[PATCH] git-format-patch-script and mailinfo updates.
- avoid duplicating [PATCH] in the commit message body if the
original commit has it already (happens for commits done from
mails via applymbox).
- check if the commit author is different from the one who is
running the script, and emit an appropriate "From:" and
"Date: " lines to the output.
- with '--date', emit "Date: " line to preserve the original
author date even for the user's own commit.
- teach mailinfo to grok not just "From: " but "Date: ".
The patch e-mail output by format-patch starts with the first
line from the original commit message, prefixed with [PATCH],
and optionally a From: line if you are reformatting a patch
obtained from somebody else, a Date: line from the original
commit if (1) --date is specified or (2) for somebody else's
patch, and the rest of the commit message body.
Expected use of this is to move the title line from the commit
to Subject: when sending it via an e-mail, and leave the From:
and the Date: lines as the first lines of your message.
The mailinfo command has been changed to read Date: (in addition
to From: it already understands) and do sensible things when
running applymbox.
Signed-off-by: Junio C Hamano <junkio@cox.net> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
[PATCH] tools/applymbox: allow manual fixing and continuing after a failure.
With "-c .dotest/0002" flag, the applymbox command can be told
to use existing .dotest/patch file after hand-fixing the patch
conflicts for the second patch in the mailbox, and continue on
from that message, skipping the already applied first patch in
.dotest/0001.
Signed-off-by: Junio C Hamano <junkio@cox.net> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Adds --exclude=pattern option to the "git-apply" command. This
was useful while reimporting the BKCVS patchset dump of the
Linux kernel, starting at 2.4.0 and ending at 2.6.12-rc2 Ingo
announced some time ago to exclude BitKeeper directory.
Signed-off-by: Junio C Hamano <junkio@cox.net> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
The parsing code had a bug that failed to recognize an
incomplete line at the end of a fragment, and the fragment
application code had a comparison bug to recognize such. Fix
them to handle incomplete lines correctly.
Add a test script for patches with various combinations of
complete and incomplete lines to make sure the fix works.
Signed-off-by: Junio C Hamano <junkio@cox.net> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
git-send-pack: verify that sender is a proper superset of receiver
This should make sure that if you have multiple people pushing to the
same tree, they cannot overwrite each others work, but have to merge
before being able to update the common reference tree.
We display the files where the child differs from the first parent
first in black in the file list window, followed by the second parent
in blue, and so on using different colors for each parent (provided
you don't have more than 6 parents; if you do we cycle around).
Make searching in files changed faster, and fix some bugs.
We now kick off a single git-diff-tree -r --stdin and feed it all
the commit pairs we want to know about, instead of doing a separate
git-diff-tree invocation for each.
Merge git-tools repository under "tools" subdirectory
Thanks to Ryan Anderson for setting me up to do this. I'd have used his
work, but I wanted to clean up the old git-tools repository before
merging it: it had old-style file modes etc that needed a round of
git-convert-cache to fix up.
Prepare git-tools for merging into the main git archive
Rename into a "tools" subdirectory, and change name of "dotest" to "applymbox".
Remove stripspace (which was already copied into git) and cvs2git (which
was likewise already copied into git, and then replaced by a much better
perl version).
All of this was brought on by Ryan Anderson shaming me into it. Thanks.
I guess.
[PATCH] fetch/pull: short-hand notation for remote repositories.
Since pull and fetch are done often against the same remote
repository repeatedly, keeping the URL to pull from along with
the name of the head to use in $GIT_DIR/branches/$name makes a
lot of sense. Adopt that convention from Cogito, and try to be
compatible when possible; storing a partial URL and completing
it with a trailing path may not be understood by Cogito.
While we are at it, fix pulling a tag. Earlier, we updated only
refs/tags/$tag without updating FETCH_HEAD, and called
resolve-script using a stale (or absent) FETCH_HEAD.
Signed-off-by: Junio C Hamano <junkio@cox.net> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
"Files" matches the find string against each of the files modified
by each commit, and can do exact, case-ignoring or regexp matching.
"Pickaxe" uses git-diff-tree -S'string' and can only do exact
matching. I called it "pickaxe" rather than "find within patch"
since it only finds commits where the string is present in the child
but not the parents or vice versa, and "pickaxe" is what the author
of that feature calls it.
Be slightly smarter about git-daemon client shutdown
Shut down connections that haven't even identified themselves as git
clients first. That should get rid of people who just connect to the
port and wait for something to happen.
The cvsimport example in the cvs migration document was still
using the old syntax for target repository after new and
improved cvsimport-script was merged.
Signed-off-by: Junio C Hamano <junkio@cox.net> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Now, not all projects can be as refined as Linux. Before the final 1.0
release, we went through fifteen 0.99 patchfiles, and pl14 alone went
through subreleases 'a' through 'z'. Now _that_ is a release process.
Not to mention the odd-ball releases, like 0.96c+
Sadly, in this day and age of RPM's etc, we have silly limitations, and
I cannot call this release '0.99pl5a or some such awe-inspiring name
just because "rpmbuild" is such a party pooper. So it's just 0.99.1.
Oh well. Aspiring to such greatness as the Linux release numbering is
hubris anyway. You can attain such perfection only once in your life.
And finally what all of this has been leading up to.
The 2 line code change to record who made a tag,
and the 8 line code change to check that we recorded
the tag.
Gosh the error checking is always so much bigger than the code :)
[PATCH] Update git-tag-script to create the .git/refs/tags if it does not already exist
When testing tags I ran into an interesting problem.
git-tag-script dies if .git/refs/tags/ does not exist.
And that directory didn't get created when I build my repository,
so we need to create it if it doesn't exist.
Signed-of-by: Eric W. Biederman <ebiederm@xmission.com> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
[PATCH] Add git-var a tool for reading interesting git variables.
Sharing code between shell scripts and C is a challenge. The program
git-var allows us to have a set of named values that a shell script can
interrogate and a normal C program can simply call the functions that
compute them. Allowing sharing when computing plain test values.
git-daemon: re-organize code a bit for --inetd flag
Alexey Nezhdanov sent a patch that made git-daemon usable from inetd (ie
where inetd has already done the accept on the new connection, the fork,
and the setup of stdin/stdout). I wanted to organize the thing slightly
differently, though.
Split up "diff_format" into "format" and "line_termination".
This removes the separate "formats" for name and name-with-zero-
termination.
It also removes the difference between HUMAN and MACHINE formats, and
they both become DIFF_FORMAT_RAW, with the difference being just in the
line and inter-filename termination.
... and make git-diff-files use it too. This all _should_ make the
diffcore-pathspec.c phase unnecessary, since the diff'ers now all do the
path matching early interally.
Start using the partial tree reading in "git-diff-cache"
The reason I say "start using" is that we really should also limit the
index checking by name - now we limit the tree object accesses by name,
but we still check the whole index.
Make "git prune" use the "--full" flag to git-fsck-cache
It's too dangerous not to. We need to follow alternate object
directories etc, or we might say something is unreachable just because
we didn't look it up completely.
[PATCH] Documentation: push-pull commands into a separate category.
This splits push-pull related commands into a separate
category. I think a bigger overhaul of the main index is
needed, but have not got around to it. Help is welcome.
Signed-off-by: Junio C Hamano <junkio@cox.net> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Oh, well.. FC4 has UTF-8 as the default environment, and I applaud
that, but then it sometimes results in these characters that aren't
actually visible as a problem.
[PATCH] clone-pack and clone-script: documentation and add a missing parameter.
While adding the documentation for these two commands, I noticed
that the name of the program on the other end (git-upload-pack)
is already almost configurable but git-clone-pack lacked command
line parameter parsing to actually use anything but default, so
I introduced --exec= like other remote commands while I was at it.
Signed-off-by: Junio C Hamano <junkio@cox.net> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
[PATCH] cat-file: be consistent in usage string and documentation.
Now that we have something called tag object, and a notion of
"tags" stored in .git/refs/tags/ directory, the word "tagname"
has become misleading in the usage string. The documentation
already calls that <type>.
Signed-off-by: Junio C Hamano <junkio@cox.net> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Useful for pulling stuff off a dedicated server. Instead of connecting
with ssh or just starting a local pipeline, we connect over TCP to the
other side and try to see if there's a git server listening.
Of course, since I haven't written the git server yet, that will never
happen. But the server really just needs to listen on a port, and
execute a "git-upload-pack" when somebody connects.
(It should read one packet-line, which should be of the format
"git-upload-pack directoryname\n"
and eventually we migth have other commands the server might accept).
git-apply: be a lot more careful when writing files
We write them under another name and rename them to their destination,
so that if something bad happens in the middle, we won't have caused any
bigger harm.
Also, this makes the writing be NFS "intr" safe, and as a side effects
makes sure that if the target is hardlinked (or symlinked) we will have
broken the link.
[PATCH] git-diff-*: --name-only and --name-only-z.
Porcelain layers often want to find only names of changed files,
and even with diff-raw output format they end up having to pick
out only the filename. Support --name-only (and --name-only-z
for xargs -0 and cpio -0 users that want to treat filenames with
embedded newlines sanely) flag to help them.
Signed-off-by: Junio C Hamano <junkio@cox.net> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
That way we avoid any confusion with "GNU Interactive Tools", and it's
more descriptive anyway (the rpm documentation talks about how git is
split into a "core" part and an "SCM" part, this makes it clear that
this is the core one).
and it creates a new branch of name <branchname>. If a starting point
is specified, that will be where the branch is created, otherwise it
will be created at the current HEAD.
Make "git checkout" verify that the argument refers to a commit
We still need to create a new branch if it didn't refer to an existing
branch, otherwise our HEAD will continue to point to something totally
different than what we just checked out.
I'll need to think about it. Maybe only do it with "-f" and force it to
the "master" branch?
git-rev-parse: Allow a "zeroth" parent of a commit - the commit itself.
This sounds nonsensical, but it's useful to make sure that the result is
a commit.
For example, "git-rev-parse v2.6.12" will return the _tag_ object for
v2.6.12, but "git-rev-parse v2.6.12^0" will return the _commit_ object
associated with that tag (and v2.6.12^1 will return the first parent).
Also, since the "parent" code will actually parse the commit, this,
together with the "--verify" flag, will verify not only that the result
is a single SHA1, but will also have verified that it's a proper commit
that we can see.
Cut-and-paste dup noticed by Junio. It's not even harmless, since a
match also causes that match to be invalidated, so this made it
impossible to update an existing branch by name.
I'd only tested the case of "ref doesn't exist at all on the other end",
which worked fine.
[PATCH] git-clone-script local optimization tweaks
- When local optimization is used, the variable repo has
already been passed through get_repo_base so there is no need
to check for .git subdirectory in there.
- Use cpio -l instead of "cp -l".
Signed-off-by: Junio C Hamano <junkio@cox.net> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This reverses the order of object lookup, to check pack index first and
then go to the filesystem to find .git/objects/??/ hierarchy.
When most of the objects are packed, this saves quite many stat() calls
and negative dcache entries; while the price this approach has to pay is
negligible, even when most of the objects are outside pack, because
checking pack index file is quite cheap.
Signed-off-by: Junio C Hamano <junkio@cox.net> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
[PATCH] Dereference tag repeatedly until we get a non-tag.
When we allow a tag object in place of a commit object, we only
dereferenced the given tag once, which causes a tag that points at a tag
that points at a commit to be rejected. Instead, dereference tag
repeatedly until we get a non-tag.
This patch makes change to two functions:
- commit.c::lookup_commit_reference() is used by merge-base,
rev-tree and rev-parse to convert user supplied SHA1 to that of
a commit.
- rev-list uses its own get_commit_reference() to do the same.
Dereferencing tags this way helps both of these uses.
Signed-off-by: Junio C Hamano <junkio@cox.net> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
The location alt_odb[j].name[0..] is filled with ??/?{38} to form a sha1
filename to try, but I was too lazy to allocate a copy, so while
fsck_object_dir() is running for the directory, the filenames ??/?{38}
are filled after NUL (usually and always the location should have '/'),
making them "not found".
git-unpack-objects: show progress report by default
This ends up being very calming for big "git clone"s, since otherwise
you just get very frustrated with a long silence, wondering whether it's
working at all.
Use "-q" to quiet it down.
Now if we could just do the same for the initial "figure out what to
pack" phase, which can also be quite slow if the other end is busy (or
not packed and not in cache)...