Code

git.git
19 years agoDon't add parents to the commit list if we have already
Linus Torvalds [Sun, 24 Apr 2005 02:21:28 +0000 (19:21 -0700)]
Don't add parents to the commit list if we have already
seen them.

Otherwise any merges will make the parent list explode.

19 years agoAdd "rev-list" program that uses the new time-based commit listing.
Linus Torvalds [Sun, 24 Apr 2005 02:04:40 +0000 (19:04 -0700)]
Add "rev-list" program that uses the new time-based commit listing.

This is probably what you'd want to see for "git log".

19 years ago[PATCH] Various transport programs
Daniel Barkalow [Sun, 24 Apr 2005 01:47:23 +0000 (18:47 -0700)]
[PATCH] Various transport programs

This patch adds three similar and related programs. http-pull downloads
objects from an HTTP server; rpull downloads objects by using ssh and
rpush on the other side; and rpush uploads objects by using ssh and rpull
on the other side.

The algorithm should be sufficient to make the network throughput required
depend only on how much content is new, not at all on how much content the
repository contains.

The combination should enable people to have remote repositories by way of
ssh login for authenticated users and HTTP for anonymous access.

Signed-Off-By: Daniel Barkalow <barkalow@iabervon.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] Replace merge-base implementation
Daniel Barkalow [Sun, 24 Apr 2005 01:47:23 +0000 (18:47 -0700)]
[PATCH] Replace merge-base implementation

The old implementation was a nice algorithm, but, unfortunately, it could
be confused in some cases and would not necessarily do the obvious thing
if one argument was decended from the other. This version fixes that by
changing the criterion to the most recent common ancestor.

Signed-Off-By: Daniel Barkalow <barkalow@iabervon.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] Additional functions for the objects database
Daniel Barkalow [Sun, 24 Apr 2005 01:47:23 +0000 (18:47 -0700)]
[PATCH] Additional functions for the objects database

This adds two functions: one to check if an object is present in the local
database, and one to add an object to the local database by reading it
from a file descriptor and checking its hash.

Signed-Off-By: Daniel Barkalow <barkalow@iabervon.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] Parse tree objects completely
Daniel Barkalow [Sun, 24 Apr 2005 01:47:23 +0000 (18:47 -0700)]
[PATCH] Parse tree objects completely

This adds the contents of trees to struct tree.

Signed-Off-By: Daniel Barkalow <barkalow@iabervon.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] Add some functions for commit lists
Daniel Barkalow [Sun, 24 Apr 2005 01:47:23 +0000 (18:47 -0700)]
[PATCH] Add some functions for commit lists

This adds a function for inserting an item in a commit list, a function
for sorting a commit list by date, and a function for progressively
scanning a commit history from most recent to least recent.

Signed-Off-By: Daniel Barkalow <barkalow@iabervon.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] Simplify building of programs
Jonas Fonseca [Sun, 24 Apr 2005 01:41:48 +0000 (18:41 -0700)]
[PATCH] Simplify building of programs

Do not first build .o files when building programs.

Signed-off-by: Jonas Fonseca <fonseca@diku.dk>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] Fix broken diff-cache output on added files
Petr Baudis [Sun, 24 Apr 2005 01:05:07 +0000 (18:05 -0700)]
[PATCH] Fix broken diff-cache output on added files

Added files were errorneously reported with the - prefix by diff-cache,
obviously leading to great confusion.

Signed-off-by: Petr Baudis <pasky@ucw.cz>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years agoMake a hack to convert-cache for missing author dates in old
Linus Torvalds [Sat, 23 Apr 2005 23:48:32 +0000 (16:48 -0700)]
Make a hack to convert-cache for missing author dates in old
archives (notably the old sparse one).

Very hacky. But hopefully we can do the conversion once, and never
worry about this ever again.

19 years agoMake "convert-cache" able to handle the really old archive formats
Linus Torvalds [Sat, 23 Apr 2005 23:37:31 +0000 (16:37 -0700)]
Make "convert-cache" able to handle the really old archive formats

This includes the old-style "flat tree" object, and the old broken
date format. Well, enough of the date format to convert the sparse
archive, at least.

19 years agoUse O_NOATIME when opening the sha1 files.
Linus Torvalds [Sat, 23 Apr 2005 18:09:32 +0000 (11:09 -0700)]
Use O_NOATIME when opening the sha1 files.

We really don't care about atime, and it sucks to dirty the
inode cache just for it.

This is more than a one-liner only because we need to be able to
clear the O_NOATIME flag in case some of the objects are owned
by others (in which case open will return EPERM), and because not
everybody has the O_NOATIME flag.

19 years ago[PATCH] PPC assembly implementation of SHA1
Paul Mackerras [Sat, 23 Apr 2005 06:08:43 +0000 (23:08 -0700)]
[PATCH] PPC assembly implementation of SHA1

Here is a SHA1 implementation with the core written in PPC assembly.
On my 2GHz G5, it does 218MB/s, compared to 135MB/s for the openssl
version or 45MB/s for the mozilla version.

Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years agoNew "diff-cache" implementation.
Linus Torvalds [Sat, 23 Apr 2005 00:15:28 +0000 (17:15 -0700)]
New "diff-cache" implementation.

This one is about a million times simpler, and much more likely to be
correct too.

Instead of trying to match up a tree object against the index, we just
read in the tree object side-by-side into the index, and just walk the
resulting index file. This was what all the read-tree cleanups were
all getting to.

19 years agoMove "read_tree()" to "tree.c" to be used as a generic helper function.
Linus Torvalds [Fri, 22 Apr 2005 23:42:37 +0000 (16:42 -0700)]
Move "read_tree()" to "tree.c" to be used as a generic helper function.

Next step: make "diff-cache" use it.

19 years agoClean up and simplify read-tree a bit.
Linus Torvalds [Fri, 22 Apr 2005 23:35:06 +0000 (16:35 -0700)]
Clean up and simplify read-tree a bit.

This is preparation for moving parts of it into "tree.c" to be used
as a library function.

19 years agoAdd support for alternate SHA1 library implementations.
Linus Torvalds [Thu, 21 Apr 2005 19:33:22 +0000 (12:33 -0700)]
Add support for alternate SHA1 library implementations.

This one includes the Mozilla SHA1 implementation sent in by Edgar Toernig.
It's dual-licenced under MPL-1.1 or GPL, so in the context of git, we
obviously use the GPL version.

Side note: the Mozilla SHA1 implementation is about twice as fast as the
default openssl one on my G5, but the default openssl one has optimized
x86 assembly language on x86. So choose wisely.

19 years agoSplit up Makefile library list handling with separate entries for
Linus Torvalds [Thu, 21 Apr 2005 19:14:46 +0000 (12:14 -0700)]
Split up Makefile library list handling with separate entries for
zlib and libssl.

I'll start giving people choices here..

19 years agoInclude <limits.h> in commit.c for ULONG_MAX. Remove old "revision.h".
Linus Torvalds [Thu, 21 Apr 2005 18:21:33 +0000 (11:21 -0700)]
Include <limits.h> in commit.c for ULONG_MAX. Remove old "revision.h".

The old revision.h helper header isn't used any more, but I never noticed
it until I started grepping for ULONG_MAX users.

19 years agoAdd support for a "GIT_INDEX_FILE" environment variable.
Linus Torvalds [Thu, 21 Apr 2005 17:55:18 +0000 (10:55 -0700)]
Add support for a "GIT_INDEX_FILE" environment variable.

We use that to specify alternative index files, which can be useful
if you want to (for example) generate a temporary index file to do
some specific operation that you don't want to mess with your main
one with.

It defaults to the regular ".git/index" if it hasn't been specified.

19 years agoAdd the ability to prefix something to the pathname to "checkout-cache.c"
Linus Torvalds [Thu, 21 Apr 2005 17:15:02 +0000 (10:15 -0700)]
Add the ability to prefix something to the pathname to "checkout-cache.c"

This basically makes it trivial to use checkout-cache as a "export as
tree" function. Just read the desired tree into the index, and do a

checkout-cache --prefix=export-dir/ -a

and checkout-cache will "export" the cache into the specified directory.

NOTE! The final "/" is important. The exported name is literally just
prefixed with the specified string, so you can also do something like

checkout-cache --prefix=.merged- Makefile

to check out the currently cached copy of "Makefile" into the file
".merged-Makefile".

19 years agoFix NSEC compile problem, and properly parse the rev-tree cmd line.
Linus Torvalds [Thu, 21 Apr 2005 16:58:24 +0000 (09:58 -0700)]
Fix NSEC compile problem, and properly parse the rev-tree cmd line.

The rev-tree thing just happened to work. It shouldn't have.

19 years ago[PATCH] Usage-string fixes.
Junio C Hamano [Thu, 21 Apr 2005 02:49:16 +0000 (19:49 -0700)]
[PATCH] Usage-string fixes.

Usage string fixes to make maintenance easier (only one instance
of a string to update not multiple copies).  I've spotted and
corrected inconsistent usage text in diff-tree while doing this.

Also diff-cache and read-tree usage text have been corrected to
match their up-to-date features.  Earlier, neither "--cached"
form of diff-cache nor "-m single-merge" form of read-tree were
described.

Signed-off-by: Junio C Hamano <junkio@cox.net>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years agoDuh, just make git-export.c use the proper syntax, everything is fine.
Linus Torvalds [Thu, 21 Apr 2005 02:19:11 +0000 (19:19 -0700)]
Duh, just make git-export.c use the proper syntax, everything is fine.

19 years agoFix up some problems from the commit->tree helper patch
Linus Torvalds [Thu, 21 Apr 2005 01:49:41 +0000 (18:49 -0700)]
Fix up some problems from the commit->tree helper patch

19 years agoFix up git-export to use the lower-level interfaces for diff generation.
Linus Torvalds [Thu, 21 Apr 2005 01:48:27 +0000 (18:48 -0700)]
Fix up git-export to use the lower-level interfaces for diff generation.

The high-level helpers seem to have forgotten what to do with sha1
names.

19 years ago[PATCH] Teach read-tree about commit objects
Junio C Hamano [Thu, 21 Apr 2005 01:06:50 +0000 (18:06 -0700)]
[PATCH] Teach read-tree about commit objects

Updates read-tree to use read_tree_with_tree_or_commit_sha1()
function.  The command can take either tree or commit IDs with
this patch.

The change involves a slight modification of how it recurses down
the tree.  Earlier the caller only supplied SHA1 and the recurser
read the object using it, but now it is the caller's responsibility
to read the object and give it to the recurser.  This matches the
way recursive behaviour is done in other tree- related commands.

Signed-off-by: Junio C Hamano <junkio@cox.net>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] Teach ls-tree about commit objects
Junio C Hamano [Thu, 21 Apr 2005 01:06:50 +0000 (18:06 -0700)]
[PATCH] Teach ls-tree about commit objects

Updates ls-tree.c to use read_tree_with_tree_or_commit_sha1()
function.  The command can take either tree or commit IDs with
this patch.

Signed-off-by: Junio C Hamano <junkio@cox.net>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] Teach diff-tree about commit objects
Junio C Hamano [Thu, 21 Apr 2005 01:06:50 +0000 (18:06 -0700)]
[PATCH] Teach diff-tree about commit objects

Updates diff-tree.c to use read_tree_with_tree_or_commit_sha1()
function.  The command can take either tree or commit IDs with this patch.

Signed-off-by: Junio C Hamano <junkio@cox.net>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] Teach diff-cache about commit objects
Junio C Hamano [Thu, 21 Apr 2005 01:06:49 +0000 (18:06 -0700)]
[PATCH] Teach diff-cache about commit objects

Updates diff-cache.c to use read_tree_with_tree_or_commit_sha1()
function.  The end-user visible result is the same --- the command
takes either tree or commit ID.

Signed-off-by: Junio C Hamano <junkio@cox.net>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] Accept commit in some places when tree is needed.
Junio C Hamano [Thu, 21 Apr 2005 01:06:49 +0000 (18:06 -0700)]
[PATCH] Accept commit in some places when tree is needed.

This patch implements read_tree_with_tree_or_commit_sha1(),
which can be used when you are interested in reading an unpacked
raw tree data but you do not know nor care if the SHA1 you
obtained your user is a tree ID or a commit ID.  Before this
function's introduction, you would have called read_sha1_file(),
examined its type, parsed it to call read_sha1_file() again if
it is a commit, and verified that the resulting object is a
tree.  Instead, this function does that for you.  It returns
NULL if the given SHA1 is not either a tree or a commit.

Signed-off-by: Junio C Hamano <junkio@cox.net>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] simplify Makefile
Andre Noll [Wed, 20 Apr 2005 20:10:13 +0000 (13:10 -0700)]
[PATCH] simplify Makefile

Use a generic rule for executables that depend only on the corresponding
.o and on $(LIB_FILE).

Signed-Off-By: Andre Noll <maan@systemlinux.org>
Signed-Off-By: Linus Torvalds <torvalds@osdl.org>
19 years agoImprove build: add <unistd.h> and use -O2 instead of -O3
Linus Torvalds [Wed, 20 Apr 2005 20:00:08 +0000 (13:00 -0700)]
Improve build: add <unistd.h> and use -O2 instead of -O3

(Nobody should use -O3. It just makes bad inlining decisions).

19 years agoMake the sha1 of the index file go at the very end of the file.
Linus Torvalds [Wed, 20 Apr 2005 19:36:41 +0000 (12:36 -0700)]
Make the sha1 of the index file go at the very end of the file.

This allows us to both calculate it and verify it faster.

19 years agoSpeed up index file writing by chunking it nicely.
Linus Torvalds [Wed, 20 Apr 2005 19:16:57 +0000 (12:16 -0700)]
Speed up index file writing by chunking it nicely.

No point in making 17,000 small writes when you can make just
a couple of hundred nice 8kB writes instead and save a lot
of time.

19 years agoMake "write_sha1_file()" exit early if the file already exists.
Linus Torvalds [Wed, 20 Apr 2005 16:28:05 +0000 (09:28 -0700)]
Make "write_sha1_file()" exit early if the file already exists.

Avoid the compression.

19 years agoThe recent hash/compression switch-over missed the blob creation.
Linus Torvalds [Wed, 20 Apr 2005 08:34:54 +0000 (01:34 -0700)]
The recent hash/compression switch-over missed the blob creation.

Happily, convert-cache just magically fixes all errors.

19 years agoDo SHA1 hash _before_ compression. conversion
Linus Torvalds [Wed, 20 Apr 2005 08:10:46 +0000 (01:10 -0700)]
Do SHA1 hash _before_ compression.

And add a "convert-cache" program to convert from old-style
to new-style.

19 years ago[PATCH] init-db.c: create and use safe_create_dir helper
Zach Welch [Wed, 20 Apr 2005 04:48:15 +0000 (21:48 -0700)]
[PATCH] init-db.c: create and use safe_create_dir helper

Factor mkdir calls into common safe_create_dir subroutine.

Signed-Off-By: Zach Welch <zw@superlucidity.net>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] init-db.c: normalize env var handling.
Zach Welch [Wed, 20 Apr 2005 04:48:15 +0000 (21:48 -0700)]
[PATCH] init-db.c: normalize env var handling.

Normalize init-db environment variable handling, allowing the creation
of object directories with something other than DEFAULT_DB_ENVIRONMENT.

Signed-Off-By: Zach Welch <zw@superlucidity.net>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] init-db.c: cleanup comments
Zach Welch [Wed, 20 Apr 2005 04:48:15 +0000 (21:48 -0700)]
[PATCH] init-db.c: cleanup comments

Consolidate comments at top of main.

Signed-Off-By: Zach Welch <zw@superlucidity.net>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years agoMake "diff-tree" take commit objects too, like "diff-cache" does.
Linus Torvalds [Wed, 20 Apr 2005 04:39:28 +0000 (21:39 -0700)]
Make "diff-tree" take commit objects too, like "diff-cache" does.

Sometimes it's just easier to not have to look up the "commit"->"tree"
translation by hand first. It's trivial to do inside diff-tree, and
it's just being polite.

19 years agoAdd "diff-cache" helper program to compare a tree (or commit) with
Linus Torvalds [Wed, 20 Apr 2005 04:00:09 +0000 (21:00 -0700)]
Add "diff-cache" helper program to compare a tree (or commit) with
the current cache state and/or working directory.

Very useful to see what has changed since the last commit, either in
the index file or in the whole working directory.

Also very possibly very buggy. Matching the two up is not entirely
trivial.

19 years agoMake us be better at guessing a good hostname for the email.
Linus Torvalds [Tue, 19 Apr 2005 22:46:06 +0000 (15:46 -0700)]
Make us be better at guessing a good hostname for the email.

It's still just a guess, and the result is not a real email
address anyway. If you want to, you can use COMMIT_AUTHOR_EMAIL
to correct for any git guesses.

19 years agoRemove duplicate getenv(DB_ENVIRONMENT) call
Linus Torvalds [Tue, 19 Apr 2005 21:11:06 +0000 (14:11 -0700)]
Remove duplicate getenv(DB_ENVIRONMENT) call

Noted by Tony Luck.

19 years agoAdd stupid "git export" thing, which can export a git archive
Linus Torvalds [Tue, 19 Apr 2005 21:00:34 +0000 (14:00 -0700)]
Add stupid "git export" thing, which can export a git archive
as a set of patches and commentary.

You'd want something like this if you are tracking a git archive
in another SCM format. Notably, we want something like that for
BK users.

19 years agoFix init-db shared database case
Linus Torvalds [Tue, 19 Apr 2005 20:59:17 +0000 (13:59 -0700)]
Fix init-db shared database case

Noted by Aaron Straus

19 years agoUpdate "git-pull-script" to use "read-tree -m" for
Linus Torvalds [Tue, 19 Apr 2005 19:56:47 +0000 (12:56 -0700)]
Update "git-pull-script" to use "read-tree -m" for
reading a single tree too. That should speed up a
trivial merge noticeably.

Also, don't bother reading back the tree we just wrote
when we committed a real merge. It had better be the
same one we still have..

19 years agoMake "read-tree" know how to do a "1-way merge".
Linus Torvalds [Tue, 19 Apr 2005 18:41:18 +0000 (11:41 -0700)]
Make "read-tree" know how to do a "1-way merge".

This one just reads one tree, but picks up any matching stat information
from the old index.

19 years agoMake "read-tree" take the 'stat' information for a merge result from the
Linus Torvalds [Tue, 19 Apr 2005 18:16:12 +0000 (11:16 -0700)]
Make "read-tree" take the 'stat' information for a merge result from the
old index state if the result matches.

This leaves the stat information in the result tree for any trivial
merges, which is just the way we like it.

19 years agoMake git-pull-script do the right thing for symlinked HEAD's.
Linus Torvalds [Tue, 19 Apr 2005 16:53:58 +0000 (09:53 -0700)]
Make git-pull-script do the right thing for symlinked HEAD's.

Also exit gracefully if the HEAD pull failed, rather than use
a possibly stale MERGE_HEAD.

19 years agoDon't parse commit objects more than once.
Linus Torvalds [Tue, 19 Apr 2005 16:52:24 +0000 (09:52 -0700)]
Don't parse commit objects more than once.

Yes, the "parse_commit()" already checks for this condition, but
we need to check for it in rev-tree too, so that we don't start
walking the parent chain unnecessarily.

19 years ago[PATCH] SCSI trees, merges and git status
James Bottomley [Tue, 19 Apr 2005 02:55:19 +0000 (19:55 -0700)]
[PATCH] SCSI trees, merges and git status

Doing the latest SCSI merge exposed two bugs in your merge script:

1) It doesn't like a completely new directory (the misc tree contains a
   new drivers/scsi/lpfc)
2) the merge testing logic is wrong.  You only want to exit 1 if the
   merge fails.

19 years ago[PATCH] provide better committer information to commit-tree.c
Greg KH [Tue, 19 Apr 2005 00:37:32 +0000 (17:37 -0700)]
[PATCH] provide better committer information to commit-tree.c

Here's a small patch to commit-tree.c that does two things:
- allows the committer email address and name to be overridden
  by environment variables (if you don't like the environment
  variable names I've used (COMMIT_AUTHOR_NAME,
  COMMIT_AUTHOR_EMAIL), feel free to change them.)
- provide the proper domainname to the author/committer email
  address (otherwise, my address was only showing up as from the
  hostname.)

This allows people to set sane values for the commit names and email
addresses, preventing odd, private hostnames and domains from being
exposed to the world.

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years agoMake fsck-cache print the object type for unreachable objects.
Linus Torvalds [Tue, 19 Apr 2005 00:35:31 +0000 (17:35 -0700)]
Make fsck-cache print the object type for unreachable objects.

This got lost when I updated to Daniel's new object model.

19 years ago[PATCH] show-diff: Remove stale comments
Junio C Hamano [Mon, 18 Apr 2005 23:51:09 +0000 (16:51 -0700)]
[PATCH] show-diff: Remove stale comments

Patch 1/6 in the series has already cleaned the interface to
call sq_expand(), but the comment before that function still
carries the stale interface warning.  Remove it.

Signed-off-by: Junio C Hamano <junkio@cox.net>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] Do not let rsync obliterate .git/object symbolic link.
Junio C Hamano [Mon, 18 Apr 2005 23:49:21 +0000 (16:49 -0700)]
[PATCH] Do not let rsync obliterate .git/object symbolic link.

Signed-off-by: Junio C Hamano <junkio@cox.net>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] show-diff.c: -R option for reverse diff.
Junio C Hamano [Mon, 18 Apr 2005 22:07:24 +0000 (15:07 -0700)]
[PATCH] show-diff.c: -R option for reverse diff.

This adds -R option to obtain reverse diff.

It may be useful in the merge workflow.  After the base of the working
directory is merged and commited, in the working directory:

    $ read-tree <tree-id-of-merged-tree>
    $ show-diff -R

to re-validate if upstream changes make sense, and/or revert or
conflict with local changes you have in the working files.

Signed-off-by: Junio C Hamano <junkio@cox.net>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] show-diff.c: make diff options customizable.
Junio C Hamano [Mon, 18 Apr 2005 22:07:24 +0000 (15:07 -0700)]
[PATCH] show-diff.c: make diff options customizable.

This makes the diff output formatting options customizable via the
environment variables.  The default is still the Linux kernel style.

Signed-off-by: Junio C Hamano <junkio@cox.net>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] show-diff.c: adjust default format for the Linux kernel.
Junio C Hamano [Mon, 18 Apr 2005 22:07:24 +0000 (15:07 -0700)]
[PATCH] show-diff.c: adjust default format for the Linux kernel.

This patch adjusts the default output format of show-diff to match
the Linux kernel style, recommended in Documentation/SubmittingPatches.

Signed-off-by: Junio C Hamano <junkio@cox.net>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] show-diff.c: simplify show_diff_empty.
Junio C Hamano [Mon, 18 Apr 2005 22:07:24 +0000 (15:07 -0700)]
[PATCH] show-diff.c: simplify show_diff_empty.

This patch removes the custom diff generation code from the
show_diff_empty() function.  Instead, just use show_differences().

This reduces the code size; but more importantly, it is needed for
the later patch to give diff options.

Signed-off-by: Junio C Hamano <junkio@cox.net>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] show-diff.c: check unreadbale blob.
Junio C Hamano [Mon, 18 Apr 2005 22:07:24 +0000 (15:07 -0700)]
[PATCH] show-diff.c: check unreadbale blob.

This patch fixes show-diff to detect unreadable blob and warn
instead of going ahead and crashing.

Signed-off-by: Junio C Hamano <junkio@cox.net>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] show-diff.c: clean up private buffer use.
Junio C Hamano [Mon, 18 Apr 2005 22:07:24 +0000 (15:07 -0700)]
[PATCH] show-diff.c: clean up private buffer use.

This patch fixes sq_expand() and show_differences() not to use and
hold onto its privately allocated buffer, which was a misguided
attempt to reduce calls to malloc but made later changes harder.

Signed-off-by: Junio C Hamano <junkio@cox.net>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years agoAdd "update-cache --refresh" to git-pull-script to make sure
Linus Torvalds [Mon, 18 Apr 2005 22:01:48 +0000 (15:01 -0700)]
Add "update-cache --refresh" to git-pull-script to make sure
out index is all ready to go after a pull.

Noted by Russell King

19 years agoChange merge-cache and git-merge-one-file to use the SHA1 of the file
Linus Torvalds [Mon, 18 Apr 2005 21:17:58 +0000 (14:17 -0700)]
Change merge-cache and git-merge-one-file to use the SHA1 of the file
instead of a checked-out temporary copy.

If merging requires a checked-out-copy, we now do so with "unpack-file".

19 years agoAdd "unpack-file" helper that unpacks a sha1 blob into a tmpfile.
Linus Torvalds [Mon, 18 Apr 2005 21:11:01 +0000 (14:11 -0700)]
Add "unpack-file" helper that unpacks a sha1 blob into a tmpfile.

19 years agoAdd more header dependencies.
Linus Torvalds [Mon, 18 Apr 2005 20:12:21 +0000 (13:12 -0700)]
Add more header dependencies.

Yeah, my Makefiles are always a total disaster. Better this than autotools
or some horror like that, though.

19 years agoSplit up read-cache.c into more logical clumps.
Linus Torvalds [Mon, 18 Apr 2005 20:04:43 +0000 (13:04 -0700)]
Split up read-cache.c into more logical clumps.

Do the usage and error reporting in "usage.c", and the sha1 file
accesses in "sha1_file.c".

Small, nice, easily separated parts. Good.

19 years agoClean up the Makefile a bit.
Linus Torvalds [Mon, 18 Apr 2005 19:49:39 +0000 (12:49 -0700)]
Clean up the Makefile a bit.

This introduces the concept of git "library" objects that
the real programs use, and makes it easier to add such things
to a "libgit.a".

This will also make it trivial to split the current "read-cache.o"
into more aptly named pieces (it does a lot more than just read
the index file).

19 years agoAdd the simple scripts I used to do a merge with content conflicts.
Linus Torvalds [Mon, 18 Apr 2005 19:15:10 +0000 (12:15 -0700)]
Add the simple scripts I used to do a merge with content conflicts.

They sure as hell aren't perfect, but they allow you to do:

./git-pull-script {other-git-directory}

to do the initial merge, and if that had content clashes, you do

merge-cache ./git-merge-one-file-script -a

which tries to auto-merge. When/if the auto-merge fails, it will
leave the last file in your working directory, and you can edit
it and then when you're happy you can do "update-cache filename"
on it. Re-do the merge-cache thing until there are no files left
to be merged, and now you can write the tree and commit:

write-tree
commit-tree .... -p $(cat .git/HEAD) -p $(cat .git/MERGE_HEAD)

and you're done.

19 years agoMerge the new object model thing from Daniel Barkalow
Linus Torvalds [Mon, 18 Apr 2005 19:12:00 +0000 (12:12 -0700)]
Merge the new object model thing from Daniel Barkalow

This was a real git merge with conflicts. I'll commit the scripts I used
to do the merge next.

Not pretty, but it's half-way functional.

19 years ago[PATCH] fix bug in read-cache.c which loses files when merging a tree
James Bottomley [Mon, 18 Apr 2005 18:44:31 +0000 (11:44 -0700)]
[PATCH] fix bug in read-cache.c which loses files when merging a tree

I noticed this when I tried a non-trivial scsi merge and checked the
results against BK.  The problem is that remove_entry_at() actually
decrements active_nr, so decrementing it in add_cache_entry() before
calling remove_entry_at() is a double decrement (hence we lose cache
entries at the end).

19 years ago[PATCH] Switch implementations of merge-base, port to parsing
Daniel Barkalow [Mon, 18 Apr 2005 18:39:48 +0000 (11:39 -0700)]
[PATCH] Switch implementations of merge-base, port to parsing

This switches to my implementation of merge-base, but with the new parsing
library.

Signed-Off-By: Daniel Barkalow <barkalow@iabervon.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] Port fsck-cache to use parsing functions
Daniel Barkalow [Mon, 18 Apr 2005 18:39:48 +0000 (11:39 -0700)]
[PATCH] Port fsck-cache to use parsing functions

This ports fsck-cache to use parsing functions. Note that performance
could be improved here by only reading each object once, but this requires
somewhat more complicated flow control.

Signed-Off-By: Daniel Barkalow <barkalow@iabervon.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] Port rev-tree to parsing functions
Daniel Barkalow [Mon, 18 Apr 2005 18:39:48 +0000 (11:39 -0700)]
[PATCH] Port rev-tree to parsing functions

This ports rev-tree to use the parsing functions introduced in the
previous patches.

Signed-Off-By: Daniel Barkalow <barkalow@iabervon.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] Implementations of parsing functions
Daniel Barkalow [Mon, 18 Apr 2005 18:39:48 +0000 (11:39 -0700)]
[PATCH] Implementations of parsing functions

This implements the parsing functions.

Signed-Off-By: Daniel Barkalow <barkalow@iabervon.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] Header files for object parsing
Daniel Barkalow [Mon, 18 Apr 2005 18:39:48 +0000 (11:39 -0700)]
[PATCH] Header files for object parsing

This adds the structs and function declarations for parsing git objects.

Signed-Off-By: Daniel Barkalow <barkalow@iabervon.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] Fix confusing behaviour of update-cache --refresh on unmerged paths.
Junio C Hamano [Mon, 18 Apr 2005 17:42:48 +0000 (10:42 -0700)]
[PATCH] Fix confusing behaviour of update-cache --refresh on unmerged paths.

The "update-cache --refresh" command attempts refresh_entry()
on unmerged path, which results in as many "needs update" messages
as there are unmerged stages for that path.  This does not do
any harm to the working directory, but it is confusing.

Here is a fix.

Signed-off-by: Junio C Hamano <junkio@cox.net>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years agoUpdate README to reflect the hierarchical tree objects,
Linus Torvalds [Mon, 18 Apr 2005 04:52:23 +0000 (21:52 -0700)]
Update README to reflect the hierarchical tree objects,
and other newfangled things like merging.

Also, talk more about the actual operations, and give some
rough examples of what you can do.

19 years ago[PATCH] (resend) show-diff.c off-by-one fix
Junio C Hamano [Mon, 18 Apr 2005 03:31:40 +0000 (20:31 -0700)]
[PATCH] (resend) show-diff.c off-by-one fix

The patch to introduce shell safety to show-diff has an
off-by-one error.  Here is an fix.

Signed-off-by: Junio C Hamano <junkio@cox.net>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years agoAss a "merge-cache" helper program to execute a merge on
Linus Torvalds [Mon, 18 Apr 2005 02:52:54 +0000 (19:52 -0700)]
Ass a "merge-cache" helper program to execute a merge on
any unmerged files.

This one doesn't actually do the merging, but it makes it
easy to script the programs that do using it.

19 years ago[PATCH] fork optional branch point normazilation
Brad Roberts [Mon, 18 Apr 2005 02:12:14 +0000 (19:12 -0700)]
[PATCH] fork optional branch point normazilation

Fix remove_specials for real.  The second half logic needs the original
head of the string.

Signed-off-by: Brad Roberts <braddr@puremagic.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years agoIgnore any unmerged entries for "checkout-cache -a".
Linus Torvalds [Mon, 18 Apr 2005 01:39:14 +0000 (18:39 -0700)]
Ignore any unmerged entries for "checkout-cache -a".

They don't "exist" yet, and you need to merge them first.

19 years agoRemove extraneous ',' ';' and '.' characters from the full name gecos field.
Linus Torvalds [Sun, 17 Apr 2005 23:52:54 +0000 (16:52 -0700)]
Remove extraneous ',' ';' and '.' characters from the full name gecos field.

Apparently some distributions tend to have space for phone numbers etc there.

19 years agoMake the revision tracking track the object types too.
Linus Torvalds [Sun, 17 Apr 2005 23:19:32 +0000 (16:19 -0700)]
Make the revision tracking track the object types too.

This allows fsck to verify not just that an object exists, but
also that it has the type it was expected to have.

19 years agoMake "commit-tree" check the input objects more carefully.
Linus Torvalds [Sun, 17 Apr 2005 22:26:13 +0000 (15:26 -0700)]
Make "commit-tree" check the input objects more carefully.

Let's not allow trivially bogus commits. I did one for the first
trial of the first kernel git merge. fsck found it ok, but..

19 years agoMake "parse_commit" return the "struct revision" for the commit.
Linus Torvalds [Sun, 17 Apr 2005 19:40:18 +0000 (12:40 -0700)]
Make "parse_commit" return the "struct revision" for the commit.

Also, make it a fatal error to pass in a non-commit object. The callers
never checked, so better check here.

This simplifies merge-base further. It's now so trivial that it's almost
ridiculous.

19 years agoDo a very simple "merge-base" that finds the most recent common
Linus Torvalds [Sun, 17 Apr 2005 19:18:17 +0000 (12:18 -0700)]
Do a very simple "merge-base" that finds the most recent common
parent of two commits.

The question of "best" commit can probably be tweaked almost arbitrarily.
In particular, trying to take things like how big the tree differences
are into account migt be a good idea. This one is just very simple.

19 years agoMake "rev-tree.c" use the new-and-improved "mark_reachable()"
Linus Torvalds [Sun, 17 Apr 2005 19:09:20 +0000 (12:09 -0700)]
Make "rev-tree.c" use the new-and-improved "mark_reachable()"

It used to have its own specialized version for marking the
sub-reachability bits.

19 years agoMake "revision.h" slightly better to use.
Linus Torvalds [Sun, 17 Apr 2005 19:07:00 +0000 (12:07 -0700)]
Make "revision.h" slightly better to use.

 - mark_reachable() can be more generic, marking the reachable revisions
   with an arbitrary mask.
 - date parsing will parse to a date of 0 rather than ULONG_MAX for the
   bad old case, sorting the dates correctly.

19 years agoMove "parse_commit()" into common revision.h file.
Linus Torvalds [Sun, 17 Apr 2005 18:39:18 +0000 (11:39 -0700)]
Move "parse_commit()" into common revision.h file.

This also drops the old-style date parsing. We just don't care
enough, since we dropped that format pretty early.

Yes, this could do with some cleanup, and a common library file.
Some day.

19 years ago[PATCH] fix for memory leak in write-tree.c
Brad Roberts [Sun, 17 Apr 2005 17:55:12 +0000 (10:55 -0700)]
[PATCH] fix for memory leak in write-tree.c

Fix a memory leak in write-tree.c, not freeing the directory buffer.

19 years ago[PATCH] Fix +x-related show-diff false positives
Petr Baudis [Sun, 17 Apr 2005 17:04:48 +0000 (10:04 -0700)]
[PATCH] Fix +x-related show-diff false positives

This fixes show-diff listing all +x files as differring.

Signed-off-by: Petr Baudis <pasky@ucw.cz>
[ That's what I get for working on a G5 - my testing was all
  big-endian in the first place.  -- Linus ]

Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] Remove unused arguments from index_fd()
Junio C Hamano [Sun, 17 Apr 2005 17:02:22 +0000 (10:02 -0700)]
[PATCH] Remove unused arguments from index_fd()

The function index_fd() in update-cache.c takes 5 arguments, but
two is not necessary and one that is a pointer to a structure
really needs to be a pointer to one member of that structure.
This patch cleans it up.

Also it removes printf() apparently left after initial
debugging.

Signed-off-by: Junio C Hamano <junkio@cox.net>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] Better error message from checkout-cache for unmerged files.
Junio C Hamano [Sun, 17 Apr 2005 17:02:21 +0000 (10:02 -0700)]
[PATCH] Better error message from checkout-cache for unmerged files.

The checkout-cache command says "file is not in the cache" when
an unmerged path is given.  This patch adds code to distinguish
the unmerged and the nonexistent cases and gives an appropriate
error message.

Signed-off-by: Junio C Hamano <junkio@cox.net>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years agoFix total permission bogosity in "checkout-cache.c".
Linus Torvalds [Sun, 17 Apr 2005 16:55:36 +0000 (09:55 -0700)]
Fix total permission bogosity in "checkout-cache.c".

Use the proper octal mode naming instead of random decimal
crud, and don't reset the mode after the create with fchmod:
the whole point was to let "umask" do its thing.

Duh.

19 years ago[PATCH] update-cache --remove marks the path merged.
Junio C Hamano [Sun, 17 Apr 2005 16:53:35 +0000 (09:53 -0700)]
[PATCH] update-cache --remove marks the path merged.

When update-cache --remove is run, resolve unmerged state for
the path.  This is consistent with the update-cache --add
behaviour.  Essentially, the user is telling us how he wants to
resolve the merge by running update-cache.

Signed-off-by: Junio C Hamano <junkio@cox.net>
Fixed to do the right thing at the end.

Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] show-diff.c: do not include unused header file
Junio C Hamano [Sun, 17 Apr 2005 16:53:35 +0000 (09:53 -0700)]
[PATCH] show-diff.c: do not include unused header file

This is my bad.  I added #include <ctype.h> to the file,
which I ended up not using and failed to remove it.

Signed-off-by: Junio C Hamano <junkio@cox.net>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years agoBe much more liberal about the file mode bits.
Linus Torvalds [Sun, 17 Apr 2005 05:26:31 +0000 (22:26 -0700)]
Be much more liberal about the file mode bits.

We only really care about the difference between a file being executable
or not (by its owner). Everything else we leave for the user umask to
decide.

19 years ago[PATCH] Do not run useless show-diff on unmerged paths repeatedly.
Junio C Hamano [Sun, 17 Apr 2005 04:29:45 +0000 (21:29 -0700)]
[PATCH] Do not run useless show-diff on unmerged paths repeatedly.

When run on unmerged dircache, show-diff compares the working
file with each non-empty stage for that path.  Two out of three
times, this is not very helpful.  This patch makes it report the
unmergedness only once per each path and avoids running the
actual diff.

Upper layer SCMs like Cogito are expected to find out mode/SHA1
for each stage by using "show-files --stage" and run the diff
itself.  This would result in more sensible diffs.

Signed-off-by: Junio C Hamano <junkio@cox.net>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>