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>
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>
[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>
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>
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.
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.
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>
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.
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.
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.
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>
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>
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>
[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>
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.
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.
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.
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.
[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.
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>
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>
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:
[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).
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>
[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>
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.
- 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.
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>
[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>
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.
[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.
[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>
This fixes some stylistic problems introduced by my previous set
of patches. I'll be sending my last patch to show-diff next,
which depends on this cleanup.
Signed-off-by: Junio C Hamano <junkio@cox.net> Signed-off-by: Linus Torvalds <torvalds@osdl.org>