-grep: -w fix
Junio C Hamano [Wed, 10 May 2006 01:27:56 +0000 (18:27 -0700)] 
builtin-grep: -w fix

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agobuiltin-grep: typofix
Junio C Hamano [Wed, 10 May 2006 01:15:21 +0000 (18:15 -0700)] 
builtin-grep: typofix

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoMerge branch 'jc/clean'
Junio C Hamano [Tue, 9 May 2006 23:52:54 +0000 (16:52 -0700)] 
Merge branch 'jc/clean'

* jc/clean:
  Teach git-clean optional <paths>... parameters.

18 years agoMerge branch 'mw/alternates'
Junio C Hamano [Tue, 9 May 2006 23:45:45 +0000 (16:45 -0700)] 
Merge branch 'mw/alternates'

* mw/alternates:
  clone: don't clone the info/alternates file
  test case for transitive info/alternates
  Transitively read alternatives

18 years agoMerge branch 'jc/xsha1'
Junio C Hamano [Tue, 9 May 2006 23:44:59 +0000 (16:44 -0700)] 
Merge branch 'jc/xsha1'

* jc/xsha1:
  get_sha1() - fix infinite loop on nonexistent stage.
  get_sha1(): :path and :[0-3]:path to extract from index.

18 years agoMerge branch 'jc/again'
Junio C Hamano [Tue, 9 May 2006 23:40:53 +0000 (16:40 -0700)] 
Merge branch 'jc/again'

* jc/again:
  Fix users of prefix_path() to free() only when necessary
  update-index --again: take optional pathspecs
  update-index --again

18 years agoMerge branch 'np/delta'
Junio C Hamano [Tue, 9 May 2006 23:40:28 +0000 (16:40 -0700)] 
Merge branch 'np/delta'

* np/delta:
  improve diff-delta with sparse and/or repetitive data
  tiny optimization to diff-delta
  replace adler32 with Rabin's polynomial in diff-delta
  use delta index data when finding best delta matches
  split the diff-delta interface

18 years agoMerge branch 'jc/bindiff'
Junio C Hamano [Tue, 9 May 2006 21:16:56 +0000 (14:16 -0700)] 
Merge branch 'jc/bindiff'

* jc/bindiff:
  improve base85 generated assembly code
  binary diff and apply: testsuite.
  binary diff: further updates.
  binary patch.

18 years agoMerge branch 'fix'
Junio C Hamano [Tue, 9 May 2006 20:54:42 +0000 (13:54 -0700)] 
Merge branch 'fix'

* fix:
  read-cache.c: use xcalloc() not calloc()
  apply: fix infinite loop with multiple patches with --index

18 years agoMerge branch 'tojunio' of http://locke.catalyst.net.nz/git/git-martinlanghoff into...
Junio C Hamano [Tue, 9 May 2006 19:25:21 +0000 (12:25 -0700)] 
Merge branch 'tojunio' of locke.catalyst.net.nz/git/git-martinlanghoff into ml/cvs

* 'tojunio' of http://locke.catalyst.net.nz/git/git-martinlanghoff:
  Change to allow subdir updates from Eclipse
  Many fixes for most operations in Eclipse.
  Added logged warnings for CVS error returns
  cvsserver: use git-rev-list instead of git-log
  git-cvsexportcommit: Add -f(orce) and -m(essage prefix) flags, small cleanups.

18 years agoread-cache.c: use xcalloc() not calloc()
Yakov Lerner [Tue, 9 May 2006 16:14:00 +0000 (16:14 +0000)] 
read-cache.c: use xcalloc() not calloc()

Elsewhere we use xcalloc(); we should consistently do so.

Signed-off-by: Yakov Lerner <iler.ml@gmail.com>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoapply: fix infinite loop with multiple patches with --index
Eric Wong [Tue, 9 May 2006 08:08:23 +0000 (01:08 -0700)] 
apply: fix infinite loop with multiple patches with --index

When multiple patches are passed to git-apply, it will attempt
to open multiple file descriptors to an index, which means
multiple entries will be in the circular cache_file_list.

This change makes git-apply only open the index once and
write the index at exit.

Signed-off-by: Eric Wong <normalperson@yhbt.net>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agobuiltin-grep: tighten argument parsing.
Junio C Hamano [Tue, 9 May 2006 06:55:47 +0000 (23:55 -0700)] 
builtin-grep: tighten argument parsing.

I mistyped

git grep next -e '"^@"' '*.c'

and got many hits that contain "next" without complaint.
Obviously what I meant to say was:

git grep -e '"^@"' next -- '*.c'

This tightens the argument parsing rule a bit:

 - All "grep" parameters should come first;

 - If there is no -e nor -f to specify pattern, the first non
   option string is the parameter;

 - After that, zero or more revs can follow.

 - An optional '--' can be present, and is skipped.

 - All the rest are pathspecs.  If '--' was not there, they must
   be paths that exist in the working tree.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoTeach git-clean optional <paths>... parameters.
Junio C Hamano [Mon, 8 May 2006 19:02:44 +0000 (12:02 -0700)] 
Teach git-clean optional <paths>... parameters.

When optional paths arguments are given, git-clean passes them
to underlying git-ls-files; with this, you can say:

git clean 'temp-*'

to clean only the garbage files whose names begin with 'temp-'.

Signed-off-by: Junio C Hamano <junkio@cox.net>
Signed-off-by: Pavel Roskin <proski@gnu.org>
18 years agoMerge branch 'fix'
Junio C Hamano [Mon, 8 May 2006 23:40:23 +0000 (16:40 -0700)] 
Merge branch 'fix'

* fix:
  Separate object name errors from usage errors
  Documentation: {caret} fixes (git-rev-list.txt)
  Fix "git diff --stat" with long filenames
  Fix repo-config set-multivar error return path.

18 years agoSeparate object name errors from usage errors
Dmitry V. Levin [Mon, 8 May 2006 21:43:38 +0000 (01:43 +0400)] 
Separate object name errors from usage errors

Separate object name errors from usage errors.

Signed-off-by: Dmitry V. Levin <ldv@altlinux.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoget_sha1() - fix infinite loop on nonexistent stage.
Junio C Hamano [Mon, 8 May 2006 22:44:06 +0000 (15:44 -0700)] 
get_sha1() - fix infinite loop on nonexistent stage.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoDocumentation: {caret} fixes (git-rev-list.txt)
Junio C Hamano [Mon, 8 May 2006 20:46:53 +0000 (13:46 -0700)] 
Documentation: {caret} fixes (git-rev-list.txt)

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agobuiltin-grep: documentation
Junio C Hamano [Mon, 8 May 2006 20:28:49 +0000 (13:28 -0700)] 
builtin-grep: documentation

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoTeach -f <file> option to builtin-grep.
Junio C Hamano [Mon, 8 May 2006 20:28:27 +0000 (13:28 -0700)] 
Teach -f <file> option to builtin-grep.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoimprove base85 generated assembly code
Nicolas Pitre [Mon, 8 May 2006 15:31:11 +0000 (11:31 -0400)] 
improve base85 generated assembly code

This code is arguably pretty hot, if you use binary patches of course.
This patch helps gcc generate both smaller and faster code especially in
the error free path.

Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoFix "git diff --stat" with long filenames
Linus Torvalds [Mon, 8 May 2006 16:46:53 +0000 (09:46 -0700)] 
Fix "git diff --stat" with long filenames

When we cut off the front of a filename to make it fit on the line, we add
a "..." in front. However, the way the "git diff" code was written, we
will never reset the prefix back to the empty string, so every single
filename afterwards will have the "..." prefix, whether appropriate or
not.

You can see this with "git diff v2.6.16.." on the current kernel tree,
since there are filenames with long names that changed there:

 [ snip snip ]
 Documentation/filesystems/vfs.txt                  |  229
 .../firmware_class/firmware_sample_driver.c        |    3
 .../firmware_sample_firmware_class.c               |    1
 ...Documentation/fujitsu/frv/kernel-ABI.txt           |  192
 ...Documentation/hwmon/w83627hf                       |    4
 [ snip snip ]

notice how the two Documentation/firmware** filenames caused the "..." to
be added, but then the later filenames don't want it, and it also screws
up the alignment of the line numbering afterwards.

Trivially fixed by moving the declaration (and initial setting) of the
"prefix" variable into the for-loop where it is used.

Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoFix repo-config set-multivar error return path.
Junio C Hamano [Mon, 8 May 2006 04:27:30 +0000 (21:27 -0700)] 
Fix repo-config set-multivar error return path.

This hopefully fixes the problem an earlier commit 5d8ee9ceb attemted
to fix.

Signed-off-by: Junio C Hamano <junkio@cox.net>
18 years agoclone: don't clone the info/alternates file
Martin Waitz [Sun, 7 May 2006 18:19:33 +0000 (20:19 +0200)] 
clone: don't clone the info/alternates file

Now that the cloned alternates file is parsed, too we don't need to
copy it into our new repository, we just reference it.

Signed-off-by: Martin Waitz <tali@admingilde.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>