The result should be consistent between fetch and push, so we ought to
use the same code in both cases, even though it's short.
Signed-off-by: NDaniel Barkalow <barkalow@iabervon.org>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

And especially do not use // comment.
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

The existing output of "git remote show <remote>" with respect to push
ref specs is basically just to show the raw refspec. This patch teaches
the command to interpret the refspecs and show how each branch will be
pushed to the destination. The output gives the user an idea of what
"git push" should do if it is run w/o any arguments.

Example new output:

1a. Typical output with no push refspec (i.e. matching branches only)

$ git remote show origin
* remote origin
  [...]
  Local refs configured for 'git push':
    master pushes to master (up to date)
    next   pushes to next   (local out of date)

1b. Same as above, w/o querying the remote:

$ git remote show origin -n
* remote origin
  [...]
  Local ref configured for 'git push' (status not queried):
    (matching) pushes to (matching)

2a. With a forcing refspec (+), and a new topic
    (something like push = refs/heads/*:refs/heads/*):

$ git remote show origin
* remote origin
  [...]
  Local refs configured for 'git push':
    master     pushes to master    (fast forwardable)
    new-topic  pushes to new-topic (create)
    next       pushes to next      (local out of date)
    pu         forces to pu        (up to date)

2b. Same as above, w/o querying the remote

$ git remote show origin -n
* remote origin
  [...]
  Local refs configured for 'git push' (status not queried):
    master     pushes to master
    new-topic  pushes to new-topic
    next       pushes to next
    pu         forces to pu

3. With a remote configured as a mirror:

* remote backup
  [...]
  Local refs will be mirrored by 'git push'
Signed-off-by: NJay Soffian <jaysoffian@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

The existing output of "git remote show <remote>" is too verbose for the
information it provides. This patch teaches it to provide more
information in less space.

The output for push refspecs is addressed in the next patch.

Before the patch:

$ git remote show origin
* remote origin
  URL: git://git.kernel.org/pub/scm/git/git.git
  HEAD branch: master
  Remote branch merged with 'git pull' while on branch master
    master
  Remote branch merged with 'git pull' while on branch next
    next
  Remote branches merged with 'git pull' while on branch octopus
    foo bar baz frotz
  New remote branch (next fetch will store in remotes/origin)
    html
  Stale tracking branch (use 'git remote prune')
    bogus
  Tracked remote branches
    maint
    man
    master
    next
    pu
    todo

After this patch:

$ git remote show origin
* remote origin
  URL: git://git.kernel.org/pub/scm/git/git.git
  HEAD branch: master
  Remote branches:
    bogus  stale (use 'git remote prune' to remove)
    html   new (next fetch will store in remotes/origin)
    maint  tracked
    man    tracked
    master tracked
    next   tracked
    pu     tracked
    todo   tracked
  Local branches configured for 'git pull':
    master  rebases onto remote master
    next    rebases onto remote next
    octopus  merges with remote foo
                and with remote bar
                and with remote baz
                and with remote frotz

$ git remote show origin -n
* remote origin
  URL: git://git.kernel.org/pub/scm/git/git.git
  HEAD branch: (not queried)
  Remote branches: (status not queried)
    bogus
    maint
    man
    master
    next
    pu
    todo
  Local branches configured for 'git pull':
    master  rebases onto remote master
    next    rebases onto remote next
    octopus  merges with remote foo
                and with remote bar
                and with remote baz
                and with remote frotz
Signed-off-by: NJay Soffian <jaysoffian@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Our usual method for determining the ref pointed to by HEAD
is to compare HEAD's sha1 to the sha1 of all refs, trying to
find a unique match.

However, some transports actually get to look at HEAD
directly; we should make use of that information when it is
available.  Currently, only http remotes support this
feature.
Signed-off-by: NJeff King <peff@peff.net>
Signed-off-by: NJay Soffian <jaysoffian@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Provide a porcelain command for setting and deleting
$GIT_DIR/remotes/<remote>/HEAD.

While we're at it, document what $GIT_DIR/remotes/<remote>/HEAD is all
about.
Signed-off-by: NJay Soffian <jaysoffian@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

This is in preparation for teaching remote how to set
refs/remotes/<remote>/HEAD to match what HEAD is set to at <remote>, but
is useful in its own right.
Signed-off-by: NJay Soffian <jaysoffian@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Remote and stale branches are emitted in alphabetical order, but new and
tracked branches are not. So sort the latter to be consistent with the
former. This also lets us use more efficient string_list_has_string()
instead of unsorted_string_list_has_string().

"show <remote>" prunes symrefs, but "show <remote> -n" does not. Fix the
latter to match the former.
Signed-off-by: NJay Soffian <jaysoffian@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

When not querying the remote, show() was having to populate
states.tracked itself. It makes more sense for get_remote_ref_states()
to do this consistently. Since show() is the only caller of
get_remote_ref_states() with query=0, this change does not affect
other callers.
Signed-off-by: NJay Soffian <jaysoffian@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

- The variable name "remote" is used as both a "char *" and as a "struct
  remote *"; this is confusing, so rename the former to remote_name.

- Consistently refer to the refs returned by transport_get_remote_refs()
  as remote_refs.

- There is no need to call "sort_string_list(&branch_list)" as
  branch_list is populated via string_list_insert(), which maintains its
  order.
Signed-off-by: NJay Soffian <jaysoffian@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

get_ref_states() populates the util pointer of the string_list_item's
that it adds to states->new and states->tracked, but nothing ever uses
the pointer, so we can get rid of the extra code.
Signed-off-by: NJay Soffian <jaysoffian@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

This patch moves identical lines of code into a cleanup function. The
function has two callers and is about to gain a third.

Also removed a bogus NEEDSWORK comment per Daniel Barkalow:

  Actually, the comment is wrong; "remote" comes from remote_get(),
  which returns things from a cache in remote.c; there could be a
  remote_put() to let the code know that the caller is done with the
  object, but it wouldn't presently do anything.
Signed-off-by: NJay Soffian <jaysoffian@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Add a convenience function for iterating over a string_list's items via
a callback.
Signed-off-by: NJay Soffian <jaysoffian@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

match_refs() returns non-zero if there is an error in
match_explicit_refs(), without handling any remaining pattern ref specs.

Its existing callers exit upon receiving non-zero, so a partial result
is of no consequence to them; however a new caller, builtin-remote, is
interested in the complete result even if there are errors in
match_explicit_refs().
Signed-off-by: NJay Soffian <jaysoffian@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

In some instances, match_refs() sets the peer_ref field of refs in the
dst list such that it points to a ref in the src list. This prevents
callers from freeing both the src and dst lists, as doing so would cause
a double-free since free_refs() frees the peer_ref.

As well, the following configuration causes two refs in the dst list to
have the same peer_ref, which can also lead to a double-free:

  push = refs/heads/master:refs/heads/backup
  push = refs/heads/master:refs/heads/master

Existing callers of match_heads() call it only once and then terminate,
w/o ever bothering to free the src or dst lists, so this is not
currently a problem.

This patch modifies match_refs() to first copy any refs it plucks from
the src list before assigning them as a peer_ref. This allows
builtin-remote, a future caller, to free the src and dst lists.
Signed-off-by: NJay Soffian <jaysoffian@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Determining HEAD is ambiguous since it is done by comparing SHA1s.

In the case of multiple matches we return refs/heads/master if it
matches, else we return the first match we encounter. builtin-remote
needs all matches returned to it, so add a flag for it to request such.

To be simple and consistent, the return value is now a copy (including
peer_ref) of the matching refs.

Originally contributed by Jeff King along with the prior commit as a
single patch.
Signed-off-by: NJay Soffian <jaysoffian@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

To ensure that copied refs can always be freed w/o causing a
double-free, make copy_ref() perform a deep copy.

Also have copy_ref() return NULL if asked to copy NULL to simplify
things for the caller.

Background: currently copy_ref() performs a shallow copy. This is fine
for current callers who never free the result and/or only copy refs
which contain NULL pointers. But copy_ref() is about to gain a new
caller (guess_remote_head()) which copies refs where peer_ref is not
NULL and the caller of guess_remote_head() will want to free the result.
Signed-off-by: NJay Soffian <jaysoffian@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

This function had complications which made it hard to extend.

- It used to do two things: find the HEAD ref, and then find a
  matching ref, optionally returning the former via assignment to a
  passed-in pointer. Since finding HEAD is a one-liner, just have a
  caller do it themselves and pass it as an argument.

- It used to manually search through the ref list for
  refs/heads/master; this can be a one-line call to
  find_ref_by_name.

Originally contributed by Jeff King along with the next commit as a
single patch.
Signed-off-by: NJay Soffian <jaysoffian@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Move locate_head() to remote.c and rename it to guess_remote_head() to
more accurately reflect what it does. This is in preparation for being
able to call it from builtin-remote.c
Signed-off-by: NJay Soffian <jaysoffian@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

ref_newer() appears to have been copied from builtin-send-pack.c to
http-push.c via cut and paste. This patch moves the function and its
helper unmark_and_free() to remote.c. There was a slight difference
between the two implementations, one used TMP_MARK for the mark, the
other used 1. Per Jeff King, I went with TMP_MARK as more correct.

This is in preparation for being able to call it from builtin-remote.c
Signed-off-by: NJay Soffian <jaysoffian@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

get_local_heads() appears to have been copied from builtin-send-pack.c
to http-push.c via cut and paste. This patch moves the function and its
helper one_local_ref() to remote.c.

The two copies of one_local_ref() were not identical. I used the more
recent version from builtin-send-pack.c after confirming with Jeff King
that it was an oversight that commit 30affa1e did not update both
copies.

This is in preparation for being able to call it from builtin-remote.c
Signed-off-by: NJay Soffian <jaysoffian@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Since it doesn't actually touch its argument, this makes
sense.

However, we still want to return a non-const version (which
requires a cast) so that this:

  struct ref *a, *b;
  a = find_ref_by_name(b);

works. Unfortunately, you can also silently strip the const
from a variable:

  struct ref *a;
  const struct ref *b;
  a = find_ref_by_name(b);

This is a classic C const problem because there is no way to
say "return the type with the same constness that was passed
to us"; we provide the same semantics as standard library
functions like strchr.
Signed-off-by: NJeff King <peff@peff.net>
Signed-off-by: NJay Soffian <jaysoffian@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

This was mostly being tested implicitly by the "http push"
tests. But making a separate test script means that:

  - we will run fetch tests even when http pushing support
    is not built

  - when there are failures on fetching, they are easier to
    see and isolate, as they are not in the middle of push
    tests

This script defaults to running the webserver on port 5550,
and puts the original t5540 on port 5540, so that the two
can be run simultaneously without conflict (but both still
respect an externally set LIB_HTTPD_PORT).
Signed-off-by: NJeff King <peff@peff.net>
Signed-off-by: NJay Soffian <jaysoffian@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

There are some redirects and some error checking that need
to be done by the caller; let's move both into the
start_httpd function so that all callers don't have to
repeat them (there is only one caller now, but another will
follow in this series).

This doesn't violate any assumptions that aren't already
being made by lib-httpd, which is happy to say "skipping"
and call test_done for a number of other cases.
Signed-off-by: NJeff King <peff@peff.net>
Signed-off-by: NJay Soffian <jaysoffian@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Signed-off-by: NJunio C Hamano <gitster@pobox.com>

* jc/maint-1.6.0-pack-directory:
  Make sure objects/pack exists before creating a new pack

In a repository created with git older than f49fb35d (git-init-db: create
"pack" subdirectory under objects, 2005-06-27), objects/pack/ directory is
not created upon initialization.  It was Ok because subdirectories are
created as needed inside directories init-db creates, and back then,
packfiles were recent invention.

After the said commit, new codepaths started relying on the presense of
objects/pack/ directory in the repository.  This was exacerbated with
8b4eb6b6 (Do not perform cross-directory renames when creating packs,
2008-09-22) that moved the location temporary pack files are created from
objects/ directory to objects/pack/ directory, because moving temporary to
the final location was done carefully with lazy leading directory creation.

Many packfile related operations in such an old repository can fail
mysteriously because of this.

This commit introduces two helper functions to make things work better.

 - odb_mkstemp() is a specialized version of mkstemp() to refactor the
   code and teach it to create leading directories as needed;

 - odb_pack_keep() refactors the code to create a ".keep" file while
   create leading directories as needed.
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

This patch allows the HTTP tests to run on OS X 10.5. It is not
sufficient to be able to pass in LIB_HTTPD_PATH and
LIB_HTTPD_MODULE_PATH alone, as the apache.conf also needs a couple
tweaks.

These changes are put into an <IfDefine> to keep them Darwin specific,
but this means lib-httpd.sh needs to be modified to pass -DDarwin to
apache when running on Darwin. As long as we're making this change to
lib-httpd.sh, we may as well set LIB_HTTPD_PATH and
LIB_HTTPD_MODULE_PATH to appropriate default values for the platform.

Note that we now pass HTTPD_PARA to apache at shutdown as well.
Otherwise apache will emit a harmless, but noisy warning that LogFormat
is an unknown directive.
Signed-off-by: NJay Soffian <jaysoffian@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Signed-off-by: NMichael J Gruber <git@drmicha.warpmail.net>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

The rules how the patterns are matched against path names are the same
for .gitattribute and .gitignore files.

This also replace the notion "glob pattern" by "pattern" because
gitignore.txt talks about "glob" only in some contexts where the pattern
is mentioned.
Signed-off-by: NJohannes Sixt <j6t@kdbg.org>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

it's != its
Signed-off-by: NFelipe Contreras <felipe.contreras@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Global static variables don't need to be initialized to 0/NULL.
Signed-off-by: NFelipe Contreras <felipe.contreras@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

README suggested to look at "Documentation/gittutorial.txt" for the
tutorial and to use "man git-commandname" for documentation of each
command.

This was not consistent because the tutorial can also be available with
"man gittutorial" once git is installed, and the documentation for each
command can be available at "Documentation/git-commandname.txt" before
installing git.

This patch tries to make the description more consistent.  It also fixes
the path to the cvs-migration documentation that changed from
"Documentation/cvs-migration.txt" to "Documentation/gitcvs-migration.txt".
Signed-off-by: NChristian Couder <chriscool@tuxfamily.org>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Signed-off-by: NAllan Caffee <allan.caffee@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Signed-off-by: NMarc Branchaud <marcnarc@xiplink.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

When run without --author and it fails to determine an author, git
quiltimport tries `read patch_author` to get user input, but standard
input has been redirected to the patch series file.  This commit lets
quiltimport read the series file through file descriptor 3 so that the
standard input is preserved.

Reported by Uwe Kleine-König through http://bugs.debian.org/515910Signed-off-by: NGerrit Pape <pape@smarden.org>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

If 'make install' was run with sufficient privileges, then the installed
builtins in gitexecdir, which are either hardlinked, symlinked, or copied,
would receive the user and group of whoever built git.  With this commit
the initial hardlink or copy is done from the installation tree and not
the build tree to fix this.
Signed-off-by: NGerrit Pape <pape@smarden.org>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

* git://git.bogomips.org/git-svn:
  git-svn fix to avoid using strftime %z

After these changes, git-svnimport worked fine for me.
Signed-off-by: NAbhijit Menon-Sen <ams@toroid.org>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

%z isn't available on all platforms in the date formatting
routines.  Provide a workalike capability that should be
more portable.
Signed-off-by: NBen Walton <bwalton@artsci.utoronto.ca>
Acked-by: NEric Wong <normalperson@yhbt.net>