Re: fix french translation

[Matthieu Moy]

乙酸鋰  writes:

> Sorry, I reversed the revisions. This one should be correct.

Yes, it is. But your patch is still not in the form expected by
Documentation/SubmittingPatches (missing sign-off-by, and incorrect
commit message).

The good news is: this is already fixed in git-gui's repo, it will land
in git.git next time Junio pulls from it:

http://repo.or.cz/w/git-gui.git/commitdiff/ab571ef02f5fc0d08bf2dce4e455d324e40fb220

-- 
Matthieu Moy
http://www-verimag.imag.fr/~moy/
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: git-svn too slow, contacts upstream svn repo

[Quark]

- Original Message -
> From: Quark 
> To: Git List 
> Cc: 
> Sent: Monday, 27 May 2013 8:55 PM
> Subject: git-svn too slow, contacts upstream svn repo
> 
> hello list,
> 
> I have been using git-svn in an corporate environment where svn repo has lot 
> of 
> branches, (lot means > 100). To avoid cloning all branches my config looks as 
> below
> 
> [svn-remote "svn"]
>     url = svn+ssh://url
>     fetch = srcroot/trunk:refs/remotes/trunk
>     branches = srcroot/branches/{branch_1, branch_2, 
> branch_3}:refs/remotes/*
> 
> Now, when a new branch of my interest is added in svn repo, I had like it to 
> be 
> in my git-repo as well. In an ideal world, one shall simply add that branch 
> name 
> to list in curly braces above, but that does not work. I had love to be 
> proved 
> wrong here. Somebody on stackoverflow.com suggested (to else's question, not 
> mine) to add one more "fetch" as show below.
> 
> [svn-remote "svn"]
>     url = svn+ssh://url
>     fetch = srcroot/trunk:refs/remotes/trunk
>     fetch = srcroot/branch_4:refs/remotes/* # NEW BRANCH
>     branches = srcroot/branches/{branch_1, branch_2, 
> branch_3}:refs/remotes/*
> 
> haven't tried it, neither do I like it.
> 
> I do have a solution which WORKS, it is modified version of an example from 
> \doc\git\html\git-svn.html page. Below I have pasted the example 
> and modified it to reflect my specifics.
> 
> # assume an existing git-svn repo
>     D:\sourcecode
> # Clone locally - make sure the refs/remotes/ space matches the server
>     mkdir project
>     cd project
>     git init
>     git remote add origin file:///D:/sourcecode # file:// is delibrate, 
> want 
> true n/w behaviour
>     git config --replace-all remote.origin.fetch 
> '+refs/remotes/*:refs/remotes/*'
>     git fetch
> # Prevent fetch/pull from local git repo,
> # we only want to use git svn for form here
>    git config --remove-section remote.origin
> # Create a local branch from one of the branches just fetched
>     git checkout -b master FETCH_HEAD
> # Initialize 'git svn' locally (be sure to use the same URL and -T/-b/-t 
> options as were used on server)
>     git svn init http://svn.example.com/project
> 
> # edit config to include newly added branch in curly braces
> # Pull the latest changes from Subversion
>     git svn fetch -r :HEAD
> 
> Above works exactly as I want, except the last step of "git-svn fetch" 
> connects to svn repo for each revision from BEGIN to HEAD. I want fairly long 
> history (for pretty blame). Which implies my BEGIN is not so recentish 
> adterall. 
> Hence, it takes awfully long to finish, 2 days typically.
> 
> When GIT_TRACE is 1, below is output received
> 
> trace: exec: 'git-svn' 'fetch'
> trace: run_command: 'git-svn' 'fetch'
> trace: built-in: git 'config' '--bool' '--get' 
> 'svn.fetchall'
> trace: built-in: git 'config' '--bool' '--get' 
> 'svn.parent'
> trace: built-in: git 'config' '--bool' '--get' 
> 'svn.noauthcache'
> trace: built-in: git 'config' '--get' 'svn.revision'
> trace: built-in: git 'config' '--bool' '--get' 
> 'svn.nocheckout'
> trace: built-in: git 'config' '--get' 'svn.authorsprog'
> trace: built-in: git 'config' '--bool' '--get' 
> 'svn.followparent'
> trace: built-in: git 'config' '--get' 'svn.authorsfile'
> trace: built-in: git 'config' '--bool' '--get' 
> 'svn.useSvmProps'
> trace: built-in: git 'config' '--get' 'svn.username'
> trace: built-in: git 'config' '--get' 'svn.repackflags'
> trace: built-in: git 'config' '--bool' '--get' 
> 'svn.localtime'
> trace: built-in: git 'config' '--int' '--get' 
> 'svn.repack'
> trace: built-in: git 'config' '--get' 'svn.ignorepaths'
> trace: built-in: git 'config' '--int' '--get' 
> 'svn.logwindowsize'
> trace: built-in: git 'config' '--bool' '--get' 
> 'svn.quiet'
> trace: built-in: git 'config' '--get' 'svn.ignorerefs'
> trace: built-in: git 'config' '--get' 'svn.configdir'
> trace: built-in: git 'config' '--bool' '--get' 
> 'svn.addauthorfrom'
> trace: built-in: git 'config' '--bool' '--get' 
> 'svn.useSvnsyncProps'
> trace: built-in: git 'config' '--bool' '--get' 
> 'svn.noMetadata'
> trace: built-in: git 'config' '--bool' '--get' 
> 'svn.uselogauthor'
> trace: built-in: git 'rev-parse' '--symbolic' '--all'
> trace: built-in: git 'config' '-l'
> trace: built-in: git 'config' '-l'
> trace: built-in: git 'config' '--bool' 'svn.useSvmProps'
> trace: built-in: git 'config' '-l'
> trace: built-in: git 'config' 'svn-remote.svn.branches-maxRev' 
> '524908'
> trace: built-in: git 'rev-list' '--pretty=raw' 
> '--reverse' 
> '9f1414be94ab007b62ace31bf4d210a069276127..refs/remotes/branch_1' 
> '--'
> trace: built-in: git 'rev-list' '--pretty=raw' 
> '--reverse' 
> '5eb0a454bcf066a8199b851add9ec07cde80119d..refs/remotes/branch_2' 
> '--'
> trace: built-in: git 'rev-list' '--pretty=raw' 
> '--reverse' 
> '383b68b8514010a71efe10821e5ccc3541903ceb..refs/remotes/branch_3' 
> '--'
> trace: built-in:

Re: [PATCH] fast-import: Remove redundant assignment of 'oe' to itself.

[Joachim Schmitz]

Stefano Lattarini wrote:

On 05/26/2013 10:05 PM, Stefan Beller wrote:

Reported by cppcheck.

Signed-off-by: Stefan Beller 
---
 fast-import.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/fast-import.c b/fast-import.c
index 5f539d7..0142e3a 100644
--- a/fast-import.c
+++ b/fast-import.c
@@ -2914,7 +2914,7 @@ static void cat_blob(struct object_entry *oe,
 unsigned char sha1[20]) static void parse_cat_blob(void)
 {
 const char *p;
- struct object_entry *oe = oe;


This was done on purpose, to avoid spurious warnings with (at least)
some versions of GCC.


+ struct object_entry *oe;
 unsigned char sha1[20];

 /* cat-blob SP  LF */




This strange construct has been removed in other places meanwhile. It is 
violating C-standards (C89, C99) and as such causes warnings with other 
compilers, so this is fighting fire with fire. As it is a pointer it may be 
more sensible to initialize with NULL, should appease all compilers and 
still be correct.


Bye, Jojo 



--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: git-send-email doesn't deal with quoted names

[John Keeping]

On Tue, May 28, 2013 at 01:40:20AM +0200, Jason A. Donenfeld wrote:
> My commit author name is "Jason A. Donenfeld". Because this has a dot,
> SMTP handling likes to put it in quotes.
> 
> git-send-email has this line:
> if (defined $author and $author ne $sender) {
> 
> With my name, this always winds up false, because it's comparing
> '"Jason A. Donenfeld" ' with 'Jason A. Donenfeld
> '.
> 
> So, the logic needs to be fixed somehow.

There was a patch for this recently, although it appears to be still
under discussion:

http://article.gmane.org/gmane.comp.version-control.git/225247
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH] git clone depth of 0 not possible.

[Matthijs Kooijman]

Hi Junio,

I'm interested in getting a fetch tip commit only feature into git, I'll
probably look into creating a patch for this.

> >>> Sounds buggy.  Would anything break if we were to make --depth=1 mean
> >>> "1 deep, including the tip commit"?
> >>
> >> As long as we do not change the meaning of the "shallow" count going
> >> over the wire (i.e. the number we receive from the user will be
> >> fudged, so that user's "depth 1" that used to mean "the tip and one
> >> behind it" is expressed as "depth 2" at the end-user level, and we
> >> send over the wire the number that corresponded to the old "depth
> >> 1"), I do not think anything will break, and then --depth=0 may
> >> magically start meaning "only the tip; its immediate parents will
> >> not be transferred and recorded as the shallow boundary in the
> >> receiving repository".
> >
> > I'd rather we reserve 0 for unlimited fetch, something we haven't done
> > so far [1]. And because "unlimited clone" with --depth does not make
> > sense, --depth=0 should be rejected by git-clone.
> 
> I actually was thinking about changing --depth=1 to mean "the tip,
> with zero commits behind it" (and that was consistent with my
> description of "fudging"), but ended up saying "--depth=0" by
> mistake.  I too think "--depth=0" or "--depth<0" does not make
> sense, so we are in agreement.

Did you consider how to implement this? Looking at the code, it seems
the "deepen" parameter in the wire protocol now means:
 - 0: Do not change anything about the shallowness (i.e., fetch
   everything from the shallow root to the tip).
 - > 0: Create new shallow commits at depth commits below the tip (so
   depth == 1 means tip and one below).
 - INFINITE_DEPTH (0x7fff): Remove all shallowness and fetch
   complete history.

Given this, I'm not sure how one can express "fetch the tip and nothing
below that", since depth == 0 already has a different meaning.

Of course, one could using depth == 1 in this case to receive two
commits and then drop one, but this would seem a bit pointless to me
(especially if the commit below the tip is very different from the tip
leading to a lot of useless data transfer).

Or did I misunderstand something here?

Gr.

Matthijs
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Feature-request: Ordering `git log --graph` based on *author's* timestamp

[Elliott Cable]

I use a fairly complex `git-log` command involving `--date-order` to
get an overview of my repository's status; but unfortunately,
`--date-order` seems to use the *committer* date, not the *author*
date. That means that each time I bring my topic branches up to date
by rebasing them onto the current upstream, I lose the helpful
chronological ordering in my `git-log` of relative commits in my topic
branches (that is, each branch becomes a single long line, because all
of its commits got rebased to sequential and nearly-identical
committer timestamps.)

After a bit of activity on [a Stack Overflow question][1], I've
realized that this is basically impossible without 1. dumping `git
log`'s `--graph` feature *entirely*, and writing a very complex custom
system to order and re-graph commits, or 2. getting `git log`
**itself** patched to support an `--author-order` flag or something.

(I've never posted here before, forgive me if this is the wrong way to
go about this … the #git channel on Freenode pointed me here. :)

   [1]: 

--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [QUERY] How do you sort completions?

[SZEDER Gábor]

Hi,

On Fri, May 24, 2013 at 08:11:00AM -0500, Felipe Contreras wrote:
> AFAIK bash needs the completions sorted, so we added some | sort |
> uniq.

Actually, it seems to sort and remove duplicates automatically:

$ _foo () { COMPREPLY=("bar" "foo" "bar" "baz") ; }
$ complete -F _foo foo
$ foo 
bar  baz  foo  

So that | sort | uniq might not be necessary after all, and by
removing them we could spare two fork()+exec()s.

I'm not sure all supported Bash versions behave this way, though.


Best,
Gábor

--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 2/2] cherry-pick: add --skip-commits option

[Joachim Schmitz]

Felipe Contreras wrote:

Pretty much what it says on the tin.


Only that it add --skip-empty and not --skip-commit ?!?

Bye, Jojo

--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH] prompt: fix show upstream with svn and zsh

[SZEDER Gábor]

Hi,

On Wed, May 22, 2013 at 09:32:44AM +0200, Thomas Gummerer wrote:
> SZEDER Gábor  writes:
> > On Tue, May 21, 2013 at 10:54:27PM +0200, Thomas Gummerer wrote:
> >> -  svn_remote[ $((${#svn_remote[@]} + 1)) ]="$value"
> >> +  svn_remote[$((${#svn_remote[@]} + 1))]="$value"
> >
> > I mean, did zsh really complained because of the space after the '[' ?!
> 
> Yes, removing the spaces after the '[' fixes the problem.  I'm not very
> proficient in shell scripting, so I can't tell if there is another
> cause.

Thanks.  I first suspected that this was just an independent style
cleanup, that's why I asked for clarification.


Gábor

--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 1/2] sequencer: trivial fix

[Neil Horman]

On Mon, May 27, 2013 at 11:52:18AM -0500, Felipe Contreras wrote:
> We should free objects before leaving.
> 
> Signed-off-by: Felipe Contreras 
> ---
>  sequencer.c | 7 +--
>  1 file changed, 5 insertions(+), 2 deletions(-)
> 
> diff --git a/sequencer.c b/sequencer.c
> index ab6f8a7..7eeae2f 100644
> --- a/sequencer.c
> +++ b/sequencer.c
> @@ -626,12 +626,15 @@ static int do_pick_commit(struct commit *commit, struct 
> replay_opts *opts)
>   rerere(opts->allow_rerere_auto);
>   } else {
>   int allow = allow_empty(opts, commit);
> - if (allow < 0)
> - return allow;
> + if (allow < 0) {
> + res = allow;
> + goto leave;
> + }
>   if (!opts->no_commit)
>   res = run_git_commit(defmsg, opts, allow);
>   }
>  
> +leave:
>   free_message(&msg);
>   free(defmsg);
>  
> -- 
> 1.8.3.rc3.312.g47657de
> 
> 
Acked-by: Neil Horman 
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 2/2] cherry-pick: add --skip-commits option

[Neil Horman]

On Mon, May 27, 2013 at 11:52:19AM -0500, Felipe Contreras wrote:
> Pretty much what it says on the tin.
> 
> Signed-off-by: Felipe Contreras 
> ---
>  Documentation/git-cherry-pick.txt   |  3 +++
>  builtin/revert.c|  2 ++
>  sequencer.c |  5 -
>  sequencer.h |  1 +
>  t/t3508-cherry-pick-many-commits.sh | 13 +
>  5 files changed, 23 insertions(+), 1 deletion(-)
> 
> diff --git a/Documentation/git-cherry-pick.txt 
> b/Documentation/git-cherry-pick.txt
> index c205d23..fccd936 100644
> --- a/Documentation/git-cherry-pick.txt
> +++ b/Documentation/git-cherry-pick.txt
> @@ -129,6 +129,9 @@ effect to your index in a row.
>   redundant commits are ignored.  This option overrides that behavior and
>   creates an empty commit object.  Implies `--allow-empty`.
>  
> +--skip-empty::
> + Instead of failing, skip commits that are or become empty.
> +
>  --strategy=::
>   Use the given merge strategy.  Should only be used once.
>   See the MERGE STRATEGIES section in linkgit:git-merge[1]
> diff --git a/builtin/revert.c b/builtin/revert.c
> index 0401fdb..0e5ce71 100644
> --- a/builtin/revert.c
> +++ b/builtin/revert.c
> @@ -118,6 +118,7 @@ static void parse_args(int argc, const char **argv, 
> struct replay_opts *opts)
>   OPT_END(),
>   OPT_END(),
>   OPT_END(),
> + OPT_END(),
>   };
>  
>   if (opts->action == REPLAY_PICK) {
> @@ -127,6 +128,7 @@ static void parse_args(int argc, const char **argv, 
> struct replay_opts *opts)
>   OPT_BOOLEAN(0, "allow-empty", &opts->allow_empty, 
> N_("preserve initially empty commits")),
>   OPT_BOOLEAN(0, "allow-empty-message", 
> &opts->allow_empty_message, N_("allow commits with empty messages")),
>   OPT_BOOLEAN(0, "keep-redundant-commits", 
> &opts->keep_redundant_commits, N_("keep redundant, empty commits")),
> + OPT_BOOLEAN(0, "skip-empty", &opts->skip_empty, 
> N_("skip empty commits")),
>   OPT_END(),
>   };
I like the idea, but this option seems a bit awkward to me.  At the very least
here, don't you now need to check for conflicts if --keep-redundant-commits and
skip-empty are both specified (as iirc git doens't see the difference between
empty commits and commits made empty by prior commits in the current history).
what if we merged the two options to an OPT_STRING, something like
--empty-commits=[keep|skip|ask].  The default currently is an impiled, since the
sequencer stops on an empty commit.

Neil

--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: git-svn too slow, contacts upstream svn repo

[Konstantin Khomoutov]

On Tue, 28 May 2013 15:42:57 +0800 (SGT)
Quark  wrote:

> > I have been using git-svn in an corporate environment where svn
> > repo has lot of branches, (lot means > 100). To avoid cloning all
> > branches my config looks as below
[...]
> is this not right forum?

As a matter of fact, this mailing list is the only correct place to ask
questions like yours.  But this is free software after all -- people
who could answer your question may be busy/absent at the moment or even
not involved in the project anymore (in the worst case).  So be
prepared to wait some time.  Also be prepared for your particular
trouble not being solved.

In the meantime, I think Thomas provided you with valuable suggestions
in reply to your mirror post on git-users, so you could possibly
explore them.
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH] wildmatch: properly fold case everywhere

[Anthony Ramine]

Case folding is not done correctly when matching against the [:upper:]
character class and uppercased character ranges (e.g. A-Z).
Specifically, an uppercase letter fails to match against any of them
when case folding is requested because plain characters in the pattern
and the whole string and preemptively lowercased to handle the base case
fast.

That optimization is kept and ISLOWER() is used in the [:upper:] case
when case folding is requested, while matching against a character range
is retried with toupper() if the character was lowercase.

Signed-off-by: Anthony Ramine 
---
 t/t3070-wildmatch.sh | 43 +--
 wildmatch.c  |  7 +++
 2 files changed, 44 insertions(+), 6 deletions(-)

diff --git a/t/t3070-wildmatch.sh b/t/t3070-wildmatch.sh
index 4c37057..17315aa 100755
--- a/t/t3070-wildmatch.sh
+++ b/t/t3070-wildmatch.sh
@@ -6,20 +6,20 @@ test_description='wildmatch tests'
 
 match() {
 if [ $1 = 1 ]; then
-   test_expect_success "wildmatch:match '$3' '$4'" "
+   test_expect_success "wildmatch: match '$3' '$4'" "
test-wildmatch wildmatch '$3' '$4'
"
 else
-   test_expect_success "wildmatch: no match '$3' '$4'" "
+   test_expect_success "wildmatch:  no match '$3' '$4'" "
! test-wildmatch wildmatch '$3' '$4'
"
 fi
 if [ $2 = 1 ]; then
-   test_expect_success "fnmatch:  match '$3' '$4'" "
+   test_expect_success "fnmatch:   match '$3' '$4'" "
test-wildmatch fnmatch '$3' '$4'
"
 elif [ $2 = 0 ]; then
-   test_expect_success "fnmatch:   no match '$3' '$4'" "
+   test_expect_success "fnmatch:no match '$3' '$4'" "
! test-wildmatch fnmatch '$3' '$4'
"
 #else
@@ -29,13 +29,25 @@ match() {
 fi
 }
 
+imatch() {
+if [ $1 = 1 ]; then
+   test_expect_success "iwildmatch:match '$2' '$3'" "
+   test-wildmatch iwildmatch '$2' '$3'
+   "
+else
+   test_expect_success "iwildmatch: no match '$2' '$3'" "
+   ! test-wildmatch iwildmatch '$2' '$3'
+   "
+fi
+}
+
 pathmatch() {
 if [ $1 = 1 ]; then
-   test_expect_success "pathmatch:match '$2' '$3'" "
+   test_expect_success "pathmatch: match '$2' '$3'" "
test-wildmatch pathmatch '$2' '$3'
"
 else
-   test_expect_success "pathmatch: no match '$2' '$3'" "
+   test_expect_success "pathmatch:  no match '$2' '$3'" "
! test-wildmatch pathmatch '$2' '$3'
"
 fi
@@ -235,4 +247,23 @@ pathmatch 1 abcXdefXghi '*X*i'
 pathmatch 1 ab/cXd/efXg/hi '*/*X*/*/*i'
 pathmatch 1 ab/cXd/efXg/hi '*Xg*i'
 
+# Case-sensitivy features
+match 0 x 'a' '[A-Z]'
+match 1 x 'A' '[A-Z]'
+match 0 x 'A' '[a-z]'
+match 1 x 'a' '[a-z]'
+match 0 x 'a' '[[:upper:]]'
+match 1 x 'A' '[[:upper:]]'
+match 0 x 'A' '[[:lower:]]'
+match 1 x 'a' '[[:lower:]]'
+
+imatch 1 'a' '[A-Z]'
+imatch 1 'A' '[A-Z]'
+imatch 1 'A' '[a-z]'
+imatch 1 'a' '[a-z]'
+imatch 1 'a' '[[:upper:]]'
+imatch 1 'A' '[[:upper:]]'
+imatch 1 'A' '[[:lower:]]'
+imatch 1 'a' '[[:lower:]]'
+
 test_done
diff --git a/wildmatch.c b/wildmatch.c
index 7192bdc..ea318d3 100644
--- a/wildmatch.c
+++ b/wildmatch.c
@@ -196,6 +196,11 @@ static int dowild(const uchar *p, const uchar *text, 
unsigned int flags)
}
if (t_ch <= p_ch && t_ch >= prev_ch)
matched = 1;
+   else if ((flags & WM_CASEFOLD) && 
ISLOWER(t_ch)) {
+   t_ch = toupper(t_ch);
+   if (t_ch <= p_ch && t_ch >= 
prev_ch)
+   matched = 1;
+   }
p_ch = 0; /* This makes "prev_ch" get 
set to 0. */
} else if (p_ch == '[' && p[1] == ':') {
const uchar *s;
@@ -245,6 +250,8 @@ static int dowild(const uchar *p, const uchar *text, 
unsigned int flags)
} else if (CC_EQ(s,i, "upper")) {
if (ISUPPER(t_ch))
matched = 1;
+   else if ((flags & WM_CASEFOLD) 
&& ISLOWER(t_ch))
+   matched = 1;
} else if (CC_EQ(s,i, "xdigit")) {
if (ISXDIGIT(t_ch))
matched = 1;
-- 
1.8.3

--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH] wildmatch: properly fold case everywhere

[Duy Nguyen]

On Tue, May 28, 2013 at 7:32 PM, Anthony Ramine  wrote:
> @@ -196,6 +196,11 @@ static int dowild(const uchar *p, const uchar *text, 
> unsigned int flags)
> }
> if (t_ch <= p_ch && t_ch >= prev_ch)
> matched = 1;
> +   else if ((flags & WM_CASEFOLD) && 
> ISLOWER(t_ch)) {
> +   t_ch = toupper(t_ch);

This happens in a while loop where t_ch may be used again. Should we
make a local copy of toupper(t_ch) and leave t_ch untouched?

> +   if (t_ch <= p_ch && t_ch >= 
> prev_ch)
> +   matched = 1;
> +   }
> p_ch = 0; /* This makes "prev_ch" get 
> set to 0. */
> } else if (p_ch == '[' && p[1] == ':') {
> const uchar *s;
--
Duy
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH 0/5] Trivial patches

[Felipe Contreras]

Hi,

Here's a bunch of trivial patches.

Felipe Contreras (5):
  remote: trivial style cleanup
  sequencer: trivial fix
  test: trivial cleanups
  test: improve rebase -q test
  test: rebase: fix --interactive test

 remote.c  |  3 +--
 sequencer.c   |  7 +--
 t/t3400-rebase.sh |  1 +
 t/t3403-rebase-skip.sh|  7 ---
 t/t3404-rebase-interactive.sh |  2 +-
 t/t3505-cherry-pick-empty.sh  | 18 +-
 6 files changed, 17 insertions(+), 21 deletions(-)

-- 
1.8.3.rc3.312.g47657de

--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH 1/5] remote: trivial style cleanup

[Felipe Contreras]

Signed-off-by: Felipe Contreras 
---
 remote.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/remote.c b/remote.c
index 68eb99b..e71f66d 100644
--- a/remote.c
+++ b/remote.c
@@ -1474,8 +1474,7 @@ struct branch *branch_get(const char *name)
ret->remote = remote_get(ret->remote_name);
if (ret->merge_nr) {
int i;
-   ret->merge = xcalloc(sizeof(*ret->merge),
-ret->merge_nr);
+   ret->merge = xcalloc(ret->merge_nr, 
sizeof(*ret->merge));
for (i = 0; i < ret->merge_nr; i++) {
ret->merge[i] = xcalloc(1, 
sizeof(**ret->merge));
ret->merge[i]->src = 
xstrdup(ret->merge_name[i]);
-- 
1.8.3.rc3.312.g47657de

--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH 2/5] sequencer: trivial fix

[Felipe Contreras]

We should free objects before leaving.

Signed-off-by: Felipe Contreras 
---
 sequencer.c | 7 +--
 1 file changed, 5 insertions(+), 2 deletions(-)

diff --git a/sequencer.c b/sequencer.c
index ab6f8a7..7eeae2f 100644
--- a/sequencer.c
+++ b/sequencer.c
@@ -626,12 +626,15 @@ static int do_pick_commit(struct commit *commit, struct 
replay_opts *opts)
rerere(opts->allow_rerere_auto);
} else {
int allow = allow_empty(opts, commit);
-   if (allow < 0)
-   return allow;
+   if (allow < 0) {
+   res = allow;
+   goto leave;
+   }
if (!opts->no_commit)
res = run_git_commit(defmsg, opts, allow);
}
 
+leave:
free_message(&msg);
free(defmsg);
 
-- 
1.8.3.rc3.312.g47657de

--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH 3/5] test: trivial cleanups

[Felipe Contreras]

No functional changes.

Signed-off-by: Felipe Contreras 
---
 t/t3403-rebase-skip.sh   |  7 ---
 t/t3505-cherry-pick-empty.sh | 18 +-
 2 files changed, 9 insertions(+), 16 deletions(-)

diff --git a/t/t3403-rebase-skip.sh b/t/t3403-rebase-skip.sh
index 826500b..3968020 100755
--- a/t/t3403-rebase-skip.sh
+++ b/t/t3403-rebase-skip.sh
@@ -64,10 +64,11 @@ test_expect_success 'rebase with --merge' '
 
 test_expect_success 'rebase --skip with --merge' '
git rebase --skip
-   '
+'
 
-test_expect_success 'merge and reference trees equal' \
-   'test -z "`git diff-tree skip-merge skip-reference`"'
+test_expect_success 'merge and reference trees equal' '
+   test -z "`git diff-tree skip-merge skip-reference`"
+'
 
 test_expect_success 'moved back to branch correctly' '
test refs/heads/skip-merge = $(git symbolic-ref HEAD)
diff --git a/t/t3505-cherry-pick-empty.sh b/t/t3505-cherry-pick-empty.sh
index a0c6e30..fbdc47c 100755
--- a/t/t3505-cherry-pick-empty.sh
+++ b/t/t3505-cherry-pick-empty.sh
@@ -28,29 +28,21 @@ test_expect_success setup '
 '
 
 test_expect_success 'cherry-pick an empty commit' '
-   git checkout master && {
-   git cherry-pick empty-branch^
-   test "$?" = 1
-   }
+   git checkout master &&
+   test_expect_code 1 git cherry-pick empty-branch^
 '
 
 test_expect_success 'index lockfile was removed' '
-
test ! -f .git/index.lock
-
 '
 
 test_expect_success 'cherry-pick a commit with an empty message' '
-   git checkout master && {
-   git cherry-pick empty-branch
-   test "$?" = 1
-   }
+   git checkout master &&
+   test_expect_code 1 git cherry-pick empty-branch
 '
 
 test_expect_success 'index lockfile was removed' '
-
test ! -f .git/index.lock
-
 '
 
 test_expect_success 'cherry-pick a commit with an empty message with 
--allow-empty-message' '
@@ -101,7 +93,7 @@ test_expect_success 'cherry-pick a no-op with 
--keep-redundant' '
git reset --hard &&
git checkout fork^0 &&
git cherry-pick --keep-redundant-commits master &&
-   git show -s --format='%s' >actual &&
+   git show -s --format=%s >actual &&
echo "add file2 on master" >expect &&
test_cmp expect actual
 '
-- 
1.8.3.rc3.312.g47657de

--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH 4/5] test: improve rebase -q test

[Felipe Contreras]

Let's show the output so it's clear why it failed.

Signed-off-by: Felipe Contreras 
---
 t/t3400-rebase.sh | 1 +
 1 file changed, 1 insertion(+)

diff --git a/t/t3400-rebase.sh b/t/t3400-rebase.sh
index b58fa1a..fb39531 100755
--- a/t/t3400-rebase.sh
+++ b/t/t3400-rebase.sh
@@ -185,6 +185,7 @@ test_expect_success 'default to @{upstream} when upstream 
arg is missing' '
 test_expect_success 'rebase -q is quiet' '
git checkout -b quiet topic &&
git rebase -q master >output.out 2>&1 &&
+   cat output.out &&
test ! -s output.out
 '
 
-- 
1.8.3.rc3.312.g47657de

--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH 5/5] test: rebase: fix --interactive test

[Felipe Contreras]

Signed-off-by: Felipe Contreras 
---
 t/t3404-rebase-interactive.sh | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/t/t3404-rebase-interactive.sh b/t/t3404-rebase-interactive.sh
index a58406d..79e8d3c 100755
--- a/t/t3404-rebase-interactive.sh
+++ b/t/t3404-rebase-interactive.sh
@@ -692,7 +692,7 @@ test_expect_success 'rebase -i can copy notes' '
test_commit n2 &&
test_commit n3 &&
git notes add -m"a note" n3 &&
-   git rebase --onto n1 n2 &&
+   git rebase -i --onto n1 n2 &&
test "a note" = "$(git notes show HEAD)"
 '
 
-- 
1.8.3.rc3.312.g47657de

--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re[4]: [PATCH 4/5] git-svn: fix bottleneck in stash_placeholder_list()

[Ilya Basin]

IB> * I think git-svn doesn't handle the case, when a tag is deleted.
IB>   I expected it to rename the ref from "tags/tagname" to
IB>   "tags/tagname@rev", but that doesn't happen.
IB>   If a tag is replaced, there's no way to tell what was the previous
IB>   state of that tag: git-svn just rewrites the ref.

OK, I figured out that git-svn creates a merge commit having one of
its parents the previous state of the tag and another parent the state
of the new copy src folder.

--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH 0/3] cherry-pick: improvments

[Felipe Contreras]

Hi,

Here's a bunch of changes to make cherry-pick (and revert) more useful.

Felipe Contreras (3):
  cherry-pick: add support to copy notes
  revert/cherry-pick: add --quiet option
  revert/cherry-pick: add --skip option

 Documentation/git-cherry-pick.txt |   7 +-
 Documentation/git-revert.txt  |   7 +-
 Documentation/sequencer.txt   |   3 +
 builtin/revert.c  |   9 +++
 sequencer.c   | 162 +-
 sequencer.h   |   6 +-
 t/t3500-cherry.sh |  32 
 t/t3510-cherry-pick-sequence.sh   |  12 +++
 8 files changed, 232 insertions(+), 6 deletions(-)

-- 
1.8.3.rc3.312.g47657de

--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH 1/3] cherry-pick: add support to copy notes

[Felipe Contreras]

Signed-off-by: Felipe Contreras 
---
 builtin/revert.c  |   2 +
 sequencer.c   | 136 --
 sequencer.h   |   2 +
 t/t3500-cherry.sh |  32 +
 4 files changed, 169 insertions(+), 3 deletions(-)

diff --git a/builtin/revert.c b/builtin/revert.c
index 0e5ce71..b977124 100644
--- a/builtin/revert.c
+++ b/builtin/revert.c
@@ -119,6 +119,7 @@ static void parse_args(int argc, const char **argv, struct 
replay_opts *opts)
OPT_END(),
OPT_END(),
OPT_END(),
+   OPT_END(),
};
 
if (opts->action == REPLAY_PICK) {
@@ -129,6 +130,7 @@ static void parse_args(int argc, const char **argv, struct 
replay_opts *opts)
OPT_BOOLEAN(0, "allow-empty-message", 
&opts->allow_empty_message, N_("allow commits with empty messages")),
OPT_BOOLEAN(0, "keep-redundant-commits", 
&opts->keep_redundant_commits, N_("keep redundant, empty commits")),
OPT_BOOLEAN(0, "skip-empty", &opts->skip_empty, 
N_("skip empty commits")),
+   OPT_BOOLEAN(0, "copy-notes", &opts->copy_notes, 
N_("copy notes")),
OPT_END(),
};
if (parse_options_concat(options, ARRAY_SIZE(options), 
cp_extra))
diff --git a/sequencer.c b/sequencer.c
index edf141d..35a84ad 100644
--- a/sequencer.c
+++ b/sequencer.c
@@ -20,6 +20,18 @@
 const char sign_off_header[] = "Signed-off-by: ";
 static const char cherry_picked_prefix[] = "(cherry picked from commit ";
 
+struct rewritten_list_item {
+   unsigned char from[20];
+   unsigned char to[20];
+};
+
+struct rewritten_list {
+   struct rewritten_list_item *items;
+   unsigned int nr, alloc;
+};
+
+static struct rewritten_list rewritten;
+
 static int is_rfc2822_line(const char *buf, int len)
 {
int i;
@@ -102,6 +114,40 @@ static int has_conforming_footer(struct strbuf *sb, struct 
strbuf *sob,
return 1;
 }
 
+static void add_rewritten(unsigned char *from, unsigned char *to)
+{
+   struct rewritten_list_item *item;
+   if (rewritten.nr + 1 >= rewritten.alloc) {
+   rewritten.alloc += 32;
+   rewritten.items = xrealloc(rewritten.items, rewritten.alloc * 
sizeof(*rewritten.items));
+   }
+   item = &rewritten.items[rewritten.nr];
+   hashcpy(item->from, from);
+   hashcpy(item->to, to);
+   rewritten.nr++;
+}
+
+static void copy_notes(void)
+{
+   unsigned char note_commit[20];
+   struct strbuf buf = STRBUF_INIT;
+   struct notes_tree *t = &default_notes_tree;
+   int i;
+
+   init_notes(t, NULL, NULL, 0);
+
+   for (i = 0; i < rewritten.nr; i++) {
+   struct rewritten_list_item *item = &rewritten.items[i];
+   copy_note(t, item->from, item->to, 0, NULL);
+   }
+
+   strbuf_addstr(&buf, "Notes added by 'git cherry-pick'\n");
+   create_notes_commit(&default_notes_tree, NULL, &buf, note_commit);
+   strbuf_insert(&buf, 0, "notes: ", 7);
+   update_ref(buf.buf, t->ref, note_commit, NULL, 0, MSG_ON_ERR);
+   strbuf_release(&buf);
+}
+
 static void remove_sequencer_state(void)
 {
struct strbuf seq_dir = STRBUF_INIT;
@@ -641,6 +687,14 @@ static int do_pick_commit(struct commit *commit, struct 
replay_opts *opts)
res = run_git_commit(defmsg, opts, allow);
}
 
+   if (!res && opts->copy_notes) {
+   unsigned char to[20];
+
+   if (read_ref("HEAD", to))
+   goto leave;
+
+   add_rewritten(commit->object.sha1, to);
+   }
 leave:
free_message(&msg);
free(defmsg);
@@ -786,6 +840,40 @@ static void read_populate_todo(struct commit_list 
**todo_list,
die(_("Unusable instruction sheet: %s"), todo_file);
 }
 
+static void read_populate_rewritten(void)
+{
+   const char *rewritten_file = git_path(SEQ_REWR_FILE);
+   struct strbuf buf = STRBUF_INIT;
+   char *p;
+   int fd;
+
+   fd = open(rewritten_file, O_RDONLY);
+   if (fd < 0)
+   return;
+   if (strbuf_read(&buf, fd, 0) < 0) {
+   close(fd);
+   strbuf_release(&buf);
+   return;
+   }
+   close(fd);
+
+   for (p = buf.buf; *p;) {
+   unsigned char from[20];
+   unsigned char to[20];
+   char *eol = strchrnul(p, '\n');
+   if (eol - p != 81)
+   /* wrong size */
+   break;
+   if (get_sha1_hex(p, from))
+   break;
+   if (get_sha1_hex(p + 41, to))
+   break;
+   add_rewritten(from, to);
+   p = *eol ? eol + 1 : eol;
+   }
+   strbuf_release(&buf);
+}
+
 static int populate_opts_cb(const char *key, const char *value, void *data)
 {
struct replay_opts 

[PATCH 2/3] revert/cherry-pick: add --quiet option

[Felipe Contreras]

Signed-off-by: Felipe Contreras 
---
 Documentation/git-cherry-pick.txt | 6 +-
 Documentation/git-revert.txt  | 6 +-
 builtin/revert.c  | 1 +
 sequencer.c   | 2 ++
 sequencer.h   | 1 +
 5 files changed, 14 insertions(+), 2 deletions(-)

diff --git a/Documentation/git-cherry-pick.txt 
b/Documentation/git-cherry-pick.txt
index fccd936..da0bd81 100644
--- a/Documentation/git-cherry-pick.txt
+++ b/Documentation/git-cherry-pick.txt
@@ -8,7 +8,7 @@ git-cherry-pick - Apply the changes introduced by some existing 
commits
 SYNOPSIS
 
 [verse]
-'git cherry-pick' [--edit] [-n] [-m parent-number] [-s] [-x] [--ff] ...
+'git cherry-pick' [-q] [--edit] [-n] [-m parent-number] [-s] [-x] [--ff] 
...
 'git cherry-pick' --continue
 'git cherry-pick' --quit
 'git cherry-pick' --abort
@@ -51,6 +51,10 @@ OPTIONS
feed all ... arguments to a single revision walk
(see a later example that uses 'maint master..next').
 
+-q::
+--quiet::
+   Quiet, suppress feedback messages.
+
 -e::
 --edit::
With this option, 'git cherry-pick' will let you edit the commit
diff --git a/Documentation/git-revert.txt b/Documentation/git-revert.txt
index f79c9d8..98a8e7a 100644
--- a/Documentation/git-revert.txt
+++ b/Documentation/git-revert.txt
@@ -8,7 +8,7 @@ git-revert - Revert some existing commits
 SYNOPSIS
 
 [verse]
-'git revert' [--[no-]edit] [-n] [-m parent-number] [-s] ...
+'git revert' [-q] [--[no-]edit] [-n] [-m parent-number] [-s] ...
 'git revert' --continue
 'git revert' --quit
 'git revert' --abort
@@ -40,6 +40,10 @@ OPTIONS
default, see linkgit:git-rev-list[1] and its '--no-walk'
option.
 
+-q::
+--quiet::
+   Quiet, suppress feedback messages.
+
 -e::
 --edit::
With this option, 'git revert' will let you edit the commit
diff --git a/builtin/revert.c b/builtin/revert.c
index b977124..d63b4a6 100644
--- a/builtin/revert.c
+++ b/builtin/revert.c
@@ -100,6 +100,7 @@ static void parse_args(int argc, const char **argv, struct 
replay_opts *opts)
int contin = 0;
int rollback = 0;
struct option options[] = {
+   OPT__QUIET(&opts->quiet, N_("suppress progress reporting")),
OPT_BOOLEAN(0, "quit", &remove_state, N_("end revert or 
cherry-pick sequence")),
OPT_BOOLEAN(0, "continue", &contin, N_("resume revert or 
cherry-pick sequence")),
OPT_BOOLEAN(0, "abort", &rollback, N_("cancel revert or 
cherry-pick sequence")),
diff --git a/sequencer.c b/sequencer.c
index 35a84ad..b4e395a 100644
--- a/sequencer.c
+++ b/sequencer.c
@@ -435,6 +435,8 @@ static int run_git_commit(const char *defmsg, struct 
replay_opts *opts,
argv_array_init(&array);
argv_array_push(&array, "commit");
argv_array_push(&array, "-n");
+   if (opts->quiet)
+   argv_array_push(&array, "-q");
 
if (opts->signoff)
argv_array_push(&array, "-s");
diff --git a/sequencer.h b/sequencer.h
index 6cc072c..41e19d0 100644
--- a/sequencer.h
+++ b/sequencer.h
@@ -37,6 +37,7 @@ struct replay_opts {
int keep_redundant_commits;
int skip_empty;
int copy_notes;
+   int quiet;
 
int mainline;
 
-- 
1.8.3.rc3.312.g47657de

--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH 3/3] revert/cherry-pick: add --skip option

[Felipe Contreras]

Akin to 'am --skip' and 'rebase --skip'.

Signed-off-by: Felipe Contreras 
---
 Documentation/git-cherry-pick.txt |  1 +
 Documentation/git-revert.txt  |  1 +
 Documentation/sequencer.txt   |  3 +++
 builtin/revert.c  |  6 ++
 sequencer.c   | 24 
 sequencer.h   |  3 ++-
 t/t3510-cherry-pick-sequence.sh   | 12 
 7 files changed, 49 insertions(+), 1 deletion(-)

diff --git a/Documentation/git-cherry-pick.txt 
b/Documentation/git-cherry-pick.txt
index da0bd81..d95c63c 100644
--- a/Documentation/git-cherry-pick.txt
+++ b/Documentation/git-cherry-pick.txt
@@ -10,6 +10,7 @@ SYNOPSIS
 [verse]
 'git cherry-pick' [-q] [--edit] [-n] [-m parent-number] [-s] [-x] [--ff] 
...
 'git cherry-pick' --continue
+'git cherry-pick' --skip
 'git cherry-pick' --quit
 'git cherry-pick' --abort
 
diff --git a/Documentation/git-revert.txt b/Documentation/git-revert.txt
index 98a8e7a..52e146e 100644
--- a/Documentation/git-revert.txt
+++ b/Documentation/git-revert.txt
@@ -10,6 +10,7 @@ SYNOPSIS
 [verse]
 'git revert' [-q] [--[no-]edit] [-n] [-m parent-number] [-s] ...
 'git revert' --continue
+'git revert' --skip
 'git revert' --quit
 'git revert' --abort
 
diff --git a/Documentation/sequencer.txt b/Documentation/sequencer.txt
index 5747f44..df2d355 100644
--- a/Documentation/sequencer.txt
+++ b/Documentation/sequencer.txt
@@ -3,6 +3,9 @@
'.git/sequencer'.  Can be used to continue after resolving
conflicts in a failed cherry-pick or revert.
 
+--skip::
+   Skip the current commit, and then continue.
+
 --quit::
Forget about the current operation in progress.  Can be used
to clear the sequencer state after a failed cherry-pick or
diff --git a/builtin/revert.c b/builtin/revert.c
index d63b4a6..6afd990 100644
--- a/builtin/revert.c
+++ b/builtin/revert.c
@@ -99,11 +99,13 @@ static void parse_args(int argc, const char **argv, struct 
replay_opts *opts)
int remove_state = 0;
int contin = 0;
int rollback = 0;
+   int skip = 0;
struct option options[] = {
OPT__QUIET(&opts->quiet, N_("suppress progress reporting")),
OPT_BOOLEAN(0, "quit", &remove_state, N_("end revert or 
cherry-pick sequence")),
OPT_BOOLEAN(0, "continue", &contin, N_("resume revert or 
cherry-pick sequence")),
OPT_BOOLEAN(0, "abort", &rollback, N_("cancel revert or 
cherry-pick sequence")),
+   OPT_BOOLEAN(0, "skip", &skip, N_("skip current commit in the 
sequence")),
OPT_BOOLEAN('n', "no-commit", &opts->no_commit, N_("don't 
automatically commit")),
OPT_BOOLEAN('e', "edit", &opts->edit, N_("edit the commit 
message")),
OPT_NOOP_NOARG('r', NULL),
@@ -160,6 +162,8 @@ static void parse_args(int argc, const char **argv, struct 
replay_opts *opts)
opts->subcommand = REPLAY_CONTINUE;
else if (rollback)
opts->subcommand = REPLAY_ROLLBACK;
+   else if (skip)
+   opts->subcommand = REPLAY_SKIP;
else
opts->subcommand = REPLAY_NONE;
 
@@ -170,6 +174,8 @@ static void parse_args(int argc, const char **argv, struct 
replay_opts *opts)
this_operation = "--quit";
else if (opts->subcommand == REPLAY_CONTINUE)
this_operation = "--continue";
+   else if (opts->subcommand == REPLAY_SKIP)
+   this_operation = "--skip";
else {
assert(opts->subcommand == REPLAY_ROLLBACK);
this_operation = "--abort";
diff --git a/sequencer.c b/sequencer.c
index b4e395a..971cab2 100644
--- a/sequencer.c
+++ b/sequencer.c
@@ -1173,6 +1173,28 @@ static int sequencer_continue(struct replay_opts *opts)
return pick_commits(todo_list, opts);
 }
 
+static int sequencer_skip(struct replay_opts *opts)
+{
+   const char *argv[4]; /* reset --hard HEAD + NULL */
+   struct string_list merge_rr = STRING_LIST_INIT_DUP;
+   int ret;
+
+   if (setup_rerere(&merge_rr, 0) >= 0) {
+   rerere_clear(&merge_rr);
+   string_list_clear(&merge_rr, 1);
+   }
+
+   argv[0] = "reset";
+   argv[1] = "--hard";
+   argv[2] = "HEAD";
+   argv[3] = NULL;
+   ret = run_command_v_opt(argv, RUN_GIT_CMD);
+   if (ret)
+   return ret;
+
+   return sequencer_continue(opts);
+}
+
 static int single_pick(struct commit *cmit, struct replay_opts *opts)
 {
int ret;
@@ -1209,6 +1231,8 @@ int sequencer_pick_revisions(struct replay_opts *opts)
return sequencer_rollback(opts);
if (opts->subcommand == REPLAY_CONTINUE)
return sequencer_continue(opts);
+   if (opts->subcommand == REPLAY_SKIP)
+   return sequencer_skip(opts);
 
for (i = 0; i < opts->revs->pending.nr; i++) {

Re: [PATCH] wildmatch: properly fold case everywhere

[Anthony Ramine]

You're right, I will amend my patch. How do I make git-send-email reply to that 
thread?

-- 
Anthony Ramine

Le 28 mai 2013 à 14:53, Duy Nguyen a écrit :

> On Tue, May 28, 2013 at 7:32 PM, Anthony Ramine  wrote:
>> @@ -196,6 +196,11 @@ static int dowild(const uchar *p, const uchar *text, 
>> unsigned int flags)
>>}
>>if (t_ch <= p_ch && t_ch >= prev_ch)
>>matched = 1;
>> +   else if ((flags & WM_CASEFOLD) && 
>> ISLOWER(t_ch)) {
>> +   t_ch = toupper(t_ch);
> 
> This happens in a while loop where t_ch may be used again. Should we
> make a local copy of toupper(t_ch) and leave t_ch untouched?
> 
>> +   if (t_ch <= p_ch && t_ch >= 
>> prev_ch)
>> +   matched = 1;
>> +   }
>>p_ch = 0; /* This makes "prev_ch" get 
>> set to 0. */
>>} else if (p_ch == '[' && p[1] == ':') {
>>const uchar *s;
> --
> Duy

--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH v2] wildmatch: properly fold case everywhere

[Anthony Ramine]

Case folding is not done correctly when matching against the [:upper:]
character class and uppercased character ranges (e.g. A-Z).
Specifically, an uppercase letter fails to match against any of them
when case folding is requested because plain characters in the pattern
and the whole string and preemptively lowercased to handle the base case
fast.

That optimization is kept and ISLOWER() is used in the [:upper:] case
when case folding is requested, while matching against a character range
is retried with toupper() if the character was lowercase.

Signed-off-by: Anthony Ramine 
---
 t/t3070-wildmatch.sh | 47 +--
 wildmatch.c  |  7 +++
 2 files changed, 48 insertions(+), 6 deletions(-)

diff --git a/t/t3070-wildmatch.sh b/t/t3070-wildmatch.sh
index 4c37057..e1b45e6 100755
--- a/t/t3070-wildmatch.sh
+++ b/t/t3070-wildmatch.sh
@@ -6,20 +6,20 @@ test_description='wildmatch tests'
 
 match() {
 if [ $1 = 1 ]; then
-   test_expect_success "wildmatch:match '$3' '$4'" "
+   test_expect_success "wildmatch: match '$3' '$4'" "
test-wildmatch wildmatch '$3' '$4'
"
 else
-   test_expect_success "wildmatch: no match '$3' '$4'" "
+   test_expect_success "wildmatch:  no match '$3' '$4'" "
! test-wildmatch wildmatch '$3' '$4'
"
 fi
 if [ $2 = 1 ]; then
-   test_expect_success "fnmatch:  match '$3' '$4'" "
+   test_expect_success "fnmatch:   match '$3' '$4'" "
test-wildmatch fnmatch '$3' '$4'
"
 elif [ $2 = 0 ]; then
-   test_expect_success "fnmatch:   no match '$3' '$4'" "
+   test_expect_success "fnmatch:no match '$3' '$4'" "
! test-wildmatch fnmatch '$3' '$4'
"
 #else
@@ -29,13 +29,25 @@ match() {
 fi
 }
 
+imatch() {
+if [ $1 = 1 ]; then
+   test_expect_success "iwildmatch:match '$2' '$3'" "
+   test-wildmatch iwildmatch '$2' '$3'
+   "
+else
+   test_expect_success "iwildmatch: no match '$2' '$3'" "
+   ! test-wildmatch iwildmatch '$2' '$3'
+   "
+fi
+}
+
 pathmatch() {
 if [ $1 = 1 ]; then
-   test_expect_success "pathmatch:match '$2' '$3'" "
+   test_expect_success "pathmatch: match '$2' '$3'" "
test-wildmatch pathmatch '$2' '$3'
"
 else
-   test_expect_success "pathmatch: no match '$2' '$3'" "
+   test_expect_success "pathmatch:  no match '$2' '$3'" "
! test-wildmatch pathmatch '$2' '$3'
"
 fi
@@ -235,4 +247,27 @@ pathmatch 1 abcXdefXghi '*X*i'
 pathmatch 1 ab/cXd/efXg/hi '*/*X*/*/*i'
 pathmatch 1 ab/cXd/efXg/hi '*Xg*i'
 
+# Case-sensitivy features
+match 0 x 'a' '[A-Z]'
+match 1 x 'A' '[A-Z]'
+match 0 x 'A' '[a-z]'
+match 1 x 'a' '[a-z]'
+match 0 x 'a' '[[:upper:]]'
+match 1 x 'A' '[[:upper:]]'
+match 0 x 'A' '[[:lower:]]'
+match 1 x 'a' '[[:lower:]]'
+match 0 x 'A' '[B-Za]'
+match 1 x 'a' '[B-Za]'
+
+imatch 1 'a' '[A-Z]'
+imatch 1 'A' '[A-Z]'
+imatch 1 'A' '[a-z]'
+imatch 1 'a' '[a-z]'
+imatch 1 'a' '[[:upper:]]'
+imatch 1 'A' '[[:upper:]]'
+imatch 1 'A' '[[:lower:]]'
+imatch 1 'a' '[[:lower:]]'
+imatch 1 'A' '[B-Za]'
+imatch 1 'a' '[B-Za]'
+
 test_done
diff --git a/wildmatch.c b/wildmatch.c
index 7192bdc..ea318d3 100644
--- a/wildmatch.c
+++ b/wildmatch.c
@@ -196,6 +196,11 @@ static int dowild(const uchar *p, const uchar *text, 
unsigned int flags)
}
if (t_ch <= p_ch && t_ch >= prev_ch)
matched = 1;
+   else if ((flags & WM_CASEFOLD) && 
ISLOWER(t_ch)) {
+   t_ch = toupper(t_ch);
+   if (t_ch <= p_ch && t_ch >= 
prev_ch)
+   matched = 1;
+   }
p_ch = 0; /* This makes "prev_ch" get 
set to 0. */
} else if (p_ch == '[' && p[1] == ':') {
const uchar *s;
@@ -245,6 +250,8 @@ static int dowild(const uchar *p, const uchar *text, 
unsigned int flags)
} else if (CC_EQ(s,i, "upper")) {
if (ISUPPER(t_ch))
matched = 1;
+   else if ((flags & WM_CASEFOLD) 
&& ISLOWER(t_ch))
+   matched = 1;
} else if (CC_EQ(s,i, "xdigit")) {
if (ISXDIGIT(t_ch))
matched = 1;
-- 
1.8.3

--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger

[PATCH 0/5] rebase: improve the keep-empty

[Felipe Contreras]

Hi,

I've been analyzing 'git rebase' and found that the --keep-empty option
triggers a very very different behavior. Here's a bunch of patches that make it
behave like the 'am' does does for the most part.

There's only a few minor changes, after which it might be possible to replace
the whole 'am' mode to use cherr-pick instead.

Felipe Contreras (5):
  rebase: split the cherry-pick stuff
  rebase: fix 'cherry' mode storage
  rebase: fix sequence continuation
  rebase: fix abort of cherry mode
  rebase: fix cherry-pick invocations

 .gitignore|  1 +
 Makefile  |  1 +
 git-rebase--am.sh | 65 ++-
 git-rebase--cherry.sh | 55 +++
 git-rebase.sh |  8 +++
 5 files changed, 93 insertions(+), 37 deletions(-)
 create mode 100644 git-rebase--cherry.sh

-- 
1.8.3.rc3.312.g47657de

--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH 1/5] rebase: split the cherry-pick stuff

[Felipe Contreras]

They do something completely different from 'git am', it belongs in a
different file.

Signed-off-by: Felipe Contreras 
---
 .gitignore|  1 +
 Makefile  |  1 +
 git-rebase--am.sh | 65 ++-
 git-rebase--cherry.sh | 34 +++
 git-rebase.sh |  4 
 5 files changed, 68 insertions(+), 37 deletions(-)
 create mode 100644 git-rebase--cherry.sh

diff --git a/.gitignore b/.gitignore
index 6669bf0..284fc8f 100644
--- a/.gitignore
+++ b/.gitignore
@@ -113,6 +113,7 @@
 /git-read-tree
 /git-rebase
 /git-rebase--am
+/git-rebase--cherry
 /git-rebase--interactive
 /git-rebase--merge
 /git-receive-pack
diff --git a/Makefile b/Makefile
index 0f931a2..a3cd4bc 100644
--- a/Makefile
+++ b/Makefile
@@ -469,6 +469,7 @@ SCRIPT_SH += git-web--browse.sh
 SCRIPT_LIB += git-mergetool--lib
 SCRIPT_LIB += git-parse-remote
 SCRIPT_LIB += git-rebase--am
+SCRIPT_LIB += git-rebase--cherry
 SCRIPT_LIB += git-rebase--interactive
 SCRIPT_LIB += git-rebase--merge
 SCRIPT_LIB += git-sh-setup
diff --git a/git-rebase--am.sh b/git-rebase--am.sh
index f84854f..ee1b1b9 100644
--- a/git-rebase--am.sh
+++ b/git-rebase--am.sh
@@ -19,52 +19,43 @@ esac
 test -n "$rebase_root" && root_flag=--root
 
 ret=0
-if test -n "$keep_empty"
-then
-   # we have to do this the hard way.  git format-patch completely squashes
-   # empty commits and even if it didn't the format doesn't really lend
-   # itself well to recording empty patches.  fortunately, cherry-pick
-   # makes this easy
-   git cherry-pick --allow-empty "$revisions"
-   ret=$?
-else
-   rm -f "$GIT_DIR/rebased-patches"
-
-   git format-patch -k --stdout --full-index --ignore-if-in-upstream \
-   --src-prefix=a/ --dst-prefix=b/ --no-renames --no-cover-letter \
-   $root_flag "$revisions" >"$GIT_DIR/rebased-patches"
-   ret=$?
 
-   if test 0 != $ret
-   then
-   rm -f "$GIT_DIR/rebased-patches"
-   case "$head_name" in
-   refs/heads/*)
-   git checkout -q "$head_name"
-   ;;
-   *)
-   git checkout -q "$orig_head"
-   ;;
-   esac
+rm -f "$GIT_DIR/rebased-patches"
 
-   cat >&2 <<-EOF
+git format-patch -k --stdout --full-index --ignore-if-in-upstream \
+   --src-prefix=a/ --dst-prefix=b/ --no-renames --no-cover-letter \
+   $root_flag "$revisions" >"$GIT_DIR/rebased-patches"
+ret=$?
 
-   git encountered an error while preparing the patches to replay
-   these revisions:
+if test 0 != $ret
+then
+   rm -f "$GIT_DIR/rebased-patches"
+   case "$head_name" in
+   refs/heads/*)
+   git checkout -q "$head_name"
+   ;;
+   *)
+   git checkout -q "$orig_head"
+   ;;
+   esac
 
-   $revisions
+   cat >&2 <<-EOF
 
-   As a result, git cannot rebase them.
-   EOF
-   exit $?
-   fi
+   git encountered an error while preparing the patches to replay
+   these revisions:
 
-   git am $git_am_opt --rebasing --resolvemsg="$resolvemsg" 
<"$GIT_DIR/rebased-patches"
-   ret=$?
+   $revisions
 
-   rm -f "$GIT_DIR/rebased-patches"
+   As a result, git cannot rebase them.
+   EOF
+   exit $?
 fi
 
+git am $git_am_opt --rebasing --resolvemsg="$resolvemsg" 
<"$GIT_DIR/rebased-patches"
+ret=$?
+
+rm -f "$GIT_DIR/rebased-patches"
+
 if test 0 != $ret
 then
test -d "$state_dir" && write_basic_state
diff --git a/git-rebase--cherry.sh b/git-rebase--cherry.sh
new file mode 100644
index 000..cbf80f9
--- /dev/null
+++ b/git-rebase--cherry.sh
@@ -0,0 +1,34 @@
+#!/bin/sh
+#
+# Copyright (c) 2010 Junio C Hamano.
+#
+
+case "$action" in
+continue)
+   git am --resolved --resolvemsg="$resolvemsg" &&
+   move_to_original_branch
+   exit
+   ;;
+skip)
+   git am --skip --resolvemsg="$resolvemsg" &&
+   move_to_original_branch
+   exit
+   ;;
+esac
+
+test -n "$rebase_root" && root_flag=--root
+
+# we have to do this the hard way.  git format-patch completely squashes
+# empty commits and even if it didn't the format doesn't really lend
+# itself well to recording empty patches.  fortunately, cherry-pick
+# makes this easy
+git cherry-pick --allow-empty "$revisions"
+ret=$?
+
+if test 0 != $ret
+then
+   test -d "$state_dir" && write_basic_state
+   exit $ret
+fi
+
+move_to_original_branch
diff --git a/git-rebase.sh b/git-rebase.sh
index 2c692c3..2754985 100755
--- a/git-rebase.sh
+++ b/git-rebase.sh
@@ -379,6 +379,10 @@ elif test -n "$do_merge"
 then
type=merge
state_dir="$merge_dir"
+elif test -n "$keep_empty"
+then
+   type=cherry
+   state_dir="$apply_dir"
 else
type=am
state_dir="$apply_dir"
-- 
1.8.3.rc3.312.g47657de

--
To

[PATCH 2/5] rebase: fix 'cherry' mode storage

[Felipe Contreras]

We don't use the 'rebase-apply'.

Signed-off-by: Felipe Contreras 
---
 git-rebase--cherry.sh | 4 
 git-rebase.sh | 5 -
 2 files changed, 8 insertions(+), 1 deletion(-)

diff --git a/git-rebase--cherry.sh b/git-rebase--cherry.sh
index cbf80f9..ab1f8b7 100644
--- a/git-rebase--cherry.sh
+++ b/git-rebase--cherry.sh
@@ -18,6 +18,9 @@ esac
 
 test -n "$rebase_root" && root_flag=--root
 
+mkdir "$state_dir" || die "Could not create temporary $state_dir"
+: > "$state_dir"/cherry || die "Could not mark as cherry"
+
 # we have to do this the hard way.  git format-patch completely squashes
 # empty commits and even if it didn't the format doesn't really lend
 # itself well to recording empty patches.  fortunately, cherry-pick
@@ -32,3 +35,4 @@ then
 fi
 
 move_to_original_branch
+rm -rf "$state_dir"
diff --git a/git-rebase.sh b/git-rebase.sh
index 2754985..b7759d5 100755
--- a/git-rebase.sh
+++ b/git-rebase.sh
@@ -174,6 +174,9 @@ then
then
type=interactive
interactive_rebase=explicit
+   elif test -f "$merge_dir"/cherry
+   then
+   type=cherry
else
type=merge
fi
@@ -382,7 +385,7 @@ then
 elif test -n "$keep_empty"
 then
type=cherry
-   state_dir="$apply_dir"
+   state_dir="$merge_dir"
 else
type=am
state_dir="$apply_dir"
-- 
1.8.3.rc3.312.g47657de

--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH 4/5] rebase: fix abort of cherry mode

[Felipe Contreras]

Signed-off-by: Felipe Contreras 
---
 git-rebase.sh | 1 +
 1 file changed, 1 insertion(+)

diff --git a/git-rebase.sh b/git-rebase.sh
index b7759d5..48bd1b8 100755
--- a/git-rebase.sh
+++ b/git-rebase.sh
@@ -335,6 +335,7 @@ skip)
run_specific_rebase
;;
 abort)
+   test "$type" == "cherry" && git cherry-pick --abort
git rerere clear
read_basic_state
case "$head_name" in
-- 
1.8.3.rc3.312.g47657de

--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH 5/5] rebase: fix cherry-pick invocations

[Felipe Contreras]

So that all the tests pass.

Signed-off-by: Felipe Contreras 
---
 git-rebase--cherry.sh | 17 -
 1 file changed, 16 insertions(+), 1 deletion(-)

diff --git a/git-rebase--cherry.sh b/git-rebase--cherry.sh
index ca78b1b..c3a2ac9 100644
--- a/git-rebase--cherry.sh
+++ b/git-rebase--cherry.sh
@@ -23,11 +23,26 @@ test -n "$rebase_root" && root_flag=--root
 mkdir "$state_dir" || die "Could not create temporary $state_dir"
 : > "$state_dir"/cherry || die "Could not mark as cherry"
 
+if test -n "$rebase_root"
+then
+   revisions="$onto...$orig_head"
+else
+   revisions="$upstream...$orig_head"
+fi
+
 # we have to do this the hard way.  git format-patch completely squashes
 # empty commits and even if it didn't the format doesn't really lend
 # itself well to recording empty patches.  fortunately, cherry-pick
 # makes this easy
-git cherry-pick --allow-empty "$revisions"
+if test -n "$keep_empty"
+then
+   extra="--allow-empty"
+else
+   extra="--skip-empty --cherry-pick"
+fi
+test -n "$GIT_QUIET" && extra="$extra -q"
+test -z "$force_rebase" && extra="$extra --ff"
+git cherry-pick --no-merges --right-only --topo-order --do-walk --copy-notes 
$extra "$revisions"
 ret=$?
 
 if test 0 != $ret
-- 
1.8.3.rc3.312.g47657de

--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH 3/5] rebase: fix sequence continuation

[Felipe Contreras]

We are not in am mode.

Signed-off-by: Felipe Contreras 
---
 git-rebase--cherry.sh | 10 ++
 1 file changed, 6 insertions(+), 4 deletions(-)

diff --git a/git-rebase--cherry.sh b/git-rebase--cherry.sh
index ab1f8b7..ca78b1b 100644
--- a/git-rebase--cherry.sh
+++ b/git-rebase--cherry.sh
@@ -5,13 +5,15 @@
 
 case "$action" in
 continue)
-   git am --resolved --resolvemsg="$resolvemsg" &&
-   move_to_original_branch
+   git cherry-pick --continue &&
+   move_to_original_branch &&
+   rm -rf "$state_dir"
exit
;;
 skip)
-   git am --skip --resolvemsg="$resolvemsg" &&
-   move_to_original_branch
+   git cherry-pick --skip &&
+   move_to_original_branch &&
+   rm -rf "$state_dir"
exit
;;
 esac
-- 
1.8.3.rc3.312.g47657de

--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 0/5] rebase: improve the keep-empty

[Felipe Contreras]

On Tue, May 28, 2013 at 8:29 AM, Felipe Contreras
 wrote:
> Hi,
>
> I've been analyzing 'git rebase' and found that the --keep-empty option
> triggers a very very different behavior. Here's a bunch of patches that make 
> it
> behave like the 'am' does does for the most part.
>
> There's only a few minor changes, after which it might be possible to replace
> the whole 'am' mode to use cherr-pick instead.
>
> Felipe Contreras (5):
>   rebase: split the cherry-pick stuff
>   rebase: fix 'cherry' mode storage
>   rebase: fix sequence continuation
>   rebase: fix abort of cherry mode
>   rebase: fix cherry-pick invocations

To see the issue, simply enable keep_mode by default, and run the tests:

Test Summary Report
---
./t3401-rebase-partial.sh  (Wstat: 256
Tests: 7 Failed: 4)
  Failed tests:  3, 5-7
  Non-zero exit status: 1
./t3403-rebase-skip.sh (Wstat: 256
Tests: 10 Failed: 3)
  Failed tests:  4-5, 9
  Non-zero exit status: 1
./t3406-rebase-message.sh  (Wstat: 256
Tests: 6 Failed: 3)
  Failed tests:  3-5
  Non-zero exit status: 1
./t3400-rebase.sh  (Wstat: 256
Tests: 27 Failed: 10)
  Failed tests:  13-17, 22-25, 27
  Non-zero exit status: 1
./t3407-rebase-abort.sh(Wstat: 256
Tests: 11 Failed: 5)
  Failed tests:  2-6
  Non-zero exit status: 1
./t3412-rebase-root.sh (Wstat: 256
Tests: 31 Failed: 17)
  Failed tests:  4, 6-8, 10-12, 14-17, 23, 25-26, 28-29
31
  Non-zero exit status: 1
./t3417-rebase-whitespace-fix.sh   (Wstat: 256
Tests: 4 Failed: 4)
  Failed tests:  1-4
  Non-zero exit status: 1
./t3419-rebase-patch-id.sh (Wstat: 256
Tests: 9 Failed: 1)
  Failed test:  4
  Non-zero exit status: 1
./t3418-rebase-continue.sh (Wstat: 256
Tests: 6 Failed: 2)
  Failed tests:  3, 5
  Non-zero exit status: 1
./t5407-post-rewrite-hook.sh   (Wstat: 256
Tests: 13 Failed: 4)
  Failed tests:  4-6, 8
  Non-zero exit status: 1
./t5520-pull.sh(Wstat: 256
Tests: 21 Failed: 8)
  Failed tests:  12-15, 18-21
  Non-zero exit status: 1
./t7512-status-help.sh (Wstat: 256
Tests: 35 Failed: 3)
  Failed tests:  5-6, 28
  Non-zero exit status: 1
./t9106-git-svn-commit-diff-clobber.sh (Wstat: 256
Tests: 10 Failed: 3)
  Failed tests:  6, 9-10
  Non-zero exit status: 1
./t3404-rebase-interactive.sh  (Wstat: 256
Tests: 71 Failed: 2)
  Failed tests:  50, 69
  Non-zero exit status: 1
./t9137-git-svn-dcommit-clobber-series.sh  (Wstat: 256
Tests: 5 Failed: 1)
  Failed test:  4
  Non-zero exit status: 1
./t9140-git-svn-reset.sh   (Wstat: 256
Tests: 6 Failed: 1)
  Failed test:  6
  Non-zero exit status: 1
./t9903-bash-prompt.sh (Wstat: 256
Tests: 51 Failed: 20)
  Failed tests:  27-34, 36-39, 41-42, 45-49, 51
  Non-zero exit status: 1
./t9164-git-svn-dcommit-concurrent.sh  (Wstat: 256
Tests: 8 Failed: 4)
  Failed tests:  4-6, 8
  Non-zero exit status: 1
Files=629, Tests=10036, 308 wallclock secs ( 7.60 usr  1.39 sys +
556.35 cusr 382.46 csys = 947.80 CPU)
Result: FAIL

-- 
Felipe Contreras
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH v3] wildmatch: properly fold case everywhere

[Anthony Ramine]

Case folding is not done correctly when matching against the [:upper:]
character class and uppercased character ranges (e.g. A-Z).
Specifically, an uppercase letter fails to match against any of them
when case folding is requested because plain characters in the pattern
and the whole string and preemptively lowercased to handle the base case
fast.

That optimization is kept and ISLOWER() is used in the [:upper:] case
when case folding is requested, while matching against a character range
is retried with toupper() if the character was lowercase.

Signed-off-by: Anthony Ramine 
---
 t/t3070-wildmatch.sh | 47 +--
 wildmatch.c  |  7 +++
 2 files changed, 48 insertions(+), 6 deletions(-)

Please disregard PATCH v2, it is identical to the first one.

diff --git a/t/t3070-wildmatch.sh b/t/t3070-wildmatch.sh
index 4c37057..e1b45e6 100755
--- a/t/t3070-wildmatch.sh
+++ b/t/t3070-wildmatch.sh
@@ -6,20 +6,20 @@ test_description='wildmatch tests'
 
 match() {
 if [ $1 = 1 ]; then
-   test_expect_success "wildmatch:match '$3' '$4'" "
+   test_expect_success "wildmatch: match '$3' '$4'" "
test-wildmatch wildmatch '$3' '$4'
"
 else
-   test_expect_success "wildmatch: no match '$3' '$4'" "
+   test_expect_success "wildmatch:  no match '$3' '$4'" "
! test-wildmatch wildmatch '$3' '$4'
"
 fi
 if [ $2 = 1 ]; then
-   test_expect_success "fnmatch:  match '$3' '$4'" "
+   test_expect_success "fnmatch:   match '$3' '$4'" "
test-wildmatch fnmatch '$3' '$4'
"
 elif [ $2 = 0 ]; then
-   test_expect_success "fnmatch:   no match '$3' '$4'" "
+   test_expect_success "fnmatch:no match '$3' '$4'" "
! test-wildmatch fnmatch '$3' '$4'
"
 #else
@@ -29,13 +29,25 @@ match() {
 fi
 }
 
+imatch() {
+if [ $1 = 1 ]; then
+   test_expect_success "iwildmatch:match '$2' '$3'" "
+   test-wildmatch iwildmatch '$2' '$3'
+   "
+else
+   test_expect_success "iwildmatch: no match '$2' '$3'" "
+   ! test-wildmatch iwildmatch '$2' '$3'
+   "
+fi
+}
+
 pathmatch() {
 if [ $1 = 1 ]; then
-   test_expect_success "pathmatch:match '$2' '$3'" "
+   test_expect_success "pathmatch: match '$2' '$3'" "
test-wildmatch pathmatch '$2' '$3'
"
 else
-   test_expect_success "pathmatch: no match '$2' '$3'" "
+   test_expect_success "pathmatch:  no match '$2' '$3'" "
! test-wildmatch pathmatch '$2' '$3'
"
 fi
@@ -235,4 +247,27 @@ pathmatch 1 abcXdefXghi '*X*i'
 pathmatch 1 ab/cXd/efXg/hi '*/*X*/*/*i'
 pathmatch 1 ab/cXd/efXg/hi '*Xg*i'
 
+# Case-sensitivy features
+match 0 x 'a' '[A-Z]'
+match 1 x 'A' '[A-Z]'
+match 0 x 'A' '[a-z]'
+match 1 x 'a' '[a-z]'
+match 0 x 'a' '[[:upper:]]'
+match 1 x 'A' '[[:upper:]]'
+match 0 x 'A' '[[:lower:]]'
+match 1 x 'a' '[[:lower:]]'
+match 0 x 'A' '[B-Za]'
+match 1 x 'a' '[B-Za]'
+
+imatch 1 'a' '[A-Z]'
+imatch 1 'A' '[A-Z]'
+imatch 1 'A' '[a-z]'
+imatch 1 'a' '[a-z]'
+imatch 1 'a' '[[:upper:]]'
+imatch 1 'A' '[[:upper:]]'
+imatch 1 'A' '[[:lower:]]'
+imatch 1 'a' '[[:lower:]]'
+imatch 1 'A' '[B-Za]'
+imatch 1 'a' '[B-Za]'
+
 test_done
diff --git a/wildmatch.c b/wildmatch.c
index 7192bdc..f91ba99 100644
--- a/wildmatch.c
+++ b/wildmatch.c
@@ -196,6 +196,11 @@ static int dowild(const uchar *p, const uchar *text, 
unsigned int flags)
}
if (t_ch <= p_ch && t_ch >= prev_ch)
matched = 1;
+   else if ((flags & WM_CASEFOLD) && 
ISLOWER(t_ch)) {
+   uchar t_ch_upper = 
toupper(t_ch);
+   if (t_ch_upper <= p_ch && 
t_ch_upper >= prev_ch)
+   matched = 1;
+   }
p_ch = 0; /* This makes "prev_ch" get 
set to 0. */
} else if (p_ch == '[' && p[1] == ':') {
const uchar *s;
@@ -245,6 +250,8 @@ static int dowild(const uchar *p, const uchar *text, 
unsigned int flags)
} else if (CC_EQ(s,i, "upper")) {
if (ISUPPER(t_ch))
matched = 1;
+   else if ((flags & WM_CASEFOLD) 
&& ISLOWER(t_ch))
+   matched = 1;
} else if (CC_EQ(s,i, "xdigit")) {
if (ISXDIGIT(t_ch))
matched = 1;
-- 
1.8.3

--
To unsubscribe fro

Re: [PATCH v2 0/3] Towards a useable git-branch

[Ramkumar Ramachandra]

Hi Duy,

I just woke up and started looking at the series: it's rather well
done, and I'm confident that this is the way forward.  To reciprocate,
I've done some work at gh:artagnon/git for-each-ref-pretty.  See:

https://github.com/artagnon/git/commits/for-each-ref-pretty

There is one major problem though:

%>(N) doesn't work properly with f-e-r, and I'm not sure why.  I'm not
talking about your last patch where you compute * -- that works fine;
it's just that %>(N) doesn't when N is a concrete number.

Also, a couple of minor annoyances:

1. When f-e-r is invoked with refs/tags, we get stray output.  Atleast
it doesn't segfault, thanks to your ignore-commit patch.  Maybe
printing stray output is the right thing to do, as opposed to erroring
out.

2. %>(*) only works with f-e-r atoms, not with pretty-format atoms.
This is ofcourse obvious from the implementation, but isn't it a
little consistent?

Should we start off a new pretty-ref-formats document, where we
explicitly exclude things like %ae (because of the hex overriding
thing)?  I don't think it's a problem if documented properly.
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH v2 0/3] Towards a useable git-branch

[Ramkumar Ramachandra]

Oh, and by the way:

We're pretty close we are to replacing branch -v and branch -vv.

brv = for-each-ref --format='%(HEAD)
%C(green)%<(*)%(refname:short)%C(reset) %<(*)%(objectname:short)
%(subject)' refs/heads

brvv = for-each-ref --format='%(HEAD)
%C(green)%<(*)%(refname:short)%C(reset) %<(*)%(objectname:short)
%C(blue)%(upstream:short)%C(reset) %(subject)' refs/heads

There are small differences:

1. In branch -v, the green-color of the branch name is dependent on
%(HEAD).  Not worth ironing out, in my opinion.

2. In branch -vv, there are dependent square braces that come on when
%(refname:short) is set.  We might want to introduce an undocumented
%(refname:branchvv) for internal use by branch -vv, for backward
compatibility.

What do you think?
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH v2 0/3] Towards a useable git-branch

[Ramkumar Ramachandra]

Ramkumar Ramachandra wrote:
> %>(N) doesn't work properly with f-e-r, and I'm not sure why.  I'm not
> talking about your last patch where you compute * -- that works fine;
> it's just that %>(N) doesn't when N is a concrete number.

Try this:

%(refname:short)%>(30)%(upstream:short)

(assuming that you have lots of branches).  I'm noticing random
alignment problems.

However, %<(N) doesn't seem to have this problem:

%<(30)%(refname:short)%(upstream:short)

I'm not able to figure this out.
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: git-svn too slow, contacts upstream svn repo

[Quark]

- Original Message -
> From: Konstantin Khomoutov 
> To: Quark 
> Cc: Git List 
> Sent: Tuesday, 28 May 2013 5:24 PM
> Subject: Re: git-svn too slow, contacts upstream svn repo
> 
> On Tue, 28 May 2013 15:42:57 +0800 (SGT)
> Quark  wrote:
> 
>>  > I have been using git-svn in an corporate environment where svn
>>  > repo has lot of branches, (lot means > 100). To avoid cloning all
>>  > branches my config looks as below
> [...]
>>  is this not right forum?
> 
> As a matter of fact, this mailing list is the only correct place to ask
> questions like yours.  But this is free software after all -- people
> who could answer your question may be busy/absent at the moment or even
> not involved in the project anymore (in the worst case).  So be
> prepared to wait some time.  Also be prepared for your particular
> trouble not being solved.
> 
> In the meantime, I think Thomas provided you with valuable suggestions
> in reply to your mirror post on git-users, so you could possibly
> explore them.
>

I was impatient not because feeling-of-entitlement, but I thought it might get 
lost in pile of e-mails
Secondly was not sure if this is right forum, I might be making noise after 
all, thanks for your clarification.
Now will you please excuse, I have to check Thomas's suggestions? :)

--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH] git clone depth of 0 not possible.

[Jonathan Nieder]

Matthijs Kooijman wrote:

> Did you consider how to implement this? Looking at the code, it seems
> the "deepen" parameter in the wire protocol now means:
>  - 0: Do not change anything about the shallowness (i.e., fetch
>everything from the shallow root to the tip).
>  - > 0: Create new shallow commits at depth commits below the tip (so
>depth == 1 means tip and one below).
>  - INFINITE_DEPTH (0x7fff): Remove all shallowness and fetch
>complete history.
>
> Given this, I'm not sure how one can express "fetch the tip and nothing
> below that", since depth == 0 already has a different meaning.

If I remember correctly, what we discussed is just changing the
protocol to "5 means a depth of 5". The client already trusts what the
server provides.

Thanks and hope that helps,
Jonathan
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH] git clone depth of 0 not possible.

[Jonathan Nieder]

Jonathan Nieder wrote:

> If I remember correctly, what we discussed is just changing the
> protocol to "5 means a depth of 5". The client already trusts what the
> server provides.

(Or tweaking the protocol by adding a new capability, if unpredictable
behavior based on the version of the server won't fly. :))

Jonathan
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH] git clone depth of 0 not possible.

[Matthijs Kooijman]

Hi Jonathan,

> > Did you consider how to implement this? Looking at the code, it seems
> > the "deepen" parameter in the wire protocol now means:
> >  - 0: Do not change anything about the shallowness (i.e., fetch
> >everything from the shallow root to the tip).
> >  - > 0: Create new shallow commits at depth commits below the tip (so
> >depth == 1 means tip and one below).
> >  - INFINITE_DEPTH (0x7fff): Remove all shallowness and fetch
> >complete history.
> >
> > Given this, I'm not sure how one can express "fetch the tip and nothing
> > below that", since depth == 0 already has a different meaning.
> 
> If I remember correctly, what we discussed is just changing the
> protocol to "5 means a depth of 5".

The mail from Junio I replied to said:
> >> As long as we do not change the meaning of the "shallow" count
> >> going over the wire

Which seems to conflict with your suggestion. Or are the "shallow count"
and the "depth" different things?

> The client already trusts what the server provides.
In other words: we won't break existing clients if we suddenly send back
one less commit than before, since the client just sends over what it
wants and then assumes that whatever it gets back is really what it
wanted?

Gr.

Matthijs
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH] git clone depth of 0 not possible.

[Jonathan Nieder]

Matthijs Kooijman wrote:

> In other words: we won't break existing clients if we suddenly send back
> one less commit than before, since the client just sends over what it
> wants and then assumes that whatever it gets back is really what it
> wanted?

Yes, depending on your definition of "break".

An advantage of that approach is that old clients would get the new,
intuitive behavior without upgrading. A disadvantage is that it is a
confusing world where the same command produces different effects when
contacting different servers.
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH v2 7/8] remote-bzr: reorganize the way 'wanted' works

[Junio C Hamano]

Felipe Contreras  writes:

> +wanted = get_config('remote-bzr.branches').rstrip().split(', ')

Two minor nits and one design suggestion:

 - Why rstrip() not strip()?  It appears that this only is helping
   an end-user "mistake" like this:

git config remote-bzr.branches 'trunk, devel, test '

   without helping people who have done this:

git config remote-bzr.branches 'trunk,  devel, test'

 - Is

 git config remote-bzr.branches trunk,devel,test

   a grave sin?

   In other words, wouldn't we want something like this instead?

map(lambda s: s.strip(), get_config('...').split(','))

 - Doesn't allowing multi-valued variable, e.g.

[remote-bzr]
branches = trunk
branches = devel
branches = test

   make it easier for the user to manage this configuration by
   e.g. selectively removing or adding tracked branches?
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 1/2] sequencer: trivial fix

[Junio C Hamano]

Neil Horman  writes:

> On Mon, May 27, 2013 at 11:52:18AM -0500, Felipe Contreras wrote:
>> We should free objects before leaving.
>> 
>> Signed-off-by: Felipe Contreras 
>> ---
>>  sequencer.c | 7 +--
>>  1 file changed, 5 insertions(+), 2 deletions(-)
>> 
>> diff --git a/sequencer.c b/sequencer.c
>> index ab6f8a7..7eeae2f 100644
>> --- a/sequencer.c
>> +++ b/sequencer.c
>> @@ -626,12 +626,15 @@ static int do_pick_commit(struct commit *commit, 
>> struct replay_opts *opts)
>>  rerere(opts->allow_rerere_auto);
>>  } else {
>>  int allow = allow_empty(opts, commit);
>> -if (allow < 0)
>> -return allow;
>> +if (allow < 0) {
>> +res = allow;
>> +goto leave;
>> +}
>>  if (!opts->no_commit)
>>  res = run_git_commit(defmsg, opts, allow);
>>  }
>>  
>> +leave:
>>  free_message(&msg);
>>  free(defmsg);
>>  
>> -- 
>> 1.8.3.rc3.312.g47657de
>> 
>> 
> Acked-by: Neil Horman 

This is better done without "goto" in general.

The other patch 2/2/ adds one more "we need to exit from the middle
of the flow" and makes it look handier to add an exit label here,
but it would be even better to express the logic of that patch as a
normal cascade of if/else if/..., which is small enough and we do
not need the "leave:" label.

It probably is better to fold this patch into the other one when it
is rerolled to correct the option name gotcha "on the tin".

Thanks.
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH RESEND v2] path: Fix a sparse warning

[Junio C Hamano]

Ramsay Jones  writes:

> On MinGW, sparse issues an "'get_st_mode_bits' not declared. Should
> it be static?" warning. The MinGW and MSVC builds do not see the
> declaration of this function, within git-compat-util.h, due to its
> placement within an preprocessor conditional.
>
> In order to suppress the warning, we simply move the declaration to
> the top level of the header.
>
> Signed-off-by: Ramsay Jones 
> ---
>
> Hi Junio,
>
> Now that v1.8.3 is out, I note that this patch seems to have been
> dropped (or did I miss something?).
>
> This used to be
>
> [PATCH 2/6] path: Make the 'get_st_mode_bits' symbol a file static
>
> but the change in implementation required a change in title.
> This version simply moves the declaration so that the MinGW and
> MSVC builds can see it.

Will queue.

Can you tell me what the conclusion on the discussion on your two
other patches on 'pu'?

* rj/mingw-cygwin (2013-05-08) 2 commits
 - cygwin: Remove the CYGWIN_V15_WIN32API build variable
 - mingw: rename WIN32 cpp macro to GIT_WINDOWS_NATIVE

I stopped keeping track of the discussion and my vague recollection
was that it is OK for 1.5 but not verified on 1.7 or something?

>
> ATB,
> Ramsay Jones
>
>  git-compat-util.h | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/git-compat-util.h b/git-compat-util.h
> index e955bb5..0e5e4f8 100644
> --- a/git-compat-util.h
> +++ b/git-compat-util.h
> @@ -163,7 +163,6 @@
>  typedef long intptr_t;
>  typedef unsigned long uintptr_t;
>  #endif
> -int get_st_mode_bits(const char *path, int *mode);
>  #if defined(__CYGWIN__)
>  #undef _XOPEN_SOURCE
>  #include 
> @@ -176,6 +175,8 @@ int get_st_mode_bits(const char *path, int *mode);
>  #endif
>  #endif
>  
> +extern int get_st_mode_bits(const char *path, int *mode);
> +
>  /* used on Mac OS X */
>  #ifdef PRECOMPOSE_UNICODE
>  #include "compat/precompose_utf8.h"
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH] git clone depth of 0 not possible.

[Junio C Hamano]

Matthijs Kooijman  writes:

> Did you consider how to implement this? Looking at the code, it seems
> the "deepen" parameter in the wire protocol now means:
>  - 0: Do not change anything about the shallowness (i.e., fetch
>everything from the shallow root to the tip).
>  - > 0: Create new shallow commits at depth commits below the tip (so
>depth == 1 means tip and one below).
>  - INFINITE_DEPTH (0x7fff): Remove all shallowness and fetch
>complete history.
>
> Given this, I'm not sure how one can express "fetch the tip and nothing
> below that", since depth == 0 already has a different meaning.

Doing it "correctly" (in the shorter term) would involve:

 - adding a capability on the sending side "fixed-off-by-one-depth"
   to the protocol, and teaching the sending side to advertise the
   capability;
   
 - teaching the requestor that got --depth=N from the end user to
   pay attention to the new capability in such a way that:

   - when talking to an old sender (i.e. without the off-by-one
 fix), send N-1 for N greater than 1.  Punt on N==1;

   - when talking to a fixed sender, ask to enable the capability,
 and send N as is (including N==1).

 - teaching the sending side to see if the new behaviour to fix
   off-by-one is asked by the requestor, and stop at the correct
   number of commits, not oversending one more.  Otherwise retain
   the old behaviour.

In the longer term, I think we should introduce a better deepening
mechanism.  Cf.

  http://thread.gmane.org/gmane.comp.version-control.git/212912/focus=212940

> Of course, one could using depth == 1 in this case to receive two
> commits and then drop one, but this would seem a bit pointless to me
> (especially if the commit below the tip is very different from the tip
> leading to a lot of useless data transfer).

--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 1/5] remote: trivial style cleanup

[Junio C Hamano]

Felipe Contreras  writes:

> Signed-off-by: Felipe Contreras 
> ---
>  remote.c | 3 +--
>  1 file changed, 1 insertion(+), 2 deletions(-)
>
> diff --git a/remote.c b/remote.c
> index 68eb99b..e71f66d 100644
> --- a/remote.c
> +++ b/remote.c
> @@ -1474,8 +1474,7 @@ struct branch *branch_get(const char *name)
>   ret->remote = remote_get(ret->remote_name);
>   if (ret->merge_nr) {
>   int i;
> - ret->merge = xcalloc(sizeof(*ret->merge),
> -  ret->merge_nr);
> + ret->merge = xcalloc(ret->merge_nr, 
> sizeof(*ret->merge));

Yeah, calloc is nmemb first and then size, so this makes sense.

>   for (i = 0; i < ret->merge_nr; i++) {
>   ret->merge[i] = xcalloc(1, 
> sizeof(**ret->merge));
>   ret->merge[i]->src = 
> xstrdup(ret->merge_name[i]);
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 4/5] test: improve rebase -q test

[Junio C Hamano]

Felipe Contreras  writes:

> Let's show the output so it's clear why it failed.
>
> Signed-off-by: Felipe Contreras 
> ---
>  t/t3400-rebase.sh | 1 +
>  1 file changed, 1 insertion(+)
>
> diff --git a/t/t3400-rebase.sh b/t/t3400-rebase.sh
> index b58fa1a..fb39531 100755
> --- a/t/t3400-rebase.sh
> +++ b/t/t3400-rebase.sh
> @@ -185,6 +185,7 @@ test_expect_success 'default to @{upstream} when upstream 
> arg is missing' '
>  test_expect_success 'rebase -q is quiet' '
>   git checkout -b quiet topic &&
>   git rebase -q master >output.out 2>&1 &&
> + cat output.out &&
>   test ! -s output.out
>  '

It is one thing to avoid squelching output that naturally comes out
of command being tested unnecessarily, so that "./t-*.sh -v"
output can be used for debugging.  I however am not sure if adding
"cat" to random places like this is a productive direction for us to
go in.

A more preferrable alternative may be adding something like this to
test-lib.sh and call it from here and elsewhere (there are about 50
places that do "test ! -s "), perhaps?

test_must_be_an_empty_file () {
if test -s "$1"
then
cat "$1"
false
fi
}

--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 1/3] cherry-pick: add support to copy notes

[Junio C Hamano]

Felipe Contreras  writes:

> Signed-off-by: Felipe Contreras 
> ---
>  builtin/revert.c  |   2 +
>  sequencer.c   | 136 
> --
>  sequencer.h   |   2 +
>  t/t3500-cherry.sh |  32 +
>  4 files changed, 169 insertions(+), 3 deletions(-)

"git cherry-pick" should help maintaining notes just like amend and
rebase, but how should this interact with notes.rewrite.,
where the command is capable of doing this without an explicit
option once you tell which notes need to be maintained?

Thomas Rast Cc'ed as he has been the primary force behind this line
of "notes" usability.

It probably is not sensible to carry over note from a reverted
commit to its revert, but I didn't immediately spot how the patch
does this only for cherry-pick but not for revert (the codepath in
do_pick_commit() is shared between them, no?).

> diff --git a/builtin/revert.c b/builtin/revert.c
> index 0e5ce71..b977124 100644
> --- a/builtin/revert.c
> +++ b/builtin/revert.c
> @@ -119,6 +119,7 @@ static void parse_args(int argc, const char **argv, 
> struct replay_opts *opts)
>   OPT_END(),
>   OPT_END(),
>   OPT_END(),
> + OPT_END(),
>   };
>  
>   if (opts->action == REPLAY_PICK) {
> @@ -129,6 +130,7 @@ static void parse_args(int argc, const char **argv, 
> struct replay_opts *opts)
>   OPT_BOOLEAN(0, "allow-empty-message", 
> &opts->allow_empty_message, N_("allow commits with empty messages")),
>   OPT_BOOLEAN(0, "keep-redundant-commits", 
> &opts->keep_redundant_commits, N_("keep redundant, empty commits")),
>   OPT_BOOLEAN(0, "skip-empty", &opts->skip_empty, 
> N_("skip empty commits")),
> + OPT_BOOLEAN(0, "copy-notes", &opts->copy_notes, 
> N_("copy notes")),
>   OPT_END(),
>   };
>   if (parse_options_concat(options, ARRAY_SIZE(options), 
> cp_extra))
> diff --git a/sequencer.c b/sequencer.c
> index edf141d..35a84ad 100644
> --- a/sequencer.c
> +++ b/sequencer.c
> @@ -20,6 +20,18 @@
>  const char sign_off_header[] = "Signed-off-by: ";
>  static const char cherry_picked_prefix[] = "(cherry picked from commit ";
>  
> +struct rewritten_list_item {
> + unsigned char from[20];
> + unsigned char to[20];
> +};
> +
> +struct rewritten_list {
> + struct rewritten_list_item *items;
> + unsigned int nr, alloc;
> +};
> +
> +static struct rewritten_list rewritten;
> +
>  static int is_rfc2822_line(const char *buf, int len)
>  {
>   int i;
> @@ -102,6 +114,40 @@ static int has_conforming_footer(struct strbuf *sb, 
> struct strbuf *sob,
>   return 1;
>  }
>  
> +static void add_rewritten(unsigned char *from, unsigned char *to)
> +{
> + struct rewritten_list_item *item;
> + if (rewritten.nr + 1 >= rewritten.alloc) {
> + rewritten.alloc += 32;
> + rewritten.items = xrealloc(rewritten.items, rewritten.alloc * 
> sizeof(*rewritten.items));
> + }
> + item = &rewritten.items[rewritten.nr];
> + hashcpy(item->from, from);
> + hashcpy(item->to, to);
> + rewritten.nr++;
> +}
> +
> +static void copy_notes(void)
> +{
> + unsigned char note_commit[20];
> + struct strbuf buf = STRBUF_INIT;
> + struct notes_tree *t = &default_notes_tree;
> + int i;
> +
> + init_notes(t, NULL, NULL, 0);
> +
> + for (i = 0; i < rewritten.nr; i++) {
> + struct rewritten_list_item *item = &rewritten.items[i];
> + copy_note(t, item->from, item->to, 0, NULL);
> + }
> +
> + strbuf_addstr(&buf, "Notes added by 'git cherry-pick'\n");
> + create_notes_commit(&default_notes_tree, NULL, &buf, note_commit);
> + strbuf_insert(&buf, 0, "notes: ", 7);
> + update_ref(buf.buf, t->ref, note_commit, NULL, 0, MSG_ON_ERR);
> + strbuf_release(&buf);
> +}
> +
>  static void remove_sequencer_state(void)
>  {
>   struct strbuf seq_dir = STRBUF_INIT;
> @@ -641,6 +687,14 @@ static int do_pick_commit(struct commit *commit, struct 
> replay_opts *opts)
>   res = run_git_commit(defmsg, opts, allow);
>   }
>  
> + if (!res && opts->copy_notes) {
> + unsigned char to[20];
> +
> + if (read_ref("HEAD", to))
> + goto leave;
> +
> + add_rewritten(commit->object.sha1, to);
> + }
>  leave:
>   free_message(&msg);
>   free(defmsg);
> @@ -786,6 +840,40 @@ static void read_populate_todo(struct commit_list 
> **todo_list,
>   die(_("Unusable instruction sheet: %s"), todo_file);
>  }
>  
> +static void read_populate_rewritten(void)
> +{
> + const char *rewritten_file = git_path(SEQ_REWR_FILE);
> + struct strbuf buf = STRBUF_INIT;
> + char *p;
> + int fd;
> +
> + fd = open(rewritten_file, O_RDONLY);
> + if (fd < 0)
> + return;
> + if (strbuf_read(&buf, fd, 0) < 0) {
> + close

where is the opensuse version for git?

[lud...@ludgerfrhrvonstetten.de]

Dear developer trying to find a version for suse 12.3 I was unable to 
locate a version.

Please advise.

downloading x-tags error messages.
error message
504-gateway time out.

Please advise what to do.
regards
Ludger
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 0/5] rebase: improve the keep-empty

[Martin von Zweigbergk]

Hi,

I think I have some patches at home that instead teach 'git am' the
--keep-empty flag. Does that make sense? It's been a while since I
looked at it, but I'll try to take a look tonight (PST).

Martin

On Tue, May 28, 2013 at 6:29 AM, Felipe Contreras
 wrote:
> Hi,
>
> I've been analyzing 'git rebase' and found that the --keep-empty option
> triggers a very very different behavior. Here's a bunch of patches that make 
> it
> behave like the 'am' does does for the most part.
>
> There's only a few minor changes, after which it might be possible to replace
> the whole 'am' mode to use cherr-pick instead.
>
> Felipe Contreras (5):
>   rebase: split the cherry-pick stuff
>   rebase: fix 'cherry' mode storage
>   rebase: fix sequence continuation
>   rebase: fix abort of cherry mode
>   rebase: fix cherry-pick invocations
>
>  .gitignore|  1 +
>  Makefile  |  1 +
>  git-rebase--am.sh | 65 
> ++-
>  git-rebase--cherry.sh | 55 +++
>  git-rebase.sh |  8 +++
>  5 files changed, 93 insertions(+), 37 deletions(-)
>  create mode 100644 git-rebase--cherry.sh
>
> --
> 1.8.3.rc3.312.g47657de
>
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 4/5] test: improve rebase -q test

[Jonathan Nieder]

Junio C Hamano wrote:

> A more preferrable alternative may be adding something like this to
> test-lib.sh and call it from here and elsewhere (there are about 50
> places that do "test ! -s "), perhaps?
>
> test_must_be_an_empty_file () {
> if test -s "$1"
> then
> cat "$1"
> false
> fi
> }

I generally just use the two-liner

>empty &&
test_cmp empty output

directly in cases like this.

Thanks,
Jonathan
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH] fix segfault with git log -c --follow

[Junio C Hamano]

Clemens Buchacher  writes:

> In diff_tree_combined we make a copy of diffopts. In
> try_to_follow_renames, called via diff_tree_sha1, we free and
> re-initialize diffopts->pathspec->items. Since we did not make a deep
> copy of diffopts in diff_tree_combined, the original diffopts does not
> get the update. By the time we return from diff_tree_combined,
> rev->diffopt->pathspec->items points to an invalid memory address. We
> get a segfault next time we try to access that pathspec.

I am not quite sure if I follow.  Do you mean

diff_tree_combined()
- makes a shallow copy of rev->diffopt
- calls diff_tree_sha1()
  diff_tree_sha1()
  - tries to follow rename and clobbers diffopt
- tries to use the shallow copy of original rev->diffopt
  that no longer is valid, which is a problem

I wonder, just like we force recursive and disable external on the
copy before we use it to call diff_tree_sha1(), if we should disable
follow-renames on it.  "--follow" is an option that is given to the
history traversal part and it should not play any role in getting
the pairwise diff with all parents diff_tree_combined() does.

Besides,

 - "--follow" hack lets us keep track of only one path; and

 - "-c" and "--cc" make sense only when dealing with a merge commit
   and the path in the child may have come from different path in
   parents,

so I am not sure if allowing combination of "--follow -c/--cc" makes
much sense in the first place.
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 4/5] test: improve rebase -q test

[Junio C Hamano]

Jonathan Nieder  writes:

> Junio C Hamano wrote:
>
>> A more preferrable alternative may be adding something like this to
>> test-lib.sh and call it from here and elsewhere (there are about 50
>> places that do "test ! -s "), perhaps?
>>
>> test_must_be_an_empty_file () {
>> if test -s "$1"
>> then
>> cat "$1"
>> false
>> fi
>> }
>
> I generally just use the two-liner
>
>   >empty &&
>   test_cmp empty output
>
> directly in cases like this.

That would work, too.
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: git hangs on pthread_join

[Jeff King]

On Thu, May 23, 2013 at 03:01:43PM +0200, Ian Kumlien wrote:

> git 1.8.2.1 is started by xinetd
> [...]
> I have found "git receive-pack"s that has been running for days/weeks
> without terminating
> 
> Attaching gdb and doing a trace results in:
> #0  0x003261207b35 in pthread_join () from /lib64/libpthread.so.0
> #1  0x004ce58b in finish_async ()
> #2  0x0045744b in cmd_receive_pack ()
> #3  0x00404851 in handle_internal_command ()
> #4  0x00404c9d in main ()

I recently fixed a deadlock that could happen in receive-pack when
clients hung up before sending a valid pack header. The fix is commit
49ecfa1, and it's in git v1.8.2.2.

The stack trace for the deadlock fixed by 49ecfa1 would have
unpack_with_sideband between #1 and #2 above, but it is entirely
possible that it is simply inlined in your build of git, depending on
the -O level of your build (it is a static function that is only called
from one place). So it seems likely that it is the culprit.

-Peff
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

1.8.3 - gitignore not being parsed correctly on OS X; regex support is broken?

[Misty De Meo]

Hi,

Gitignore parsing no longer seems to work properly in git 1.8.3.

One of my repositories has the following gitignore:

/*
!/.gitignore
!/Library/
!/CONTRIBUTING.md
!/README.md
!/SUPPORTERS.md
!/bin
/bin/*
!/bin/brew
!/share/man/man1/brew.1
.DS_Store
/Library/LinkedKegs
/Library/PinnedKegs
/Library/Taps
/Library/Formula/.gitignore

In 1.8.2.3 and earlier, this works as expected. However, in 1.8.3 I'm
seeing every file in /bin/ being marked as an untracked file.

I asked about this in #git, and was told that the culprit was the
regex support; apparently recompiling without regex support fixes the
specific gitignore issue. However, this doesn't seem to have been
reported anywhere on the mailing list that I can see. I was also told
that the issue is OS X-specific, and doesn't happen on other
platforms.

Thanks,
Misty De Meo
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 1/3] cherry-pick: add support to copy notes

[Thomas Rast]

Junio C Hamano  writes:

> Thomas Rast Cc'ed as he has been the primary force behind this line
> of "notes" usability.

Thanks for pointing this out to me.

> Felipe Contreras  writes:
>
>> Signed-off-by: Felipe Contreras 
>> ---
>>  builtin/revert.c  |   2 +
>>  sequencer.c   | 136 
>> --
>>  sequencer.h   |   2 +
>>  t/t3500-cherry.sh |  32 +
>>  4 files changed, 169 insertions(+), 3 deletions(-)
>
> "git cherry-pick" should help maintaining notes just like amend and
> rebase, but how should this interact with notes.rewrite.,
> where the command is capable of doing this without an explicit
> option once you tell which notes need to be maintained?

Since we already have the notes.rewrite. convention, it would
seem the obvious choice to line it up with the others.  The main
bikeshedding opportunity is whether this should be an exception and
default to false (all other commands default it to true).

Also: how does this interact with notes.rewriteRef and the corresponding
env vars?  Why?

How does it interact with 'cherry-pick -n' if this is done in sequence,
effectively squashing several commits (this use-case is actually
suggested by the manpage), if multiple source commits had notes?  Should
it respect notes.rewriteMode (and by default concatenate)?  (I don't
know if the sequencer state is expressive enough already to carry this
in a meaningful way across cherry-pick commands.)

A commit message and some docs would be a nice idea, too.

> diff --git a/t/t3500-cherry.sh b/t/t3500-cherry.sh
> index f038f34..79c1219 100755
> --- a/t/t3500-cherry.sh
> +++ b/t/t3500-cherry.sh

This file starts out with

  test_description='git cherry should detect patches integrated upstream

  This test cherry-picks one local change of two into master branch, and
  checks that git cherry only returns the second patch in the local branch
  '

So either your tests should go to a different file or the description
becomes stale and needs to be updated.

-- 
Thomas Rast
trast@{inf,student}.ethz.ch
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH v2] difftool --dir-diff: always use identical working tree file

[Junio C Hamano]

Kenichi Saita  writes:

> When deciding whether or not we should link a working tree file into
> the temporary right-hand directory for a directory diff, we
> currently behave differently in the --symlink and --no-symlink
> cases.  If using symlinks any identical files are linked across but
> with --no-symlink only files that contain unstaged changes are
> copied back into the working tree.

I may have missed an earlier discussion, but I do not follow the
last sentence.  The former part (i.e. symlinks) talks about what is
done to populate the temporary tree (i.e. everything is linked), but
the latter part (i.e. not symlinks) only talks about what is copied
back, i.e. it is not a contrast between the behaviour of symlink vs
no-symlink case wrt how the temporary tree is populated.

Confused...

> Change this so that identical files are copied back as well.  This
> is beneficial because it widens the set of circumstances in which we
> copy changes made by the user back into the working tree.

Ah, OK, you meant that the set of files we keep in @working_tree
array for later copying back are different between the two.

> Signed-off-by: Kenichi Saita 
> ---
>  git-difftool.perl   |9 ++---
>  t/t7800-difftool.sh |   19 +++
>  2 files changed, 21 insertions(+), 7 deletions(-)
>
> diff --git a/git-difftool.perl b/git-difftool.perl
> index 8a75205..e57d3d1 100755
> --- a/git-difftool.perl
> +++ b/git-difftool.perl
> @@ -85,13 +85,9 @@ sub exit_cleanup
>  
>  sub use_wt_file
>  {
> - my ($repo, $workdir, $file, $sha1, $symlinks) = @_;
> + my ($repo, $workdir, $file, $sha1) = @_;
>   my $null_sha1 = '0' x 40;
>  
> - if ($sha1 ne $null_sha1 and not $symlinks) {
> - return 0;
> - }
> -
>   if (! -e "$workdir/$file") {
>   # If the file doesn't exist in the working tree, we cannot
>   # use it.
> @@ -213,8 +209,7 @@ EOF
>  
>   if ($rmode ne $null_mode) {
>   my ($use, $wt_sha1) = use_wt_file($repo, $workdir,
> -   $dst_path, $rsha1,
> -   $symlinks);
> +   $dst_path, $rsha1);
>   if ($use) {
>   push @working_tree, $dst_path;
>   $wtindex .= "$rmode $wt_sha1\t$dst_path\0";
> diff --git a/t/t7800-difftool.sh b/t/t7800-difftool.sh
> index d46f041..2418528 100755
> --- a/t/t7800-difftool.sh
> +++ b/t/t7800-difftool.sh
> @@ -385,6 +385,25 @@ test_expect_success PERL,SYMLINKS 'difftool --dir-diff 
> --symlink without unstage
>   test_cmp actual expect
>  '
>  
> +write_script modify-right-file <<\EOF
> +echo "new content" >"$2/file"
> +EOF
> +
> +run_dir_diff_test 'difftool --dir-diff syncs worktree with unstaged change' '
> + test_when_finished git reset --hard &&
> + echo "orig content" >file &&
> + git difftool -d $symlinks --extcmd "$(pwd)/modify-right-file" branch &&
> + echo "new content" >expect &&
> + test_cmp expect file
> +'
> +
> +run_dir_diff_test 'difftool --dir-diff syncs worktree without unstaged 
> change' '
> + test_when_finished git reset --hard &&
> + git difftool -d $symlinks --extcmd "$(pwd)/modify-right-file" branch &&
> + echo "new content" >expect &&
> + test_cmp expect file
> +'
> +
>  write_script modify-file <<\EOF
>  echo "new content" >file
>  EOF
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH] git-remote-mediawiki: better error message when HTTP(S) access fails

[Jeff King]

On Thu, May 23, 2013 at 10:05:03PM +0200, Matthieu Moy wrote:

> My use-case is an invalid SSL certificate. Pulling from the wiki with a
> recent version of libwww-perl fails, and git-remote-mediawiki gave no
> clue about the reason. Give the mediawiki API detailed error message, and
> since it is not so informative, hint the user about an invalid SSL
> certificate on https:// urls.

This is definitely an improvement, but it seems like it just the tip of
the iceberg.

The call in get_mw_tracked_categories already uses die() with the MW
error code and detailed string. Good. The call you fix here prints a
guess to stderr and exits 1, and your patch improves that. But the call
in get_mw_first_pages does the same broken thing, and is not fixed.
Ditto for get_all_mediafiles. Other calls do not even seem to error
check the result at all, and assume the result is not undef (which I
suspect would produce a perl runtime error).

I wonder if we can do something like:

  our $mw_operation;
  $mediawiki->{config}->{on_error} = sub {
  my $err = "fatal: ";
  if (defined $mw_operation) {
  $err .= "unable to $mw_operation: ";
  }
  err .= $mediawiki->{error}->{details};
  die "$err\n";
  };

and then callers do not have to worry about error-checking at all. If
they want a nicer human-readable indication of where the error occured,
they can do:

  local $mw_operation = "get the list of remote wiki pages";
  my $mw_pages = $mediawiki->list(...);

-Peff
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH v2] difftool --dir-diff: always use identical working tree file

[John Keeping]

On Tue, May 28, 2013 at 11:06:13AM -0700, Junio C Hamano wrote:
> Kenichi Saita  writes:
> 
> > When deciding whether or not we should link a working tree file into
> > the temporary right-hand directory for a directory diff, we
> > currently behave differently in the --symlink and --no-symlink
> > cases.  If using symlinks any identical files are linked across but
> > with --no-symlink only files that contain unstaged changes are
> > copied back into the working tree.
> 
> I may have missed an earlier discussion, but I do not follow the
> last sentence.  The former part (i.e. symlinks) talks about what is
> done to populate the temporary tree (i.e. everything is linked), but
> the latter part (i.e. not symlinks) only talks about what is copied
> back, i.e. it is not a contrast between the behaviour of symlink vs
> no-symlink case wrt how the temporary tree is populated.
> 
> Confused...

Yeah, the commit message is still quite focused on the end effect of
copying files back.  But that's not what's being changed here.

In my suggested commit message I tried to make it clear that we're
changing when we decide to copy a file across to the temporary tree.
This has the beneficial (side-)effect of changing the set of files we
consider for copying back into the working tree after the diff tool has
been run.

> > Change this so that identical files are copied back as well.  This
> > is beneficial because it widens the set of circumstances in which we
> > copy changes made by the user back into the working tree.
> 
> Ah, OK, you meant that the set of files we keep in @working_tree
> array for later copying back are different between the two.
> 
> > Signed-off-by: Kenichi Saita 
> > ---
> >  git-difftool.perl   |9 ++---
> >  t/t7800-difftool.sh |   19 +++
> >  2 files changed, 21 insertions(+), 7 deletions(-)
> >
> > diff --git a/git-difftool.perl b/git-difftool.perl
> > index 8a75205..e57d3d1 100755
> > --- a/git-difftool.perl
> > +++ b/git-difftool.perl
> > @@ -85,13 +85,9 @@ sub exit_cleanup
> >  
> >  sub use_wt_file
> >  {
> > -   my ($repo, $workdir, $file, $sha1, $symlinks) = @_;
> > +   my ($repo, $workdir, $file, $sha1) = @_;
> > my $null_sha1 = '0' x 40;
> >  
> > -   if ($sha1 ne $null_sha1 and not $symlinks) {
> > -   return 0;
> > -   }
> > -
> > if (! -e "$workdir/$file") {
> > # If the file doesn't exist in the working tree, we cannot
> > # use it.
> > @@ -213,8 +209,7 @@ EOF
> >  
> > if ($rmode ne $null_mode) {
> > my ($use, $wt_sha1) = use_wt_file($repo, $workdir,
> > - $dst_path, $rsha1,
> > - $symlinks);
> > + $dst_path, $rsha1);
> > if ($use) {
> > push @working_tree, $dst_path;
> > $wtindex .= "$rmode $wt_sha1\t$dst_path\0";
> > diff --git a/t/t7800-difftool.sh b/t/t7800-difftool.sh
> > index d46f041..2418528 100755
> > --- a/t/t7800-difftool.sh
> > +++ b/t/t7800-difftool.sh
> > @@ -385,6 +385,25 @@ test_expect_success PERL,SYMLINKS 'difftool --dir-diff 
> > --symlink without unstage
> > test_cmp actual expect
> >  '
> >  
> > +write_script modify-right-file <<\EOF
> > +echo "new content" >"$2/file"
> > +EOF
> > +
> > +run_dir_diff_test 'difftool --dir-diff syncs worktree with unstaged 
> > change' '
> > +   test_when_finished git reset --hard &&
> > +   echo "orig content" >file &&
> > +   git difftool -d $symlinks --extcmd "$(pwd)/modify-right-file" branch &&
> > +   echo "new content" >expect &&
> > +   test_cmp expect file
> > +'
> > +
> > +run_dir_diff_test 'difftool --dir-diff syncs worktree without unstaged 
> > change' '
> > +   test_when_finished git reset --hard &&
> > +   git difftool -d $symlinks --extcmd "$(pwd)/modify-right-file" branch &&
> > +   echo "new content" >expect &&
> > +   test_cmp expect file
> > +'
> > +
> >  write_script modify-file <<\EOF
> >  echo "new content" >file
> >  EOF
> --
> To unsubscribe from this list: send the line "unsubscribe git" in
> the body of a message to majord...@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH v2] difftool --dir-diff: always use identical working tree file

[Junio C Hamano]

John Keeping  writes:

> Yeah, the commit message is still quite focused on the end effect of
> copying files back.  But that's not what's being changed here.
>
> In my suggested commit message I tried to make it clear that we're
> changing when we decide to copy a file across to the temporary tree.
> This has the beneficial (side-)effect of changing the set of files we
> consider for copying back into the working tree after the diff tool has
> been run.

I actually think the effect of copying files back _is_ the primary
motivation of this change, and stressing that end effect is a much
better description.  After all, if the working tree files do not
have any difference from the RHS of the comparison, copying from the
working tree and stuffing the $rsha1 to the RHS temporary index and
running "checkout -f" should produce identical temporary directory
for the user to start viewing.

The _only_ difference in behaviour before and after this patch that
matters to the end user is if that path is in @working_tree, which
is returned to @worktree of dir_diff sub to be later copied back,
no?  I would view this change as a mere means, an implementation
detail, to achieve that end of stuffing more paths in the @worktree
array.

Perhaps

difftool --dir-diff: allow changing any clean working tree file

The temporary directory prepared by "difftool --dir-diff" to
show the result of a change can be modified by the user via
the tree diff program, and we try hard not to lose changes
to them after tree diff program returns to us.

However, the set of files to be copied back is computed
differently between --symlinks and --no-symlinks modes.  The
former checks all paths that start out as identical to the
working tree file, while the latter checks paths that
already had a local modification in the working tree,
allowing changes made in the tree diff program to paths that
did not have any local change to be lost.

or something.  This invites a few questions, though.

 - By allowing these files in the temporary directory to be
   modified, aren't we making the user's life harder by forcing them
   to handle "working tree file was already modified, made different
   changes in the temporary directory, now these changes need to be
   consolidated" case?

 - When comparing two revisions, e.g. "--dir-diff HEAD^^ HEAD^",
   that checks out (via $rsha1 to "checkout -f" codepath) a blob
   that does not match what is in the working tree of HEAD to the
   temporary directory, we still allow modifications to the copy in
   the temporary directory, but what can the user do with these
   changes that are _not_ based on HEAD, short of checking out HEAD^
   and apply the difference first?

I still cannot shake this nagging feeling that giving a writable
temporary directory might have been a mistake in the first place.
Perhaps it may be a better design to make the ones that the user
shouldn't touch (or will lead to the above confusion) read-only,
while the ones that match the working tree read-write?
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH v2] difftool --dir-diff: always use identical working tree file

[John Keeping]

On Tue, May 28, 2013 at 11:57:08AM -0700, Junio C Hamano wrote:
> John Keeping  writes:
> 
> > Yeah, the commit message is still quite focused on the end effect of
> > copying files back.  But that's not what's being changed here.
> >
> > In my suggested commit message I tried to make it clear that we're
> > changing when we decide to copy a file across to the temporary tree.
> > This has the beneficial (side-)effect of changing the set of files we
> > consider for copying back into the working tree after the diff tool has
> > been run.
> 
> I actually think the effect of copying files back _is_ the primary
> motivation of this change, and stressing that end effect is a much
> better description.  After all, if the working tree files do not
> have any difference from the RHS of the comparison, copying from the
> working tree and stuffing the $rsha1 to the RHS temporary index and
> running "checkout -f" should produce identical temporary directory
> for the user to start viewing.
> 
> The _only_ difference in behaviour before and after this patch that
> matters to the end user is if that path is in @working_tree, which
> is returned to @worktree of dir_diff sub to be later copied back,
> no?  I would view this change as a mere means, an implementation
> detail, to achieve that end of stuffing more paths in the @worktree
> array.

I agree with this, but like you I found it confusing that the patch
touched code seemingly unrelated to copying files back.  I went toward
describing the patch more literally and giving the motivation in the
final paragraph.  Your message below is better, but I think it needs to
say that the set of files considered for copying back is the set that is
copied across to begin with.

> Perhaps
> 
>   difftool --dir-diff: allow changing any clean working tree file
> 
>   The temporary directory prepared by "difftool --dir-diff" to
>   show the result of a change can be modified by the user via
>   the tree diff program, and we try hard not to lose changes
>   to them after tree diff program returns to us.
> 
> However, the set of files to be copied back is computed
>   differently between --symlinks and --no-symlinks modes.  The
>   former checks all paths that start out as identical to the
>   working tree file, while the latter checks paths that
>   already had a local modification in the working tree,
>   allowing changes made in the tree diff program to paths that
>   did not have any local change to be lost.
> 
> or something.  This invites a few questions, though.
> 
>  - By allowing these files in the temporary directory to be
>modified, aren't we making the user's life harder by forcing them
>to handle "working tree file was already modified, made different
>changes in the temporary directory, now these changes need to be
>consolidated" case?
> 
>  - When comparing two revisions, e.g. "--dir-diff HEAD^^ HEAD^",
>that checks out (via $rsha1 to "checkout -f" codepath) a blob
>that does not match what is in the working tree of HEAD to the
>temporary directory, we still allow modifications to the copy in
>the temporary directory, but what can the user do with these
>changes that are _not_ based on HEAD, short of checking out HEAD^
>and apply the difference first?
> 
> I still cannot shake this nagging feeling that giving a writable
> temporary directory might have been a mistake in the first place.
> Perhaps it may be a better design to make the ones that the user
> shouldn't touch (or will lead to the above confusion) read-only,
> while the ones that match the working tree read-write?

My ideal scenario would be that we only allow users to edit files when
they are comparing against the working tree, but that would require
git-difftool to fully understand all git-diff options since it just
passes through any it doesn't recognise.  I don't think there's an easy
way to do that, which leaves us with this confusing situation.
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH v2] difftool --dir-diff: always use identical working tree file

[Junio C Hamano]

John Keeping  writes:

>>  - When comparing two revisions, e.g. "--dir-diff HEAD^^ HEAD^",
>>that checks out (via $rsha1 to "checkout -f" codepath) a blob
>>that does not match what is in the working tree of HEAD to the
>>temporary directory, we still allow modifications to the copy in
>>the temporary directory, but what can the user do with these
>>changes that are _not_ based on HEAD, short of checking out HEAD^
>>and apply the difference first?
>> 
>> I still cannot shake this nagging feeling that giving a writable
>> temporary directory might have been a mistake in the first place.
>> Perhaps it may be a better design to make the ones that the user
>> shouldn't touch (or will lead to the above confusion) read-only,
>> while the ones that match the working tree read-write?
>
> My ideal scenario would be that we only allow users to edit files when
> they are comparing against the working tree, but that would require
> git-difftool to fully understand all git-diff options since it just
> passes through any it doesn't recognise.  I don't think there's an easy
> way to do that, which leaves us with this confusing situation.

Not necessarily.

Let's assume that changing files in "diff" tool is a sensible thing
to do, as long as we make sure such a change is not lost (which I
may not 100% agree with, but let's put it aside for now).

When you are viewing a file F in "--dir-diff HEAD^^ HEAD^", if there
is no change for F in between HEAD^ and HEAD and you notice a typo
that may or may not be related to the differences between HEAD^^ and
HEAD^, it would be tempting to fix that right there.  And as long as
F in the working tree matches that of HEAD^ and the modification you
make in the temporary directory gets copied back to the working tree,
your typofix will end up to be in the working tree.

Which I _think_ is what people, who want to change files in "diff"
tool, want to do.

Of course, your working tree may have been in the middle of doing
something entirely different and you may have to "add [-p]" to
separate such a typofix with other changes you were working on, but
that is a separate issue.

And for that to work, the only think you need is "does the blob we
show on the RHS temporary tree match what is in the working tree?"
check.  You do not need to know or care if you are comparing two old
revisions, etc.
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 2/4] remote-helpers: rename tests

[Junio C Hamano]

Felipe Contreras  writes:

> The .t extension is more standard for sharness tests.
>
> Signed-off-by: Felipe Contreras 
> ---

Is that "sharness" test the sh script testsuite forked from our
testsuite?

I do not see how it makes sense to copy how they deviate from us
back to our codebase, especially if we plan to eventually move some
of these tests out of contrib/ area, but even without such a plan in
the future.



>  contrib/remote-helpers/Makefile| 2 +-
>  contrib/remote-helpers/{test-bzr.sh => test-bzr.t} | 0
>  contrib/remote-helpers/{test-hg-bidi.sh => test-hg-bidi.t} | 0
>  contrib/remote-helpers/{test-hg-hg-git.sh => test-hg-hg-git.t} | 0
>  contrib/remote-helpers/{test-hg.sh => test-hg.t}   | 0
>  5 files changed, 1 insertion(+), 1 deletion(-)
>  rename contrib/remote-helpers/{test-bzr.sh => test-bzr.t} (100%)
>  rename contrib/remote-helpers/{test-hg-bidi.sh => test-hg-bidi.t} (100%)
>  rename contrib/remote-helpers/{test-hg-hg-git.sh => test-hg-hg-git.t} (100%)
>  rename contrib/remote-helpers/{test-hg.sh => test-hg.t} (100%)
>
> diff --git a/contrib/remote-helpers/Makefile b/contrib/remote-helpers/Makefile
> index d9b3515..2c91ec6 100644
> --- a/contrib/remote-helpers/Makefile
> +++ b/contrib/remote-helpers/Makefile
> @@ -1,4 +1,4 @@
> -TESTS := $(wildcard test*.sh)
> +TESTS := $(wildcard test-*.t)
>  SCRIPTS := $(wildcard git-remote-*.py)
>  
>  export T := $(addprefix $(CURDIR)/,$(TESTS))
> diff --git a/contrib/remote-helpers/test-bzr.sh 
> b/contrib/remote-helpers/test-bzr.t
> similarity index 100%
> rename from contrib/remote-helpers/test-bzr.sh
> rename to contrib/remote-helpers/test-bzr.t
> diff --git a/contrib/remote-helpers/test-hg-bidi.sh 
> b/contrib/remote-helpers/test-hg-bidi.t
> similarity index 100%
> rename from contrib/remote-helpers/test-hg-bidi.sh
> rename to contrib/remote-helpers/test-hg-bidi.t
> diff --git a/contrib/remote-helpers/test-hg-hg-git.sh 
> b/contrib/remote-helpers/test-hg-hg-git.t
> similarity index 100%
> rename from contrib/remote-helpers/test-hg-hg-git.sh
> rename to contrib/remote-helpers/test-hg-hg-git.t
> diff --git a/contrib/remote-helpers/test-hg.sh 
> b/contrib/remote-helpers/test-hg.t
> similarity index 100%
> rename from contrib/remote-helpers/test-hg.sh
> rename to contrib/remote-helpers/test-hg.t
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 4/4] remote-helpers: add exec-path links

[Junio C Hamano]

Felipe Contreras  writes:

> This way we don't have to modify the PATH ourselves and it's easier to
> test without 'make'.
>
> Signed-off-by: Felipe Contreras 
> ---
>  .gitignore  |  2 ++
>  contrib/remote-helpers/Makefile | 13 +
>  2 files changed, 11 insertions(+), 4 deletions(-)

If we teach the top-level .gitignore that some contrib/ stuff might
contaminate the top-level directory (which is a good idea), we
should also teach the top-level Makefile how to get rid of the built
cruft upon "make clean", no?

> diff --git a/.gitignore b/.gitignore
> index 10aee94..a79b412 100644
> --- a/.gitignore
> +++ b/.gitignore
> @@ -119,11 +119,13 @@
>  /git-reflog
>  /git-relink
>  /git-remote
> +/git-remote-bzr
>  /git-remote-http
>  /git-remote-https
>  /git-remote-ftp
>  /git-remote-ftps
>  /git-remote-fd
> +/git-remote-hg
>  /git-remote-ext
>  /git-remote-testgit
>  /git-remote-testpy
> diff --git a/contrib/remote-helpers/Makefile b/contrib/remote-helpers/Makefile
> index 55abf0b..98150b4 100644
> --- a/contrib/remote-helpers/Makefile
> +++ b/contrib/remote-helpers/Makefile
> @@ -1,9 +1,9 @@
>  TESTS := $(wildcard test-*.t)
>  SCRIPTS := $(wildcard git-remote-*.py)
> +LINKS := $(addprefix ../../,$(patsubst %.py,%,$(SCRIPTS)))
>  
>  export T := $(addprefix $(CURDIR)/,$(TESTS))
>  export MAKE := $(MAKE) -e
> -export PATH := $(CURDIR):$(PATH)
>  export TEST_LINT := test-lint-executable test-lint-shell-syntax
>  export TEST_DIRECTORY := $(CURDIR)/../../t
>  
> @@ -15,10 +15,15 @@ all: $(SCRIPTS)
>  install:
>   $(MAKE) -C ../.. install-python-script
>  
> -test: all
> +links: all $(LINKS)
> +
> +test: links
>   $(MAKE) -C ../../t $@
>  
> -$(TESTS): all
> +$(LINKS):
> + ln -sf contrib/remote-helpers/$(notdir $@) ../..
> +
> +$(TESTS): links
>   $(MAKE) -C ../../t $(CURDIR)/$@
>  
> -.PHONY: all install test $(TESTS)
> +.PHONY: all install test links $(TESTS)
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: git clone does not understand insteadOf URLs

[Junio C Hamano]

Andreas Schwab  writes:

> Gioele Barabucci  writes:
>
>> Simple, I keep all my projects on the same server, so I would like to
>> refer to that server + path using 'remote-repo'.
>>
>> "git+ssh://git.example.org//users/gioele/projects" insteadOf "remote-repo"
>
> You can use "remote-repo:" instead.

That "with colon" is pretty much how it was designed to be used, but
the code did not insist that the insteadof has at least the scheme
part, and by mistake allowed something like this

[url "k.org:/pub/scm/git"] insteadOf = g

to be used with "g/foo" to expand to "k.org:/pub/scm/git/foo".

But if "git fetch g/foo" with such a configuration happily goes to
"k.org:/pub/scm/git/foo" without problems, then I think it is very
reasonable to delay the "is that a local filename?" short-cut "git
clone" does and let the insteadof URL expand take place before it.
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: contrib/git-normal-to-bare.sh

[Junio C Hamano]

Zenaan Harkness  writes:

> I needed this quite a bit in the last few days, basic script but
> serves my need. I think it would be useful for other beginners if in
> $git/contrib/ source dir.
>
> Just a start to a basic script. Needs more tests etc, but it's enough
> to get newbies (like me) off to a reasonable start. Handles multiple
> input dirs.
>
> PLEASE CC me, as I am not subscribed.
>
> (some SMTP server rejected attachment, so pasting below instead)
>
> Thanks,
> Zenaan
>
>
> #!/bin/bash

I do not think you need (nor used) any bash-ism in this script.
Saying "#!/bin/sh" here is cleaner.

>
> # Change one or more normal repos into bare repos:
> # See also 
> https://git.wiki.kernel.org/index.php/GitFaq#How_do_I_make_existing_non-bare_repository_bare.3F
>
> for i in "$@"; do

for i
do

You do not have to say 'in "$@"'; it is implied.

>echo; echo "--"
>echo Processing $i

Forgot to dq?

>
>repo="$i"
>repo="`basename $i`"
>tmp_repo="${repo}.git"
># Insert here: may be exit if any spaces in repo fqn
># Insert here: check for non-existent repo/.git dir
># Insert here: check that we are not inside the repo
># Insert here: add exit/do-nothing if fail to mv dirs etc
>
>mv $repo/.git $tmp_repo

Forgot to dq?  i.e.

mv "$repo/.git" "$tmp_repo"

The same for all the variable references in the remainder of the
script.

More importantly, "mv" would fail if $repo is given as a full
pathname elsewhere in the filesystem that is different from your
current directory where you create $tmp_repo.

>git --git-dir=$tmp_repo config core.bare true
>mv $repo ${repo}.bak
>mv $tmp_repo $repo
> done
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: What's cooking in git.git (May 2013, #04; Wed, 15)

[Junio C Hamano]

Duy Nguyen  writes:

> Point taken. I guess the message would be something like this?
>
> Refname '%.*s' is ignored. It may be created by mistake.
>
> Or should we be more elaborate?

I dunno; with s/may be/may have been/, I think it is better than
"refname is ambiguous".
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: 1.8.3 - gitignore not being parsed correctly on OS X; regex support is broken?

[Øystein Walle]

Misty De Meo  brew.sh> writes:

> 
> Hi,
> 
> Gitignore parsing no longer seems to work properly in git 1.8.3.
> 
> One of my repositories has the following gitignore:
> 
> /*
> !/.gitignore
> !/Library/
> !/CONTRIBUTING.md
> !/README.md
> !/SUPPORTERS.md
> !/bin
> /bin/*
> !/bin/brew
> !/share/man/man1/brew.1
> .DS_Store
> /Library/LinkedKegs
> /Library/PinnedKegs
> /Library/Taps
> /Library/Formula/.gitignore
> 
> In 1.8.2.3 and earlier, this works as expected. However, in 1.8.3 I'm
> seeing every file in /bin/ being marked as an untracked file.
> 
> I asked about this in #git, and was told that the culprit was the
> regex support; apparently recompiling without regex support fixes the
> specific gitignore issue. However, this doesn't seem to have been
> reported anywhere on the mailing list that I can see. I was also told
> that the issue is OS X-specific, and doesn't happen on other
> platforms.
> 
> Thanks,
> Misty De Meo
> 

I see a similar problem using e.g. the following .gitignore to exclude
everything except C source files and header files:

*
!*/
!*.c
!*.h

In Git 1.8.3 'git status' will show other files as untracked while in
Git 1.8.2.3 I don't have that problem. I bisected to find that the
offending commit is v1.8.2.1-402-g95c6f27. 

I am not on OSX, however, but on Linux (Ubuntu 12.04 and RHEL 5.8) so
this may be a separate issue. I've also gotten the impression that this
is intentional. In any case I cannot create a .gitignore that achieves
the same for both older and newer versions of Git.

Best regards,
Øystein Walle


--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

What's cooking in git.git (May 2013, #08; Tue, 28)

[Junio C Hamano]

Here are the topics that have been cooking.  Commits prefixed with
'-' are only in 'pu' (proposed updates) while commits prefixed with
'+' are in 'next'.

There are topics that are still in 'next', not because they needed
more testing, but only because we already were in the -rc period.
Now that 1.8.3 is out, we will start merging them to 'master' (and
some to 'maint' for 1.8.3.1) probably tomorrow (i.e. after waiting
for a few days to make sure we can have a clean start of the post
1.8.3 maint branch with brown paper bag fixes and nothing else).

The post 1.8.3 cycle will start early next month, at which time the
'next' branch will be rewound and rebuilt.  Until then, I expect
that my patch queue may stay slow and leaky while I take a bit of
break.

You can find the changes described here in the integration branches
of the repositories listed at

http://git-blame.blogspot.com/p/git-public-repositories.html

--
[New Topics]

* nd/urls-doc-no-file-hyperlink-fix (2013-05-24) 1 commit
 - urls.txt: avoid auto converting to hyperlink

 Will merge to 'next'.


* cb/log-follow-with-combined (2013-05-28) 1 commit
 - fix segfault with git log -c --follow

 Will merge to 'next'.


* fc/cleanups (2013-05-28) 3 commits
 - test: rebase: fix --interactive test
 - test: trivial cleanups
 - remote: trivial style cleanup

 Will merge to 'next'.


* fc/makefile (2013-05-26) 5 commits
 - build: do not install git-remote-testpy
 - build: add NO_INSTALL variable
 - build: cleanup using $<
 - build: cleanup using $^
 - build: trivial simplification
 (this branch is used by fc/remote-helpers-use-specified-python.)

 Will merge to 'next'.


* fc/remote-helpers-use-specified-python (2013-05-28) 4 commits
 - remote-helpers: add exec-path links
 - remote-helpers: allow direct test execution
 - remote-helpers: rename tests
 - remote-helpers: generate scripts
 (this branch uses fc/makefile.)

 I do not particularly think the second from the bottom is a good
 change, but it takes the remainder of the series hostage.
 Will hopefully be rerolled without it.


* fc/send-email-chainreplyto-warning (2013-05-28) 1 commit
 - send-email: remove warning about unset chainreplyto

 An overdue removal od "behaviour changed at 1.7.0; if you were
 living in a cave, here is what you can adjust to it" message.

 Will merge to 'next'.


* ks/difftool-dirdiff-copy-all (2013-05-28) 1 commit
 - difftool --dir-diff: always use identical working tree file

 "difftool --dir-diff" populates a temporary directory with files,
 inviting the user to modify them in place, but in some cases we
 weren't expecting the user to modify them and did not take the
 change back to the working tree.

 The log message has room for clarification, but the change looked
 sane.


* nd/prune-packed-dryrun-verbose (2013-05-28) 1 commit
 - prune-packed: avoid implying "1" is DRY_RUN in prune_packed_objects()

 Will merge to 'next'.


* rj/mingw-compat-st-mode-bits (2013-05-28) 1 commit
 - path: Fix a sparse warning

 Will merge to 'next'.


* rs/commit-m-no-edit (2013-05-28) 1 commit
 - commit: don't start editor if empty message is given with -m

 Will merge to 'next'.


* xq/credential-osxkeychain (2013-05-28) 1 commit
 - credential-osxkeychain: support more protocols

 Will merge to 'next'.


* jc/core-checkstat (2013-05-06) 1 commit
 - deprecate core.statinfo at Git 2.0 boundary
 (this branch is used by jc/core-checkstat-2.0.)

 Will merge to 'next'.

--
[Stalled]

* rr/rebase-autostash (2013-05-12) 7 commits
 - rebase: implement --[no-]autostash and rebase.autostash
 - rebase --merge: return control to caller, for housekeeping
 - rebase -i: return control to caller, for housekeeping
 - am: return control to caller, for housekeeping
 - rebase: prepare to do generic housekeeping
 - rebase -i: don't error out if $state_dir already exists
 - am: tighten a conditional that checks for $dotest

 This is from v3, but after a "Fixed" message in $gmane/224111 we
 haven't seen a reroll yet.  Also there was an attempt for a
 follow-up, but it was never completed.


* rj/mingw-cygwin (2013-05-08) 2 commits
 - cygwin: Remove the CYGWIN_V15_WIN32API build variable
 - mingw: rename WIN32 cpp macro to GIT_WINDOWS_NATIVE

 Update build for Cygwin 1.[57].  There seems to have been some
 discussion but did anything concrete come out of it???


* mh/multimail (2013-04-21) 1 commit
 - git-multimail: a replacement for post-receive-email

 Waiting for the initial history to pull from.
 $gmane/223564


* jc/format-patch (2013-04-22) 2 commits
 - format-patch: --inline-single
 - format-patch: rename "no_inline" field

 A new option to send a single patch to the standard output to be
 appended at the bottom of a message.  I personally have no need for
 this, but it was easy enough to cobble together.  Tests, docs and
 stripping out more MIMEy stuff are left as exercises to interested
 parties.

 Not ready 

Re: 1.8.3 - gitignore not being parsed correctly on OS X; regex support is broken?

[Junio C Hamano]

Øystein Walle  writes:

> Misty De Meo  brew.sh> writes:
>
>> 
>> Hi,
>> 
>> Gitignore parsing no longer seems to work properly in git 1.8.3.
>> 
>> One of my repositories has the following gitignore:
>> 
>> /*
>> !/.gitignore
>> !/Library/
>> !/CONTRIBUTING.md
>> !/README.md
>> !/SUPPORTERS.md
>> !/bin
>> /bin/*
>> !/bin/brew
>> !/share/man/man1/brew.1
>> .DS_Store
>> /Library/LinkedKegs
>> /Library/PinnedKegs
>> /Library/Taps
>> /Library/Formula/.gitignore
>> 
>> In 1.8.2.3 and earlier, this works as expected. However, in 1.8.3 I'm
>> seeing every file in /bin/ being marked as an untracked file.
>> 
>> I asked about this in #git, and was told that the culprit was the
>> regex support; apparently recompiling without regex support fixes the
>> specific gitignore issue. However, this doesn't seem to have been
>> reported anywhere on the mailing list that I can see. I was also told
>> that the issue is OS X-specific, and doesn't happen on other
>> platforms.
>> 
>> Thanks,
>> Misty De Meo
>> 
>
> I see a similar problem using e.g. the following .gitignore to exclude
> everything except C source files and header files:
>
> *
> !*/
> !*.c
> !*.h
>
> In Git 1.8.3 'git status' will show other files as untracked while in
> Git 1.8.2.3 I don't have that problem. I bisected to find that the
> offending commit is v1.8.2.1-402-g95c6f27. 
>
> I am not on OSX, however, but on Linux (Ubuntu 12.04 and RHEL 5.8) so
> this may be a separate issue. I've also gotten the impression that this
> is intentional. In any case I cannot create a .gitignore that achieves
> the same for both older and newer versions of Git.

Thanks; 95c6f27 is from Karsten and I think Duy is the most clueful
one in this area in the discussion that had this change; I am
forwarding this message to them.

Cf. http://thread.gmane.org/gmane.comp.version-control.git/218440/focus=221289

--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 1/5] rebase: split the cherry-pick stuff

[Junio C Hamano]

Felipe Contreras  writes:

> They do something completely different from 'git am', it belongs in a
> different file.

I would prefer to see it called --cherry-pick, not --cherry, as they
are different commands (the latter may be useful when deciding which
one to use the former).

>
> Signed-off-by: Felipe Contreras 
> ---
>  .gitignore|  1 +
>  Makefile  |  1 +
>  git-rebase--am.sh | 65 
> ++-
>  git-rebase--cherry.sh | 34 +++
>  git-rebase.sh |  4 
>  5 files changed, 68 insertions(+), 37 deletions(-)
>  create mode 100644 git-rebase--cherry.sh
>
> diff --git a/.gitignore b/.gitignore
> index 6669bf0..284fc8f 100644
> --- a/.gitignore
> +++ b/.gitignore
> @@ -113,6 +113,7 @@
>  /git-read-tree
>  /git-rebase
>  /git-rebase--am
> +/git-rebase--cherry
>  /git-rebase--interactive
>  /git-rebase--merge
>  /git-receive-pack
> diff --git a/Makefile b/Makefile
> index 0f931a2..a3cd4bc 100644
> --- a/Makefile
> +++ b/Makefile
> @@ -469,6 +469,7 @@ SCRIPT_SH += git-web--browse.sh
>  SCRIPT_LIB += git-mergetool--lib
>  SCRIPT_LIB += git-parse-remote
>  SCRIPT_LIB += git-rebase--am
> +SCRIPT_LIB += git-rebase--cherry
>  SCRIPT_LIB += git-rebase--interactive
>  SCRIPT_LIB += git-rebase--merge
>  SCRIPT_LIB += git-sh-setup
> diff --git a/git-rebase--am.sh b/git-rebase--am.sh
> index f84854f..ee1b1b9 100644
> --- a/git-rebase--am.sh
> +++ b/git-rebase--am.sh
> @@ -19,52 +19,43 @@ esac
>  test -n "$rebase_root" && root_flag=--root
>  
>  ret=0
> +rm -f "$GIT_DIR/rebased-patches"
>  
> - cat >&2 <<-EOF
> +git format-patch -k --stdout --full-index --ignore-if-in-upstream \
> + --src-prefix=a/ --dst-prefix=b/ --no-renames --no-cover-letter \
> + $root_flag "$revisions" >"$GIT_DIR/rebased-patches"
> +ret=$?
>  
> +if test 0 != $ret
> +then
> + rm -f "$GIT_DIR/rebased-patches"
> + case "$head_name" in
> + refs/heads/*)
> + git checkout -q "$head_name"
> + ;;
> + *)
> + git checkout -q "$orig_head"
> + ;;
> + esac
>  
> - $revisions
> + cat >&2 <<-EOF
>  
> + git encountered an error while preparing the patches to replay
> + these revisions:
>  
> + $revisions
>  
> + As a result, git cannot rebase them.
> + EOF
> + exit $?
>  fi
>  
> +git am $git_am_opt --rebasing --resolvemsg="$resolvemsg" 
> <"$GIT_DIR/rebased-patches"
> +ret=$?
> +
> +rm -f "$GIT_DIR/rebased-patches"
> +
>  if test 0 != $ret
>  then
>   test -d "$state_dir" && write_basic_state

OK, this part seems a straightforward reindentation of "else" side.

> diff --git a/git-rebase--cherry.sh b/git-rebase--cherry.sh
> new file mode 100644
> index 000..cbf80f9
> --- /dev/null
> +++ b/git-rebase--cherry.sh
> @@ -0,0 +1,34 @@
> +#!/bin/sh
> +#
> +# Copyright (c) 2010 Junio C Hamano.
> +#
> +
> +case "$action" in
> +continue)
> + git am --resolved --resolvemsg="$resolvemsg" &&
> + move_to_original_branch
> + exit
> + ;;
> +skip)
> + git am --skip --resolvemsg="$resolvemsg" &&
> + move_to_original_branch
> + exit
> + ;;
> +esac
> +
> +test -n "$rebase_root" && root_flag=--root

Hmm, do we want to duplicate this part between the two?  Perhaps it
will be unified in a later patch---I'll read on.

> +# we have to do this the hard way.  git format-patch completely squashes
> +# empty commits and even if it didn't the format doesn't really lend
> +# itself well to recording empty patches.  fortunately, cherry-pick
> +# makes this easy
> +git cherry-pick --allow-empty "$revisions"
> +ret=$?
> +
> +if test 0 != $ret
> +then
> + test -d "$state_dir" && write_basic_state
> + exit $ret
> +fi
> +
> +move_to_original_branch

OK.  This is the "then" side of the original, with clean-up phase
copied.

> diff --git a/git-rebase.sh b/git-rebase.sh
> index 2c692c3..2754985 100755
> --- a/git-rebase.sh
> +++ b/git-rebase.sh
> @@ -379,6 +379,10 @@ elif test -n "$do_merge"
>  then
>   type=merge
>   state_dir="$merge_dir"
> +elif test -n "$keep_empty"
> +then
> + type=cherry
> + state_dir="$apply_dir"
>  else
>   type=am
>   state_dir="$apply_dir"

--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 2/5] rebase: fix 'cherry' mode storage

[Junio C Hamano]

Felipe Contreras  writes:

> We don't use the 'rebase-apply'.

s/.$/; we will use rebase-merge instead./ I think.

> Signed-off-by: Felipe Contreras 
> ---
>  git-rebase--cherry.sh | 4 
>  git-rebase.sh | 5 -
>  2 files changed, 8 insertions(+), 1 deletion(-)
>
> diff --git a/git-rebase--cherry.sh b/git-rebase--cherry.sh
> index cbf80f9..ab1f8b7 100644
> --- a/git-rebase--cherry.sh
> +++ b/git-rebase--cherry.sh
> @@ -18,6 +18,9 @@ esac
>  
>  test -n "$rebase_root" && root_flag=--root
>  
> +mkdir "$state_dir" || die "Could not create temporary $state_dir"
> +: > "$state_dir"/cherry || die "Could not mark as cherry"

Style:

>"$state_dir/cherry-pick"

I am not sure if the user _cares_ that internally we use cherry-pick
when s/he asks us to do a keep-empty, and I suspect "mark as cherry"
incomprehensible.  I do not even know what is going on at this point
(yet).

I _suspect_ that you are using cherry-pick when --keep-empty is given,
so it might even make sense to talk in the end-users' terms,
i.e. call this "$state_dir/keep-empty" (I am not sure if calling
this split half "git-rebase--keep-empty.sh" makes sense, yet).

> +
>  # we have to do this the hard way.  git format-patch completely squashes
>  # empty commits and even if it didn't the format doesn't really lend
>  # itself well to recording empty patches.  fortunately, cherry-pick
> @@ -32,3 +35,4 @@ then
>  fi
>  
>  move_to_original_branch
> +rm -rf "$state_dir"
> diff --git a/git-rebase.sh b/git-rebase.sh
> index 2754985..b7759d5 100755
> --- a/git-rebase.sh
> +++ b/git-rebase.sh
> @@ -174,6 +174,9 @@ then
>   then
>   type=interactive
>   interactive_rebase=explicit
> + elif test -f "$merge_dir"/cherry
> + then
> + type=cherry
>   else
>   type=merge
>   fi
> @@ -382,7 +385,7 @@ then
>  elif test -n "$keep_empty"
>  then
>   type=cherry
> - state_dir="$apply_dir"
> + state_dir="$merge_dir"
>  else
>   type=am
>   state_dir="$apply_dir"
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 3/5] rebase: fix sequence continuation

[Junio C Hamano]

Felipe Contreras  writes:

> We are not in am mode.

That may make sense, but shouldn't this part be like so from the
very beginning?  In other words, this looks like an "oops, 1/5 was
buggy and this is a hotfix".

>
> Signed-off-by: Felipe Contreras 
> ---
>  git-rebase--cherry.sh | 10 ++
>  1 file changed, 6 insertions(+), 4 deletions(-)
>
> diff --git a/git-rebase--cherry.sh b/git-rebase--cherry.sh
> index ab1f8b7..ca78b1b 100644
> --- a/git-rebase--cherry.sh
> +++ b/git-rebase--cherry.sh
> @@ -5,13 +5,15 @@
>  
>  case "$action" in
>  continue)
> - git am --resolved --resolvemsg="$resolvemsg" &&
> - move_to_original_branch
> + git cherry-pick --continue &&
> + move_to_original_branch &&
> + rm -rf "$state_dir"
>   exit
>   ;;
>  skip)
> - git am --skip --resolvemsg="$resolvemsg" &&
> - move_to_original_branch
> + git cherry-pick --skip &&
> + move_to_original_branch &&
> + rm -rf "$state_dir"
>   exit
>   ;;
>  esac
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 5/5] rebase: fix cherry-pick invocations

[Junio C Hamano]

Felipe Contreras  writes:

> So that all the tests pass.
>
> Signed-off-by: Felipe Contreras 
> ---
>  git-rebase--cherry.sh | 17 -
>  1 file changed, 16 insertions(+), 1 deletion(-)
>
> diff --git a/git-rebase--cherry.sh b/git-rebase--cherry.sh
> index ca78b1b..c3a2ac9 100644
> --- a/git-rebase--cherry.sh
> +++ b/git-rebase--cherry.sh
> @@ -23,11 +23,26 @@ test -n "$rebase_root" && root_flag=--root
>  mkdir "$state_dir" || die "Could not create temporary $state_dir"
>  : > "$state_dir"/cherry || die "Could not mark as cherry"
>  
> +if test -n "$rebase_root"
> +then
> + revisions="$onto...$orig_head"
> +else
> + revisions="$upstream...$orig_head"
> +fi

"So that all the tests pass" needs a bit more explanation to say for
cherry-pick codepath why and how two-dot range fails and why and how
three-dot variant with --right-only fixes it.  What are the problematic
cases?

>  # we have to do this the hard way.  git format-patch completely squashes
>  # empty commits and even if it didn't the format doesn't really lend
>  # itself well to recording empty patches.  fortunately, cherry-pick
>  # makes this easy
> -git cherry-pick --allow-empty "$revisions"
> +if test -n "$keep_empty"
> +then
> + extra="--allow-empty"
> +else
> + extra="--skip-empty --cherry-pick"
> +fi
> +test -n "$GIT_QUIET" && extra="$extra -q"
> +test -z "$force_rebase" && extra="$extra --ff"
> +git cherry-pick --no-merges --right-only --topo-order --do-walk --copy-notes 
> $extra "$revisions"
>  ret=$?
>  
>  if test 0 != $ret
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 5/5] rebase: fix cherry-pick invocations

[Junio C Hamano]

Junio C Hamano  writes:

> Felipe Contreras  writes:
>
>> So that all the tests pass.
>>
>> Signed-off-by: Felipe Contreras 
>> ---
>>  git-rebase--cherry.sh | 17 -
>>  1 file changed, 16 insertions(+), 1 deletion(-)
>>
>> diff --git a/git-rebase--cherry.sh b/git-rebase--cherry.sh
>> index ca78b1b..c3a2ac9 100644
>> --- a/git-rebase--cherry.sh
>> +++ b/git-rebase--cherry.sh
>> @@ -23,11 +23,26 @@ test -n "$rebase_root" && root_flag=--root
>>  mkdir "$state_dir" || die "Could not create temporary $state_dir"
>>  : > "$state_dir"/cherry || die "Could not mark as cherry"
>>  
>> +if test -n "$rebase_root"
>> +then
>> +revisions="$onto...$orig_head"
>> +else
>> +revisions="$upstream...$orig_head"
>> +fi
>
> "So that all the tests pass" needs a bit more explanation to say for
> cherry-pick codepath why and how two-dot range fails and why and how
> three-dot variant with --right-only fixes it.  What are the problematic
> cases?

Yikes, sorry, this was me being slow.  Walking A...B range with
right-only and --cherry applied will filter the duplicates, which is
wat you want, I think, and walking A..B range will not do the
filtering for you.


>>  # we have to do this the hard way.  git format-patch completely squashes
>>  # empty commits and even if it didn't the format doesn't really lend
>>  # itself well to recording empty patches.  fortunately, cherry-pick
>>  # makes this easy
>> -git cherry-pick --allow-empty "$revisions"
>> +if test -n "$keep_empty"
>> +then
>> +extra="--allow-empty"
>> +else
>> +extra="--skip-empty --cherry-pick"
>> +fi
>> +test -n "$GIT_QUIET" && extra="$extra -q"
>> +test -z "$force_rebase" && extra="$extra --ff"
>> +git cherry-pick --no-merges --right-only --topo-order --do-walk 
>> --copy-notes $extra "$revisions"
>>  ret=$?
>>  
>>  if test 0 != $ret
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH] fix segfault with git log -c --follow

[Clemens Buchacher]

On Tue, May 28, 2013 at 10:22:17AM -0700, Junio C Hamano wrote:
> Clemens Buchacher  writes:
> 
> > In diff_tree_combined we make a copy of diffopts. In
> > try_to_follow_renames, called via diff_tree_sha1, we free and
> > re-initialize diffopts->pathspec->items. Since we did not make a deep
> > copy of diffopts in diff_tree_combined, the original diffopts does not
> > get the update. By the time we return from diff_tree_combined,
> > rev->diffopt->pathspec->items points to an invalid memory address. We
> > get a segfault next time we try to access that pathspec.
> 
> I am not quite sure if I follow.  Do you mean
> 
>   diff_tree_combined()
> - makes a shallow copy of rev->diffopt
> - calls diff_tree_sha1()
>   diff_tree_sha1()
>   - tries to follow rename and clobbers diffopt

Right.

> - tries to use the shallow copy of original rev->diffopt
>   that no longer is valid, which is a problem

diff_tree_combined does not try to use it right away. It does return,
but rev->diffopt is now invalid and the next time we do any kind of diff
with it, we have a problem.

> I wonder, just like we force recursive and disable external on the
> copy before we use it to call diff_tree_sha1(), if we should disable
> follow-renames on it.  "--follow" is an option that is given to the
> history traversal part and it should not play any role in getting the
> pairwise diff with all parents diff_tree_combined() does.

Can't parse that last sentence.

In any case, I don't think disabling diff_tree_sha1 is a solution. The
bug is in diff_tree_sha1 and its subfunctions, because they manipulate a
data structures such that it becomes corrupt. And they do so in an
obfuscated and clearly unintentional manner. So we should not blame the
user for calling diff_tree_sha1 in such a way that it causes corruption.

> Besides,
> 
>  - "--follow" hack lets us keep track of only one path; and

Ok. Good to know it is considered a hack. The code is quite strange
indeed.

>  - "-c" and "--cc" make sense only when dealing with a merge commit
>and the path in the child may have come from different path in
>parents,

Sorry, I don't get it.

> so I am not sure if allowing combination of "--follow -c/--cc" makes
> much sense in the first place.

My use-case is came up with this history:

1. Code gets added to file A.
2. File A gets renamed to B in a different branch.
3. The branches get merged, and code from (1) is removed in the merge.

Later I wonder why code from (1) is gone from B even though I felt
certain it had been added before. I also remember that B was renamed at
some point. So I do git log -p --follow B, and it nicely shows that diff
where the code was added, but no diff where the code is removed.

The reason is of course, that the code was removed in the merge and that
diff is not shown. And -c is usually what I do to enable showing diffs
in merge commits.

And if the pairwise diff can also deal with file renames, I think it
absolutely does make sense to show also a three-way diff.

I can't tell far away the code is from supporting anything like that.

Cheers,
Clemens
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH] fix segfault with git log -c --follow

[Junio C Hamano]

Clemens Buchacher  writes:

>> I wonder, just like we force recursive and disable external on the
>> copy before we use it to call diff_tree_sha1(), if we should disable
>> follow-renames on it.  "--follow" is an option that is given to the
>> history traversal part and it should not play any role in getting the
>> pairwise diff with all parents diff_tree_combined() does.
>
> Can't parse that last sentence.
>
> In any case, I don't think disabling diff_tree_sha1 is a solution. The
> bug is in diff_tree_sha1 and its subfunctions, because they manipulate a
> data structures such that it becomes corrupt. And they do so in an
> obfuscated and clearly unintentional manner. So we should not blame the
> user for calling diff_tree_sha1 in such a way that it causes corruption.
>
>> Besides,
>> 
>>  - "--follow" hack lets us keep track of only one path; and
>
> Ok. Good to know it is considered a hack. The code is quite strange
> indeed.

The problem with --follow is that it only tracks one path globally.
In a history like this, suppose that a path X long time ago was
renamed to Y at commit B:

---o---A---B---C---o HEAD

and you start digging with "log --follow -c HEAD -- Y".  When
looking at C, because it and its parent B both have path Y, the
try-to-follow hack does not kick in, and when trying to show C, we
will show the change in Y (because that is the pathspec).

Then we look at B.  Because B has path we are following, i.e. Y, and
its parent A does not, try-to-follow hack kicks in, and it mangles
the pathspec that is used globally for history traversal to X while
showing the difference between A's X and B's Y.  Then we dig further
to find A; at this point the global pathspec is swapped and now it
is X.

That makes --follow a working hack for a simplest single strand of
pearls.  But if you have a mergy history, e.g.

---o---A---B---C---o HEAD
\ /
 D---E---F---G---H

it can break in interesting ways.  We are likely to have looked at H
before looking at B and used pathspec Y while inspecting H, but
after looking at B, the global pathspec is swapped to X, and then we
try to look at G, F, E and D, none of which may have renamed the
original X, so you would likely miss the change to the path Y you
wanted to follow.

To fix this, we would need to keep "what path are we following" not
in the global revs->pathspec, but per the traversal paths that are
currently active (e.g. when we look at C and H, it is Y, when we
look at B, it is X, when we look at G, that is inherited from H and
still Y, not affected by the rename at B.  And then when we look at
A (we need topo-order traversal to do this), it needs to notice that
one child (i.e. B) has been following X while the other (i.e. D) Y,
and merge the "I've been following this path" information in a
sensible way (e.g. look at its own tree and see what is available,
in this case X).
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH v2 7/8] remote-bzr: reorganize the way 'wanted' works

[Felipe Contreras]

Junio C Hamano  wrote:
> Felipe Contreras  writes:
> 
> > +wanted = get_config('remote-bzr.branches').rstrip().split(', ')
> 
> Two minor nits and one design suggestion:
> 
>  - Why rstrip() not strip()?

The purpose of the strip is to remove the _single_ "\n" at the end that
subprocess communicate. Maybe get_config() should do that.

> It appears that this only is helping
>an end-user "mistake" like this:
> 
>   git config remote-bzr.branches 'trunk, devel, test '
> 
>without helping people who have done this:
> 
>   git config remote-bzr.branches 'trunk,  devel, test'

No, that's tnot it.

>  - Is
> 
>  git config remote-bzr.branches trunk,devel,test
> 
>a grave sin?
> 
>In other words, wouldn't we want something like this instead?
> 
>   map(lambda s: s.strip(), get_config('...').split(','))

Yeah, that might make sense.

>  - Doesn't allowing multi-valued variable, e.g.
> 
>   [remote-bzr]
> branches = trunk
> branches = devel
> branches = test
> 
>make it easier for the user to manage this configuration by
>e.g. selectively removing or adding tracked branches?

How would the 'git config' command look like?

Either way, that's orthogonal to this patch.

-- 
Felipe Contreras
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 1/2] sequencer: trivial fix

[Felipe Contreras]

Junio C Hamano wrote:
> Neil Horman  writes:
> 
> > On Mon, May 27, 2013 at 11:52:18AM -0500, Felipe Contreras wrote:
> >> We should free objects before leaving.
> >> 
> >> Signed-off-by: Felipe Contreras 
> >> ---
> >>  sequencer.c | 7 +--
> >>  1 file changed, 5 insertions(+), 2 deletions(-)
> >> 
> >> diff --git a/sequencer.c b/sequencer.c
> >> index ab6f8a7..7eeae2f 100644
> >> --- a/sequencer.c
> >> +++ b/sequencer.c
> >> @@ -626,12 +626,15 @@ static int do_pick_commit(struct commit *commit, 
> >> struct replay_opts *opts)
> >>rerere(opts->allow_rerere_auto);
> >>} else {
> >>int allow = allow_empty(opts, commit);
> >> -  if (allow < 0)
> >> -  return allow;
> >> +  if (allow < 0) {
> >> +  res = allow;
> >> +  goto leave;
> >> +  }
> >>if (!opts->no_commit)
> >>res = run_git_commit(defmsg, opts, allow);
> >>}
> >>  
> >> +leave:
> >>free_message(&msg);
> >>free(defmsg);
> >>  
> >> -- 
> >> 1.8.3.rc3.312.g47657de
> >> 
> >> 
> > Acked-by: Neil Horman 
> 
> This is better done without "goto" in general.
> 
> The other patch 2/2/ adds one more "we need to exit from the middle
> of the flow" and makes it look handier to add an exit label here,
> but it would be even better to express the logic of that patch as a
> normal cascade of if/else if/..., which is small enough and we do
> not need the "leave:" label.

Linux kernel developers would disagree. In C 'goto' is quite of then the only
sane option, and you can see 'goto' used in the Linux kernel all over the place
for that reason.

In this particular case it also makes perfect sense.

> It probably is better to fold this patch into the other one when it
> is rerolled to correct the option name gotcha "on the tin".

Why? This patch is standalone and fixes an issue that is independent of the
other patch. Why squash two patches that do *two* different things?

Anyway, I'll happily drop this patch if you want this memory leak to remain.
But then I'll do the same in the other patch.

This mantra of avodiing 'goto' is not helping anybody.

-- 
Felipe Contreras
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 4/5] test: improve rebase -q test

[Felipe Contreras]

Junio C Hamano wrote:
> Felipe Contreras  writes:
> 
> > Let's show the output so it's clear why it failed.
> >
> > Signed-off-by: Felipe Contreras 
> > ---
> >  t/t3400-rebase.sh | 1 +
> >  1 file changed, 1 insertion(+)
> >
> > diff --git a/t/t3400-rebase.sh b/t/t3400-rebase.sh
> > index b58fa1a..fb39531 100755
> > --- a/t/t3400-rebase.sh
> > +++ b/t/t3400-rebase.sh
> > @@ -185,6 +185,7 @@ test_expect_success 'default to @{upstream} when 
> > upstream arg is missing' '
> >  test_expect_success 'rebase -q is quiet' '
> > git checkout -b quiet topic &&
> > git rebase -q master >output.out 2>&1 &&
> > +   cat output.out &&
> > test ! -s output.out
> >  '
> 
> It is one thing to avoid squelching output that naturally comes out
> of command being tested unnecessarily, so that "./t-*.sh -v"
> output can be used for debugging.  I however am not sure if adding
> "cat" to random places like this is a productive direction for us to
> go in.
> 
> A more preferrable alternative may be adding something like this to
> test-lib.sh and call it from here and elsewhere (there are about 50
> places that do "test ! -s "), perhaps?
> 
> test_must_be_an_empty_file () {
> if test -s "$1"
> then
> cat "$1"
> false
> fi
> }

Perhaps, but I'm not interested. I'm tired of obvious fixes getting rejected
for hypothetical "ideal" situations that we'll never reach.

-- 
Felipe Contreras
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 1/3] cherry-pick: add support to copy notes

[Felipe Contreras]

Junio C Hamano wrote:
> Felipe Contreras  writes:
> 
> > Signed-off-by: Felipe Contreras 
> > ---
> >  builtin/revert.c  |   2 +
> >  sequencer.c   | 136 
> > --
> >  sequencer.h   |   2 +
> >  t/t3500-cherry.sh |  32 +
> >  4 files changed, 169 insertions(+), 3 deletions(-)
> 
> "git cherry-pick" should help maintaining notes just like amend and
> rebase, but how should this interact with notes.rewrite.,
> where the command is capable of doing this without an explicit
> option once you tell which notes need to be maintained?
> 
> Thomas Rast Cc'ed as he has been the primary force behind this line
> of "notes" usability.
> 
> It probably is not sensible to carry over note from a reverted
> commit to its revert, but I didn't immediately spot how the patch
> does this only for cherry-pick but not for revert (the codepath in
> do_pick_commit() is shared between them, no?).

The option is enabled from the UI, and only 'git cherry-pick' has the
--copy-notes option.

-- 
Felipe Contreras
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 1/3] cherry-pick: add support to copy notes

[Felipe Contreras]

Thomas Rast wrote:
> Junio C Hamano  writes:
> 
> > Thomas Rast Cc'ed as he has been the primary force behind this line
> > of "notes" usability.
> 
> Thanks for pointing this out to me.
> 
> > Felipe Contreras  writes:
> >
> >> Signed-off-by: Felipe Contreras 
> >> ---
> >>  builtin/revert.c  |   2 +
> >>  sequencer.c   | 136 
> >> --
> >>  sequencer.h   |   2 +
> >>  t/t3500-cherry.sh |  32 +
> >>  4 files changed, 169 insertions(+), 3 deletions(-)
> >
> > "git cherry-pick" should help maintaining notes just like amend and
> > rebase, but how should this interact with notes.rewrite.,
> > where the command is capable of doing this without an explicit
> > option once you tell which notes need to be maintained?
> 
> Since we already have the notes.rewrite. convention, it would
> seem the obvious choice to line it up with the others.  The main
> bikeshedding opportunity is whether this should be an exception and
> default to false (all other commands default it to true).
> 
> Also: how does this interact with notes.rewriteRef and the corresponding
> env vars?  Why?
> 
> How does it interact with 'cherry-pick -n' if this is done in sequence,
> effectively squashing several commits (this use-case is actually
> suggested by the manpage), if multiple source commits had notes?  Should
> it respect notes.rewriteMode (and by default concatenate)?  (I don't
> know if the sequencer state is expressive enough already to carry this
> in a meaningful way across cherry-pick commands.)

Feel free to implement that. I'm just interested in 'git cherry-pick' being
usable for 'git rebase' purposes.

-- 
Felipe Contreras
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 2/4] remote-helpers: rename tests

[Felipe Contreras]

Junio C Hamano wrote:
> Felipe Contreras  writes:
> 
> > The .t extension is more standard for sharness tests.
> >
> > Signed-off-by: Felipe Contreras 
> > ---
> 
> Is that "sharness" test the sh script testsuite forked from our
> testsuite?
> 
> I do not see how it makes sense to copy how they deviate from us
> back to our codebase, especially if we plan to eventually move some
> of these tests out of contrib/ area, but even without such a plan in
> the future.

They deviate from us, we deviate from them, whatever. We are a single project,
what more than one project does is more standard.

% man prove

--ext Set the extension for tests (default '.t')

-- 
Felipe Contreras
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

RE: contrib/git-normal-to-bare.sh

[Felipe Contreras]

Zenaan Harkness wrote:
> This question comes up every now and then - how to convert from normal
> to bare, or vice versa.
> 
> This is just a start to a basic script to go one way. Needs more tests
> etc, but it's enough to get newbies (like me) off to a reasonable
> start.
> 
> PLEASE CC me, as I am not subscribed.

You do not need to ask to be CC'ed. Any mailing list properly configued would
not munge the Reply-To header, so anybody replying to this mail would reply to
you.

Git is a properly configured mailing list, just like all the mailing list in
vger.kernel.org.

Cheers.

-- 
Felipe Contreras
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

RE: What's cooking in git.git (May 2013, #08; Tue, 28)

[Felipe Contreras]

Junio C Hamano wrote:
> * fc/makefile (2013-05-26) 5 commits
>  - build: do not install git-remote-testpy
>  - build: add NO_INSTALL variable
>  - build: cleanup using $<
>  - build: cleanup using $^
>  - build: trivial simplification
>  (this branch is used by fc/remote-helpers-use-specified-python.)

No, these are independent.

> * fc/remote-helpers-use-specified-python (2013-05-28) 4 commits
>  - remote-helpers: add exec-path links
>  - remote-helpers: allow direct test execution
>  - remote-helpers: rename tests
>  - remote-helpers: generate scripts
>  (this branch uses fc/makefile.)
> 
>  I do not particularly think the second from the bottom is a good
>  change, but it takes the remainder of the series hostage.
>  Will hopefully be rerolled without it.


> * fc/remote-bzr (2013-05-28) 8 commits
>  - remote-bzr: add fallback check for a partial clone
>  - remote-bzr: reorganize the way 'wanted' works
>  - remote-bzr: trivial cleanups
>  - remote-bzr: change global repo
>  - remote-bzr: delay cloning/pulling
>  - remote-bzr: simplify get_remote_branch()
>  - remote-bzr: fix for files with spaces
>  - remote-bzr: recover from failed clones
> 
>  The ones near the tip conflicted with the hotfix for 1.8.3 so I
>  discarded them for now.
> 
>  Will merge to 'next'?

Didn't I resend these with the conflict fixed?

-- 
Felipe Contreras
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 0/5] rebase: improve the keep-empty

[Felipe Contreras]

Martin von Zweigbergk wrote:
> I think I have some patches at home that instead teach 'git am' the
> --keep-empty flag. Does that make sense? It's been a while since I
> looked at it, but I'll try to take a look tonight (PST).

I think it does make sense. But I still would prefer 'git rebase' to rely on
'git cherry-pick' primariliy.

-- 
Felipe Contreras
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH v2 0/3] Towards a useable git-branch

[Duy Nguyen]

On Tue, May 28, 2013 at 9:28 PM, Ramkumar Ramachandra
 wrote:
> Ramkumar Ramachandra wrote:
>> %>(N) doesn't work properly with f-e-r, and I'm not sure why.  I'm not
>> talking about your last patch where you compute * -- that works fine;
>> it's just that %>(N) doesn't when N is a concrete number.
>
> Try this:
>
> %(refname:short)%>(30)%(upstream:short)
>
> (assuming that you have lots of branches).  I'm noticing random
> alignment problems.

It's because you don't pad enough spaces after %(refname:short) so the
starting point of %(upstream:short) on each line is already unaligned,
I think. Try this:

%<(*)%(refname:short)%>(30)%(upstream:short)

or if you prefer at specific column (e.g. align upstream close to the
60th column, regardless of refname's length):

%(refname:short)%>|(60)%(upstream:short)
--
Duy
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 2/5] rebase: fix 'cherry' mode storage

[Felipe Contreras]

Junio C Hamano wrote:
> Felipe Contreras  writes:
> 
> > We don't use the 'rebase-apply'.
> 
> s/.$/; we will use rebase-merge instead./ I think.

We could use 'rebase-apply' or any directory, but currently we don't use any,
and 'rebase-apply' is for 'git am'.

> > Signed-off-by: Felipe Contreras 
> > ---
> >  git-rebase--cherry.sh | 4 
> >  git-rebase.sh | 5 -
> >  2 files changed, 8 insertions(+), 1 deletion(-)
> >
> > diff --git a/git-rebase--cherry.sh b/git-rebase--cherry.sh
> > index cbf80f9..ab1f8b7 100644
> > --- a/git-rebase--cherry.sh
> > +++ b/git-rebase--cherry.sh
> > @@ -18,6 +18,9 @@ esac
> >  
> >  test -n "$rebase_root" && root_flag=--root
> >  
> > +mkdir "$state_dir" || die "Could not create temporary $state_dir"
> > +: > "$state_dir"/cherry || die "Could not mark as cherry"
> 
> Style:
> 
>   >"$state_dir/cherry-pick"

Fix that in git-rebase--interactive.sh then, where this code came from.

> I am not sure if the user _cares_ that internally we use cherry-pick
> when s/he asks us to do a keep-empty, and I suspect "mark as cherry"
> incomprehensible.  I do not even know what is going on at this point
> (yet).

It's the same we do in git-rebase--interactive.sh.

> I _suspect_ that you are using cherry-pick when --keep-empty is given,
> so it might even make sense to talk in the end-users' terms,
> i.e. call this "$state_dir/keep-empty" (I am not sure if calling
> this split half "git-rebase--keep-empty.sh" makes sense, yet).

If git-rebase--cherry(-pick)?.sh does the same as 'git-rebase--am.sh', we can
replace it, and make yet another step of replacing scripts (git am) with C code
(git cherry-pick). I thought I explained that somewhow, maybe I didn't.

It still makes sense to fix the broken --keep-empty behavior.

-- 
Felipe Contreras
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH] completion: avoid ls-remote in certain scenarios

[Felipe Contreras]

It's _very_ slow in many cases, and there's really no point in fetching
*everything* from the remote just for completion. In many cases it might
be faster for the user to type the whole thing.

If the user manually specifies 'refs/*', then the full ls-remote
completion is triggered.

Signed-off-by: Felipe Contreras 
---
 contrib/completion/git-completion.bash | 10 ++
 1 file changed, 2 insertions(+), 8 deletions(-)

diff --git a/contrib/completion/git-completion.bash 
b/contrib/completion/git-completion.bash
index 1c35eef..2ce4f7d 100644
--- a/contrib/completion/git-completion.bash
+++ b/contrib/completion/git-completion.bash
@@ -427,14 +427,8 @@ __git_refs ()
done
;;
*)
-   git ls-remote "$dir" HEAD ORIG_HEAD 'refs/tags/*' 
'refs/heads/*' 'refs/remotes/*' 2>/dev/null | \
-   while read -r hash i; do
-   case "$i" in
-   *^{}) ;;
-   refs/*) echo "${i#refs/*/}" ;;
-   *) echo "$i" ;;
-   esac
-   done
+   echo "HEAD"
+   git for-each-ref --format="%(refname:short)" -- 
"refs/remotes/$dir/" | sed -e "s#^$dir/##"
;;
esac
 }
-- 
1.8.3.rc3.312.g47657de

--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH v2 0/3] Towards a useable git-branch

[Duy Nguyen]

On Tue, May 28, 2013 at 9:01 PM, Ramkumar Ramachandra
 wrote:
> Also, a couple of minor annoyances:
>
> 1. When f-e-r is invoked with refs/tags, we get stray output.  Atleast
> it doesn't segfault, thanks to your ignore-commit patch.  Maybe
> printing stray output is the right thing to do, as opposed to erroring
> out.

What format did you use?

> 2. %>(*) only works with f-e-r atoms, not with pretty-format atoms.
> This is ofcourse obvious from the implementation, but isn't it a
> little consistent?

It is not. I think it's documented as well as a known implementation
limitation. It's not hard to be fixed (we could call
format_commit_message again for all entries, this time with a single
placeholder, to collect the best width). If anybody does need %>(*)%H
to work, we could do it. BTW, the way I implement get_atom_width() is
not really optimal. For n lines, we call print_value() in
get_atom_width n^2 times. We could cache the calculated width instead
and reduce that to n times.

> Should we start off a new pretty-ref-formats document, where we
> explicitly exclude things like %ae (because of the hex overriding
> thing)?  I don't think it's a problem if documented properly.

And this one is doucmented as well, I think. I'd really like to
introduce a new --pretty option (or something) that does not accept
%xx as a hex notion, so %ae and friends won't be hidden. It's also a
good thing for compatibility because currently %H in --format produces
%H. After this, %H produces something else. It's unlikely that anybody
has done that. But it's even better if we avoid that possibility
entirely with a new option.
--
Duy
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH] completion: zsh: improve bash script loading

[Felipe Contreras]

It's better to check in multiple locations, so the user doesn't have to.

And update the documentation.

Signed-off-by: Felipe Contreras 
---
 contrib/completion/git-completion.zsh | 26 ++
 1 file changed, 18 insertions(+), 8 deletions(-)

diff --git a/contrib/completion/git-completion.zsh 
b/contrib/completion/git-completion.zsh
index 2565d2e..9555cf8 100644
--- a/contrib/completion/git-completion.zsh
+++ b/contrib/completion/git-completion.zsh
@@ -4,18 +4,17 @@
 #
 # Copyright (c) 2012-2013 Felipe Contreras 
 #
-# You need git's bash completion script installed somewhere, by default on the
-# same directory as this script.
+# You need git's bash completion script installed somewhere, by default it
+# would be the location bash-completion uses.
 #
-# If your script is on ~/.git-completion.sh instead, you can configure it on
-# your ~/.zshrc:
+# If your script is somewhere else, you can configure it on your ~/.zshrc:
 #
 #  zstyle ':completion:*:*:git:*' script ~/.git-completion.sh
 #
-# The recommended way to install this script is to copy to
-# '~/.zsh/completion/_git', and then add the following to your ~/.zshrc file:
+# The recommended way to install this script is to copy to '~/.zsh/_git', and
+# then add the following to your ~/.zshrc file:
 #
-#  fpath=(~/.zsh/completion $fpath)
+#  fpath=(~/.zsh $fpath)
 
 complete ()
 {
@@ -27,7 +26,18 @@ zstyle -T ':completion:*:*:git:*' tag-order && \
zstyle ':completion:*:*:git:*' tag-order 'common-commands'
 
 zstyle -s ":completion:*:*:git:*" script script
-test -z "$script" && script="$(dirname 
${funcsourcetrace[1]%:*})"/git-completion.bash
+if [ -z "$script" ]; then
+   local -a locations
+   locations=(
+   '/etc/bash_completion.d/git' # fedora, old debian
+   '/usr/share/bash-completion/completions/git' # arch, ubuntu, 
new debian
+   '/usr/share/bash-completion/git' # gentoo
+   $(dirname ${funcsourcetrace[1]%:*})/git-completion.bash
+   )
+   for e in $locations; do
+   test -f $e && script="$e" && break
+   done
+fi
 ZSH_VERSION='' . "$script"
 
 __gitcomp ()
-- 
1.8.3.rc3.312.g47657de

--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: 1.8.3 - gitignore not being parsed correctly on OS X; regex support is broken?

[Duy Nguyen]

On Wed, May 29, 2013 at 12:54 AM, Misty De Meo  wrote:
> Hi,
>
> Gitignore parsing no longer seems to work properly in git 1.8.3.
>
> One of my repositories has the following gitignore:
>
> /*
> !/.gitignore
> !/Library/
> !/CONTRIBUTING.md
> !/README.md
> !/SUPPORTERS.md
> !/bin
> /bin/*
> !/bin/brew
> !/share/man/man1/brew.1
> .DS_Store
> /Library/LinkedKegs
> /Library/PinnedKegs
> /Library/Taps
> /Library/Formula/.gitignore
>
> In 1.8.2.3 and earlier, this works as expected. However, in 1.8.3 I'm
> seeing every file in /bin/ being marked as an untracked file.

The changes in this area since 1.8.2.3 seem to be Karsten's (I'm not
blaming, just wanted to narrow down the problem). The patterns of
interest seem to be

!/bin
/bin/*
!/bin/brew

Without "!/bin" v1.8.3 seems to behave the same as v1.8.2.3. Can you verify it?

> I asked about this in #git, and was told that the culprit was the
> regex support; apparently recompiling without regex support fixes the
> specific gitignore issue. However, this doesn't seem to have been
> reported anywhere on the mailing list that I can see. I was also told
> that the issue is OS X-specific, and doesn't happen on other
> platforms.

Puzzled. regex has nothing to do with gitignore (glob does). How do
you recompile without regex support? Setting NO_REGEX? I'm on Linux
btw, no chance of touching OS X.
--
Duy
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH v2 0/8] cherry-pick: improvements

[Felipe Contreras]

Hi,

Here's a bunch of changes to make cherry-pick (and revert) more useful. In
particular; this makes it more friendly for 'git rebase.

Felipe Contreras (8):
  sequencer: remove useless indentation
  sequencer: trivial fix
  cherry-pick: add --skip-empty option
  cherry-pick: store rewritten commits
  sequencer: run post-rewrite hook
  cherry-pick: add support to copy notes
  revert/cherry-pick: add --quiet option
  revert/cherry-pick: add --skip option

 Documentation/git-cherry-pick.txt   |  10 +-
 Documentation/git-revert.txt|   7 +-
 Documentation/sequencer.txt |   3 +
 builtin/revert.c|  11 ++
 sequencer.c | 230 ++--
 sequencer.h |   7 +-
 t/t3500-cherry.sh   |  32 +
 t/t3508-cherry-pick-many-commits.sh |  13 ++
 t/t3510-cherry-pick-sequence.sh |  12 ++
 9 files changed, 310 insertions(+), 15 deletions(-)

-- 
1.8.3.rc3.312.g47657de

--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH v2 1/8] sequencer: remove useless indentation

[Felipe Contreras]

By using good ol' goto.

Signed-off-by: Felipe Contreras 
---
 sequencer.c | 16 +---
 1 file changed, 9 insertions(+), 7 deletions(-)

diff --git a/sequencer.c b/sequencer.c
index ab6f8a7..b4989ba 100644
--- a/sequencer.c
+++ b/sequencer.c
@@ -474,7 +474,7 @@ static int do_pick_commit(struct commit *commit, struct 
replay_opts *opts)
struct commit_message msg = { NULL, NULL, NULL, NULL, NULL };
char *defmsg = NULL;
struct strbuf msgbuf = STRBUF_INIT;
-   int res, unborn = 0;
+   int res, unborn = 0, allow;
 
if (opts->no_commit) {
/*
@@ -624,14 +624,16 @@ static int do_pick_commit(struct commit *commit, struct 
replay_opts *opts)
  msg.subject);
print_advice(res == 1, opts);
rerere(opts->allow_rerere_auto);
-   } else {
-   int allow = allow_empty(opts, commit);
-   if (allow < 0)
-   return allow;
-   if (!opts->no_commit)
-   res = run_git_commit(defmsg, opts, allow);
+   goto leave;
}
 
+   allow = allow_empty(opts, commit);
+   if (allow < 0)
+   return allow;
+   if (!opts->no_commit)
+   res = run_git_commit(defmsg, opts, allow);
+
+leave:
free_message(&msg);
free(defmsg);
 
-- 
1.8.3.rc3.312.g47657de

--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH v2 2/8] sequencer: trivial fix

[Felipe Contreras]

We should free objects before leaving.

Signed-off-by: Felipe Contreras 
---
 sequencer.c | 6 --
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/sequencer.c b/sequencer.c
index b4989ba..f7be7d8 100644
--- a/sequencer.c
+++ b/sequencer.c
@@ -628,8 +628,10 @@ static int do_pick_commit(struct commit *commit, struct 
replay_opts *opts)
}
 
allow = allow_empty(opts, commit);
-   if (allow < 0)
-   return allow;
+   if (allow < 0) {
+   res = allow;
+   goto leave;
+   }
if (!opts->no_commit)
res = run_git_commit(defmsg, opts, allow);
 
-- 
1.8.3.rc3.312.g47657de

--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH v2 3/8] cherry-pick: add --skip-empty option

[Felipe Contreras]

Pretty much what it says on the tin.

Signed-off-by: Felipe Contreras 
---
 Documentation/git-cherry-pick.txt   |  3 +++
 builtin/revert.c|  2 ++
 sequencer.c |  6 ++
 sequencer.h |  1 +
 t/t3508-cherry-pick-many-commits.sh | 13 +
 5 files changed, 25 insertions(+)

diff --git a/Documentation/git-cherry-pick.txt 
b/Documentation/git-cherry-pick.txt
index c205d23..fccd936 100644
--- a/Documentation/git-cherry-pick.txt
+++ b/Documentation/git-cherry-pick.txt
@@ -129,6 +129,9 @@ effect to your index in a row.
redundant commits are ignored.  This option overrides that behavior and
creates an empty commit object.  Implies `--allow-empty`.
 
+--skip-empty::
+   Instead of failing, skip commits that are or become empty.
+
 --strategy=::
Use the given merge strategy.  Should only be used once.
See the MERGE STRATEGIES section in linkgit:git-merge[1]
diff --git a/builtin/revert.c b/builtin/revert.c
index 0401fdb..0e5ce71 100644
--- a/builtin/revert.c
+++ b/builtin/revert.c
@@ -118,6 +118,7 @@ static void parse_args(int argc, const char **argv, struct 
replay_opts *opts)
OPT_END(),
OPT_END(),
OPT_END(),
+   OPT_END(),
};
 
if (opts->action == REPLAY_PICK) {
@@ -127,6 +128,7 @@ static void parse_args(int argc, const char **argv, struct 
replay_opts *opts)
OPT_BOOLEAN(0, "allow-empty", &opts->allow_empty, 
N_("preserve initially empty commits")),
OPT_BOOLEAN(0, "allow-empty-message", 
&opts->allow_empty_message, N_("allow commits with empty messages")),
OPT_BOOLEAN(0, "keep-redundant-commits", 
&opts->keep_redundant_commits, N_("keep redundant, empty commits")),
+   OPT_BOOLEAN(0, "skip-empty", &opts->skip_empty, 
N_("skip empty commits")),
OPT_END(),
};
if (parse_options_concat(options, ARRAY_SIZE(options), 
cp_extra))
diff --git a/sequencer.c b/sequencer.c
index f7be7d8..d3c7d72 100644
--- a/sequencer.c
+++ b/sequencer.c
@@ -627,6 +627,12 @@ static int do_pick_commit(struct commit *commit, struct 
replay_opts *opts)
goto leave;
}
 
+   if (opts->skip_empty && is_index_unchanged() == 1) {
+   warning(_("skipping %s... %s"),
+   find_unique_abbrev(commit->object.sha1, DEFAULT_ABBREV),
+   msg.subject);
+   goto leave;
+   }
allow = allow_empty(opts, commit);
if (allow < 0) {
res = allow;
diff --git a/sequencer.h b/sequencer.h
index 1fc22dc..3b04844 100644
--- a/sequencer.h
+++ b/sequencer.h
@@ -34,6 +34,7 @@ struct replay_opts {
int allow_empty;
int allow_empty_message;
int keep_redundant_commits;
+   int skip_empty;
 
int mainline;
 
diff --git a/t/t3508-cherry-pick-many-commits.sh 
b/t/t3508-cherry-pick-many-commits.sh
index 19c99d7..3dc19c6 100755
--- a/t/t3508-cherry-pick-many-commits.sh
+++ b/t/t3508-cherry-pick-many-commits.sh
@@ -187,4 +187,17 @@ test_expect_success 'cherry-pick --stdin works' '
check_head_differs_from fourth
 '
 
+test_expect_success 'cherry-pick skip empty' '
+   git clean -fxd &&
+   git checkout -b empty fourth &&
+   git commit --allow-empty -m empty &&
+   test_commit ontop &&
+   git checkout -f master &&
+   git reset --hard fourth &&
+   git cherry-pick --skip-empty fourth..empty &&
+   echo ontop > expected &&
+   git log --format=%s fourth..HEAD > actual
+   test_cmp expected actual
+'
+
 test_done
-- 
1.8.3.rc3.312.g47657de

--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH v2 4/8] cherry-pick: store rewritten commits

[Felipe Contreras]

Will be useful for the next commits.

Signed-off-by: Felipe Contreras 
---
 sequencer.c | 95 -
 sequencer.h |  1 +
 2 files changed, 95 insertions(+), 1 deletion(-)

diff --git a/sequencer.c b/sequencer.c
index d3c7d72..c217716 100644
--- a/sequencer.c
+++ b/sequencer.c
@@ -20,6 +20,18 @@
 const char sign_off_header[] = "Signed-off-by: ";
 static const char cherry_picked_prefix[] = "(cherry picked from commit ";
 
+struct rewritten_list_item {
+   unsigned char from[20];
+   unsigned char to[20];
+};
+
+struct rewritten_list {
+   struct rewritten_list_item *items;
+   unsigned int nr, alloc;
+};
+
+static struct rewritten_list rewritten;
+
 static int is_rfc2822_line(const char *buf, int len)
 {
int i;
@@ -102,6 +114,19 @@ static int has_conforming_footer(struct strbuf *sb, struct 
strbuf *sob,
return 1;
 }
 
+static void add_rewritten(unsigned char *from, unsigned char *to)
+{
+   struct rewritten_list_item *item;
+   if (rewritten.nr + 1 >= rewritten.alloc) {
+   rewritten.alloc += 32;
+   rewritten.items = xrealloc(rewritten.items, rewritten.alloc * 
sizeof(*rewritten.items));
+   }
+   item = &rewritten.items[rewritten.nr];
+   hashcpy(item->from, from);
+   hashcpy(item->to, to);
+   rewritten.nr++;
+}
+
 static void remove_sequencer_state(void)
 {
struct strbuf seq_dir = STRBUF_INIT;
@@ -641,6 +666,14 @@ static int do_pick_commit(struct commit *commit, struct 
replay_opts *opts)
if (!opts->no_commit)
res = run_git_commit(defmsg, opts, allow);
 
+   if (!res) {
+   unsigned char to[20];
+
+   if (read_ref("HEAD", to))
+   goto leave;
+
+   add_rewritten(commit->object.sha1, to);
+   }
 leave:
free_message(&msg);
free(defmsg);
@@ -786,6 +819,40 @@ static void read_populate_todo(struct commit_list 
**todo_list,
die(_("Unusable instruction sheet: %s"), todo_file);
 }
 
+static void read_populate_rewritten(void)
+{
+   const char *rewritten_file = git_path(SEQ_REWR_FILE);
+   struct strbuf buf = STRBUF_INIT;
+   char *p;
+   int fd;
+
+   fd = open(rewritten_file, O_RDONLY);
+   if (fd < 0)
+   return;
+   if (strbuf_read(&buf, fd, 0) < 0) {
+   close(fd);
+   strbuf_release(&buf);
+   return;
+   }
+   close(fd);
+
+   for (p = buf.buf; *p;) {
+   unsigned char from[20];
+   unsigned char to[20];
+   char *eol = strchrnul(p, '\n');
+   if (eol - p != 81)
+   /* wrong size */
+   break;
+   if (get_sha1_hex(p, from))
+   break;
+   if (get_sha1_hex(p + 41, to))
+   break;
+   add_rewritten(from, to);
+   p = *eol ? eol + 1 : eol;
+   }
+   strbuf_release(&buf);
+}
+
 static int populate_opts_cb(const char *key, const char *value, void *data)
 {
struct replay_opts *opts = data;
@@ -958,6 +1025,29 @@ static void save_todo(struct commit_list *todo_list, 
struct replay_opts *opts)
strbuf_release(&buf);
 }
 
+static void save_rewritten(void)
+{
+   const char *todo_file = git_path(SEQ_REWR_FILE);
+   static struct lock_file rewritten_lock;
+   struct strbuf buf = STRBUF_INIT;
+   int fd, i;
+
+   fd = hold_lock_file_for_update(&rewritten_lock, todo_file, 
LOCK_DIE_ON_ERROR);
+   for (i = 0; i < rewritten.nr; i++) {
+   struct rewritten_list_item *item = &rewritten.items[i];
+   strbuf_addf(&buf, "%s %s\n", sha1_to_hex(item->from), 
sha1_to_hex(item->to));
+   }
+   if (write_in_full(fd, buf.buf, buf.len) < 0) {
+   strbuf_release(&buf);
+   die_errno(_("Could not write to %s"), todo_file);
+   }
+   if (commit_lock_file(&rewritten_lock) < 0) {
+   strbuf_release(&buf);
+   die(_("Error wrapping up %s."), todo_file);
+   }
+   strbuf_release(&buf);
+}
+
 static void save_opts(struct replay_opts *opts)
 {
const char *opts_file = git_path(SEQ_OPTS_FILE);
@@ -1003,8 +1093,10 @@ static int pick_commits(struct commit_list *todo_list, 
struct replay_opts *opts)
for (cur = todo_list; cur; cur = cur->next) {
save_todo(cur, opts);
res = do_pick_commit(cur->item, opts);
-   if (res)
+   if (res) {
+   save_rewritten();
return res;
+   }
}
 
/*
@@ -1033,6 +1125,7 @@ static int sequencer_continue(struct replay_opts *opts)
return continue_single_pick();
read_populate_opts(&opts);
read_populate_todo(&todo_list, opts);
+   read_populate_rewritten();
 
/* Verify that the con

[PATCH v2 5/8] sequencer: run post-rewrite hook

[Felipe Contreras]

As we should.

Signed-off-by: Felipe Contreras 
---
 sequencer.c | 45 -
 1 file changed, 44 insertions(+), 1 deletion(-)

diff --git a/sequencer.c b/sequencer.c
index c217716..3aa480e 100644
--- a/sequencer.c
+++ b/sequencer.c
@@ -127,6 +127,37 @@ static void add_rewritten(unsigned char *from, unsigned 
char *to)
rewritten.nr++;
 }
 
+static void run_rewrite_hook(void)
+{
+   struct strbuf buf = STRBUF_INIT;
+   struct child_process proc;
+   const char *argv[3];
+   int code, i;
+
+   argv[0] = find_hook("post-rewrite");
+   if (!argv[0])
+   return;
+
+   argv[1] = "rebase";
+   argv[2] = NULL;
+
+   memset(&proc, 0, sizeof(proc));
+   proc.argv = argv;
+   proc.in = -1;
+   proc.stdout_to_stderr = 1;
+
+   code = start_command(&proc);
+   if (code)
+   return;
+   for (i = 0; i < rewritten.nr; i++) {
+   struct rewritten_list_item *item = &rewritten.items[i];
+   strbuf_addf(&buf, "%s %s\n", sha1_to_hex(item->from), 
sha1_to_hex(item->to));
+   }
+   write_in_full(proc.in, buf.buf, buf.len);
+   close(proc.in);
+   finish_command(&proc);
+}
+
 static void remove_sequencer_state(void)
 {
struct strbuf seq_dir = STRBUF_INIT;
@@ -1099,6 +1130,8 @@ static int pick_commits(struct commit_list *todo_list, 
struct replay_opts *opts)
}
}
 
+   run_rewrite_hook();
+
/*
 * Sequence of picks finished successfully; cleanup by
 * removing the .git/sequencer directory
@@ -1136,14 +1169,24 @@ static int sequencer_continue(struct replay_opts *opts)
}
if (index_differs_from("HEAD", 0))
return error_dirty_index(opts);
+   {
+   unsigned char to[20];
+   if (!read_ref("HEAD", to))
+   add_rewritten(todo_list->item->object.sha1, to);
+   }
todo_list = todo_list->next;
return pick_commits(todo_list, opts);
 }
 
 static int single_pick(struct commit *cmit, struct replay_opts *opts)
 {
+   int ret;
setenv(GIT_REFLOG_ACTION, action_name(opts), 0);
-   return do_pick_commit(cmit, opts);
+   ret = do_pick_commit(cmit, opts);
+   if (ret)
+   return ret;
+   run_rewrite_hook();
+   return 0;
 }
 
 int sequencer_pick_revisions(struct replay_opts *opts)
-- 
1.8.3.rc3.312.g47657de

--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html