An earlier commit made "reset --hard" chattier but leaking its
message from "git rebase" (which calls it when first rewinding
the current branch to prepare replaying our own changes) without
explanation was confusing, so add an extra message to mention
it.  Inside restorestate in merge (which is rarely exercised
codepath, where more than one strategies are attempted),
resetting to the original state uses "reset --hard" -- this can
be squelched entirely.
Signed-off-by: NJunio C Hamano <junkio@cox.net>

We have a static allocation of MAXPARENT, but that limit was not
enforced, other than by a lucky invocation of the program segfaulting.
Signed-off-by: NJunio C Hamano <junkio@cox.net>

Like the existing error() function the new warn() function can be
used to describe a situation that probably should not be occuring,
but which the user (and Git) can continue to work around without
running into too many problems.

An example situation is a bad commit SHA1 found in a reflog.
Attempting to read this record out of the reflog isn't really an
error as we have skipped over it in the past.
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>
Signed-off-by: NJunio C Hamano <junkio@cox.net>

Because I am about to introduce a global warn() function (much
like the global error function) this global declaration would
conflict with the one supplied by imap-send.  Further since
imap-send's warn function output depends on its Quiet setting
we cannot simply remove its internal definition and use the
forthcoming global one.

So refactor warn() -> imap_warn() and info() -> imap_info()
(the latter was done just to be consistent in naming).
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>
Signed-off-by: NJunio C Hamano <junkio@cox.net>

Suggested by Rocco Rutte, Marco Roeland and others.
Signed-off-by: NJunio C Hamano <junkio@cox.net>

An earlier commit a71fb0a1 implemented much requested safety
valve to refuse "git pull" or "git pull origin" without explicit
refspecs from using the first set of remote refs obtained by
reading .git/remotes/origin file or branch.*.fetch configuration
variables to create a merge.  The argument was that while on a
branch different from the default branch, it is often wrong to
merge the default remote ref suitable for merging into the master.

That is fine as a theory.  But many repositories already in use
by people in the real world do not have any of the per branch
configuration crap.  They did not need it, and they do not need
it now.  Merging with the first remote ref listed was just fine,
because they had only one ref (e.g. 'master' from linux-2.6.git)
anyway.

So this changes the safety valve to be a lot looser.  When "git
fetch" gets only one remote branch, the irritating warning would
not trigger anymore.

I think we could also make the warning trigger when branch.*.merge
is not specified for the current branch, but is for some other
branch.  That is for another commit.
Signed-off-by: NJunio C Hamano <junkio@cox.net>

* maint:
  diff --check: fix off by one error
  spurious .sp in manpages

Signed-off-by: NJohannes Schindelin <Johannes.Schindelin@gmx.de>
Signed-off-by: NJunio C Hamano <junkio@cox.net>

This imitates the behaviour of git-commit.

Noticed by Han-Wen Nienhuys.
Signed-off-by: NJohannes Schindelin <Johannes.Schindelin@gmx.de>
Signed-off-by: NJunio C Hamano <junkio@cox.net>

This reverts commit a71fb0a1.

The logic to decide when to refuse to use the default "first set of
refs fetched" for merge was utterly bogus.

In a repository that happily worked correctly without any of the
per-branch configuration crap did not have (and did not have to
have) any branch.<current>.merge.  With that broken commit, pulling
from origin no longer would work.

Now that we have decided to make 'add' behave like 'update-index'
(and therefore fully classify update-index as strictly plumbing)
the am/revert/cherry-pick family of commands should not steer the
user towards update-index.  Instead send them to the command they
probably already know, 'add'.
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>
Signed-off-by: NJunio C Hamano <junkio@cox.net>

Would you believe? I edited git-merge-one-file (note the missing ".sh"!)
when I submitted the patch which became commit e2b70087...
Signed-off-by: NJohannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: NJunio C Hamano <junkio@cox.net>

When parsing the diff line starting with '@@', the line number of the
'+' file is parsed. For the subsequent line parses, the line number
should therefore be incremented after the parse, not before it.
Signed-off-by: NJohannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: NJunio C Hamano <junkio@cox.net>

Signed-off-by: NQuy Tonthat <qtonthat@gmail.com>
Signed-off-by: NJunio C Hamano <junkio@cox.net>

For Mac OS X 10.4, _XOPEN_SOURCE defines _POSIX_C_SOURCE which
hides many symbols from the program.

Breakage noticed and initial analysis provided by Randal
L. Schwartz.
Signed-off-by: NJunio C Hamano <junkio@cox.net>

* ew/svn-pm:
  git-svn: rename 'commit' command to 'set-tree'
  git-svn: remove support for the svn command-line client
  git-svn: convert to using Git.pm

This corrects minor remaining bits that still talked about <tree-ish>;
the Porcelain users (as opposed to plumbers) are mostly interested in
commits so use <commit> consistently and keep a sentence that mentions
that <tree-ish> can be used in place of them.

* jc/leftright:
  Revert "Make left-right automatic."
  Make left-right automatic.
  Teach all of log family --left-right output.
  rev-list --left-right

* jc/blame:
  blame: -b (blame.blankboundary) and --root (blame.showroot)

* jc/branch-remove-remote:
  git-branch -d: do not stop at the first failure.
  Teach git-branch to delete tracking branches with -r -d

* jc/clone:
  Move "no merge candidate" warning into git-pull
  Use preprocessor constants for environment variable names.
  Do not create $GIT_DIR/remotes/ directory anymore.
  Introduce GIT_TEMPLATE_DIR
  Revert "fix testsuite: make sure they use templates freshly built from the source"
  fix testsuite: make sure they use templates freshly built from the source
  git-clone: lose the traditional 'no-separate-remote' layout
  git-clone: lose the artificial "first" fetch refspec
  git-pull: refuse default merge without branch.*.merge
  git-clone: use wildcard specification for tracking branches

It is pointless.
Signed-off-by: NJunio C Hamano <junkio@cox.net>

This cherry-picks 7ef04350 that has been applied to the 'master'
to fix ".sp" in generated manpages.
Signed-off-by: NJunio C Hamano <junkio@cox.net>

Signed-off-by: NJunio C Hamano <junkio@cox.net>

* maint:
  GIT 1.4.4.3

Signed-off-by: NJunio C Hamano <junkio@cox.net>

If ever new object types are added for future extensions then better
have current git version report them as "unknown" instead of
"corrupted".
Signed-off-by: NNicolas Pitre <nico@cam.org>
Signed-off-by: NJunio C Hamano <junkio@cox.net>

This is a mechanical clean-up of the way *.c files include
system header files.

 (1) sources under compat/, platform sha-1 implementations, and
     xdelta code are exempt from the following rules;

 (2) the first #include must be "git-compat-util.h" or one of
     our own header file that includes it first (e.g. config.h,
     builtin.h, pkt-line.h);

 (3) system headers that are included in "git-compat-util.h"
     need not be included in individual C source files.

 (4) "git-compat-util.h" does not have to include subsystem
     specific header files (e.g. expat.h).
Signed-off-by: NJunio C Hamano <junkio@cox.net>

It was reported by Randal L. Schwartz <merlyn@stonehenge.com> that
indexing the Linux repository ~150MB pack takes about an hour on OS x
while it's a minute on Linux.  It seems that the OS X mmap()
implementation is more than 2 orders of magnitude slower than the Linux
one.

Linus proposed a patch replacing mmap() with pread() bringing index-pack
performance on OS X in line with the Linux one.  The performances on
Linux also improved by a small margin.
Signed-off-by: NNicolas Pitre <nico@cam.org>
Signed-off-by: NJunio C Hamano <junkio@cox.net>

Show "..." links in "summary" view to shortlog, heads (if there are
any), and tags (if there are any) only if there are more items to show
than shown already.

This means that "..." link is shown below shortened shortlog if there
are more than 16 commits, "..." link below shortened heads list if
there are more than 16 heads refs (16 branches), "..." link below
shortened tags list if there are more than 16 tags.

Modified patch from Jakub to to apply cleanly to master, also preform
the same "..." link logic to the forks list.
Signed-off-by: NJakub Narebski <jnareb@gmail.com>
Signed-off-by: NRobert Fitzsimons <robfitz@273k.net>
Signed-off-by: NJunio C Hamano <junkio@cox.net>

Noticed by SungHyun Nam.
Signed-off-by: NJunio C Hamano <junkio@cox.net>

It's rather misleading to have configuration options that don't do
anything.  If someone adds collision checking they might also want to
restore this option.
Signed-off-by: NBrian Gernhardt <benji@silverinsanity.com>
Signed-off-by: NJunio C Hamano <junkio@cox.net>

I hand munged the original patch when committing 1510fea7, and
screwed up the conversion.
Signed-off-by: NJunio C Hamano <junkio@cox.net>

In the same vein as 8336afa5,
this fixes the the RCS merge to git-merge-file conversion in
commit e2b70087.
Signed-off-by: NEric Wong <normalperson@yhbt.net>
Signed-off-by: NJunio C Hamano <junkio@cox.net>

This reverts commit 57612319.

Feeding symmetric difference to gitk is so useful, and it is the
same for other graphical Porcelains.  Rather than forcing them
to pass --no-left-right, making it optional.

Noticed and reported by Jeff King.

The warning triggered even when running "git fetch" only
when resulting .git/FETCH_HEAD only contained
branches marked as 'not-for-merge'.
Signed-off-by: NJosef Weidendorfer <weidendo@gmx.de>
Signed-off-by: NJunio C Hamano <junkio@cox.net>

We broke the discipline Linus set up to allow compiler help us
avoid typos in environment names in the early days of git over
time.  This defines a handful preprocessor constants for
environment variable names used in relatively core parts of the
system.

I've left out variable names specific to subsystems such as HTTP
and SSL as I do not think they are big problems.
Signed-off-by: NJunio C Hamano <junkio@cox.net>

Because we do not use --no-separate-remote anymore, there is no
reason to create that directory from the template.

t5510 test is updated to test both $GIT_DIR/remotes/ based
configuration and $GIT_DIR/config variable (credits to
Johannes).
Signed-off-by: NJunio C Hamano <junkio@cox.net>

* jc/test-clone: (35 commits)
  Introduce GIT_TEMPLATE_DIR
  Revert "fix testsuite: make sure they use templates freshly built from the source"
  fix testsuite: make sure they use templates freshly built from the source
  rerere: fix breakage of resolving.
  Add config example with respect to branch
  Add documentation for show-branch --topics
  make git a bit less cryptic on fetch errors
  make patch_delta() error cases a bit more verbose
  racy-git: documentation updates.
  show-ref: fix --exclude-existing
  parse-remote::expand_refs_wildcard()
  vim syntax: follow recent changes to commit template
  show-ref: fix --verify --hash=length
  show-ref: fix --quiet --verify
  avoid accessing _all_ loose refs in git-show-ref --verify
  git-fetch: Avoid reading packed refs over and over again
  Teach show-branch how to show ref-log data.
  markup fix in svnimport documentation.
  Documentation: new option -P for git-svnimport
  Fix mis-mark-up in git-merge-file.txt documentation
  ...

Instead of passing --template explicitely to init-db and clone, you can
just set the environment variable GIT_TEMPLATE_DIR.

Also make use of it in the tests, to make sure that the templates are
copied.
Signed-off-by: NJohannes Schindelin <Johannes.Schindelin@gmx.de>
Signed-off-by: NJunio C Hamano <junkio@cox.net>