On Tue, May 10, 2016 at 3:12 PM, Eric Sunshine wrote:
> On Tue, May 10, 2016 at 2:39 PM, Jeff King wrote:
>> On Tue, May 10, 2016 at 01:20:54AM -0400, Eric Sunshine wrote:
>>> while :
>>> do
>>> case "$1" in
>>> -C)
Eric Sunshine writes:
> On Tue, May 10, 2016 at 3:12 PM, Eric Sunshine
> wrote:
>> On Tue, May 10, 2016 at 2:39 PM, Jeff King wrote:
>>> On Tue, May 10, 2016 at 01:20:54AM -0400, Eric Sunshine wrote:
while :
Ramsay Jones writes:
> On 10/05/16 12:52, Dennis Kaarsemaker wrote:
>> On ma, 2016-05-09 at 15:22 -0700, Junio C Hamano wrote:
>>> It passes on one box and fails on another. They both run the same
>>> Ubuntu 14.04 derivative, with same ext3 filesystem. The failing
On Tue, May 10, 2016 at 01:20:54AM -0400, Eric Sunshine wrote:
> diff --git a/t/t1500-rev-parse.sh b/t/t1500-rev-parse.sh
> index c058aa4..525e6d3 100755
> --- a/t/t1500-rev-parse.sh
> +++ b/t/t1500-rev-parse.sh
> @@ -7,11 +7,13 @@ test_description='test git rev-parse'
> test_rev_parse () {
>
Duy Nguyen writes:
> Or a simpler, more-to-the-point patch like this?
I am OK with that, even though I find it a bit too "cute" for my
taste.
> -- 8< --
> Subject: [PATCH] wrap-for-bin.sh: regenerate bin-wrappers when switching
> branches
>
> Commit e6e7530 (test helpers:
Junio C Hamano writes:
> Junio C Hamano writes:
>
>> Eric Sunshine writes:
>>
>>> This series modernizes t1500; it takes an entirely different approach
>>> than [1][2] and is intended to replace that series.
>>
>> Turns out that it
Sascha Silbe writes:
> A combination of --break-rewrites and --inter-hunk-context that merges
> changes with less than the given number of unchanged lines between them
> into a single delete/insert change would be even better. But just
> ignoring the
Johannes Schindelin writes:
> In Git for Windows' SDK, Git's source code is always checked out
> with symlinks disabled. The reason is that POSIX symlinks have no
> accurate equivalent on Windows [*1*]. More precisely, though, it is
> not just Git's source code but
Eric Sunshine writes:
> t1500: test_rev_parse: facilitate future test enhancements
> t1500: reduce dependence upon global state
> t1500: avoid changing working directory outside of tests
> t1500: avoid setting configuration options outside of tests
> t1500:
Armin Kunaschik writes:
> I fail to see why eval is really necessary here.
It is necessary to work correctly with any strategy option with $IFS
in it, I would think. The calling script "git-rebase" accumulates
--strategy-option values after passing each of them
On Tue, May 10, 2016 at 2:39 PM, Jeff King wrote:
> On Tue, May 10, 2016 at 01:20:54AM -0400, Eric Sunshine wrote:
>> diff --git a/t/t1500-rev-parse.sh b/t/t1500-rev-parse.sh
>> @@ -7,11 +7,13 @@ test_description='test git rev-parse'
>> while :
>> do
>>
On Fri, 06 May 2016, Yaroslav Halchenko wrote:
> Dear Git Folks,
> Originally this issue was mentioned in previous thread [1], and I have decided
> to bring it into a separate thread. ATM there is a dichotomy in git behavior
> between cloning non-bare repos: if I clone over ssh or just
Johannes Schindelin writes:
> To: Junio C Hamano
> Cc: git@vger.kernel.org
Probably the above is the other way around.
> The -x flag (trace commands) is a priceless tool when hunting down bugs
> that trigger test failures. It is a worthless tool
Eric Sunshine writes:
>> I don't know if it's worth worrying about or not. The usual solution is
>> something like:
>>
>> env_git_dir=$2
>> env='GIT_DIR=$env_git_dir; export GIT_DIR'
>> ...
>> eval "$env"
>
> Makes sense; I wasn't quite happy with having $2
On Tue, May 10, 2016 at 3:48 PM, Eric Sunshine wrote:
> Actually, I think we can have improved encapsulation and maintain
> readability like this:
>
> case "$1" in
> ...
> -g) env="$2"; shift; shift ;;
> ...
> esac
>
> ...
>
Yaroslav Halchenko writes:
> On Fri, 06 May 2016, Yaroslav Halchenko wrote:
>
>> Dear Git Folks,
>
>> Originally this issue was mentioned in previous thread [1], and I have
>> decided
>> to bring it into a separate thread. ATM there is a dichotomy in git behavior
>>
Johannes Schindelin writes:
> This patch makes perf-lib.sh more robust so that it can run correctly
> even inside a worktree. For example, it assumed that $GIT_DIR/objects is
> the objects directory (which is not the case for worktrees) and it used
> the commondir
On Tue, May 10, 2016 at 03:59:42PM -0400, Eric Sunshine wrote:
> On Tue, May 10, 2016 at 3:48 PM, Eric Sunshine
> wrote:
> > Actually, I think we can have improved encapsulation and maintain
> > readability like this:
> >
> > case "$1" in
> > ...
> > -g)
Just noticed a curiosity:
$ git show -W 3e3ceaa58 quote.c
shows the entire file. The commit in question adds a whole new
function at the end of the file. If I move that addition to just
before the last function the file already had before the change and
amend the commit, "show -W" would
On Tue, May 10, 2016 at 1:11 PM, Junio C Hamano wrote:
> Yaroslav Halchenko writes:
>
>> On Fri, 06 May 2016, Yaroslav Halchenko wrote:
>>
>>> Dear Git Folks,
>>
>>> Originally this issue was mentioned in previous thread [1], and I have
>>> decided
>>> to
Jeff King writes:
> On Tue, May 10, 2016 at 12:49:42PM -0700, Junio C Hamano wrote:
>
>> I wonder if we can fix "-x" instead so that we do not have to
>> butcher tests like this patch does. It was quite clear what it
>> expected to see before this patch, and it is sad that the
On Tue, May 10, 2016 at 12:49:42PM -0700, Junio C Hamano wrote:
> I wonder if we can fix "-x" instead so that we do not have to
> butcher tests like this patch does. It was quite clear what it
> expected to see before this patch, and it is sad that the workaround
> makes less readable (and
SZEDER Gábor writes:
> I wonder if is it really necessary to specify the path to the .git
> directory via $GIT_DIR. Would 'git --git-dir=/over/there' be just as
> good?
Then you are testing two different things that may go through
different codepaths.
Adding yet another
Jeff King writes:
> I think it is clear why it works. If $strategy_opts is empty, then the
> code we generate looks like:
>
> for strategy_opt in
> do
> ...
> done
Ah, of course. Thanks.
--
To unsubscribe from this list: send the line "unsubscribe git" in
the
On Tue, May 10, 2016 at 12:11:59PM -0700, Junio C Hamano wrote:
> Armin Kunaschik writes:
>
> > I fail to see why eval is really necessary here.
>
> It is necessary to work correctly with any strategy option with $IFS
> in it, I would think. The calling script
On Tue, May 10, 2016 at 01:53:56PM -0700, Junio C Hamano wrote:
> Jeff King writes:
>
> > I think it is clear why it works. If $strategy_opts is empty, then the
> > code we generate looks like:
> >
> > for strategy_opt in
> > do
> > ...
> > done
>
> Ah, of
On Tue, May 10, 2016 at 5:11 PM, Junio C Hamano wrote:
> SZEDER Gábor writes:
>> I wonder if is it really necessary to specify the path to the .git
>> directory via $GIT_DIR. Would 'git --git-dir=/over/there' be just as
>> good?
>
> Then you are testing two
On Tue, May 10, 2016 at 02:13:26PM -0700, Junio C Hamano wrote:
> > I don't think there is a scalable, portable way to do so. "-x" output is
> > going to stderr, and is inherited by any functions or subshells. So
> > either we have to ask "-x" output to go somewhere else, or we have to
> > turn
Michael Haggerty writes:
> ... I think I have addressed all of the points that were
> brought up. Plus I fixed a pre-existing bug that I noticed myself
> while adding some more tests; see the first bullet point below for
> more information.
>
> Changes between v1 and v2:
>
Jeff King writes:
> The patch itself is a trivial-looking one-liner, but there
> are a few subtleties worth mentioning:
>
> - the variable is _not_ exported; the "set -x" is local to
> our process, and so the tracefd should match
>
> - this line has to come after we do the
Quoting Eric Sunshine :
On Tue, May 10, 2016 at 3:12 PM, Eric Sunshine
wrote:
On Tue, May 10, 2016 at 2:39 PM, Jeff King wrote:
On Tue, May 10, 2016 at 01:20:54AM -0400, Eric Sunshine wrote:
while :
do
On Tue, 10 May 2016, Jacob Keller wrote:
> > The necessary update to the client might as simple as using
> > $GIVEN_URL/.git/ and attempting the request again after seeing the
> > probe for $GIVEN_URL/info/refs fails.
> I know at least Jenkin's Git plugin has a workaround to solve this
> issue
On 10/05/16 21:30, Junio C Hamano wrote:
> Ramsay Jones writes:
>
>> On 10/05/16 12:52, Dennis Kaarsemaker wrote:
>>> On ma, 2016-05-09 at 15:22 -0700, Junio C Hamano wrote:
It passes on one box and fails on another. They both run the same
Ubuntu 14.04
Yaroslav Halchenko writes:
>> The necessary update to the client might be as simple as using
>> $GIVEN_URL/.git/ and attempting the request again after seeing the
>> probe for $GIVEN_URL/info/refs fails.
>
> Sure -- workarounds are possible,...
Just so that there is no
Junio C Hamano writes:
> Just noticed a curiosity:
>
> $ git show -W 3e3ceaa58 quote.c
>
> shows the entire file. The commit in question adds a whole new
> function at the end of the file. If I move that addition to just
> before the last function the file already had
On Wed, May 11, 2016 at 6:07 AM, Junio C Hamano wrote:
> Nguyễn Thái Ngọc Duy writes:
>
>> Signed-off-by: Nguyễn Thái Ngọc Duy
>> ---
>
> This changes semantics, doesn't it? prefix_filename() seems to do a
> lot more than just
Stefan Beller writes:
> +static int in_group(int argc, const char **argv, const char *prefix)
> + ...
> + if (!group)
> + list = git_config_get_value_multi("submodule.updateGroup");
> + else {
> + string_list_split(_list, group, ',', -1);
Is
Stefan Beller writes:
> +static void split_argv_pathspec_groups(int argc, const char **argv,
> +const char ***pathspec_argv,
> +struct string_list *group)
> +{
> + int i;
> + struct argv_array ps
Duy Nguyen writes:
> On second thought, why hold patches back, lengthen the worktree-move
> series and make it a pain to review? I moved a few patches from
> worktree-move into this series and I took two other out to create
> nd/error-errno. So I'm going to take more patches
Stefan Beller writes:
> @@ -199,6 +203,7 @@ static struct submodule *lookup_or_create_by_name(struct
> submodule_cache *cache,
> submodule->update_strategy.command = NULL;
> submodule->fetch_recurse = RECURSE_SUBMODULES_NONE;
> submodule->ignore = NULL;
> +
On Wed, May 11, 2016 at 6:03 AM, Junio C Hamano wrote:
> Duy Nguyen writes:
>
>> On second thought, why hold patches back, lengthen the worktree-move
>> series and make it a pain to review? I moved a few patches from
>> worktree-move into this series and I
On Tue, 10 May 2016, Junio C Hamano wrote:
> >> The necessary update to the client might be as simple as using
> >> $GIVEN_URL/.git/ and attempting the request again after seeing the
> >> probe for $GIVEN_URL/info/refs fails.
> > Sure -- workarounds are possible,...
> Just so that there is no
Nguyễn Thái Ngọc Duy writes:
> Signed-off-by: Nguyễn Thái Ngọc Duy
> ---
This changes semantics, doesn't it? prefix_filename() seems to do a
lot more than just strbuf_vadd("%s%s", prefix, filename); would do.
It may be a good change (e.g. turn '\' into
This allows to specify a subset of all available submodules to be
initialized and cloned. It is unrelated to the `--recursive` option,
i.e. the user may still want to give `--recursive` as an option.
Originally `--recursive` implied to initialize all submodules, this
changes as well with the new
In later patches we need to tell if a submodule is in a group,
so expose a handy test function in both C and shell.
Signed-off-by: Stefan Beller
---
builtin/submodule--helper.c | 42 +++-
submodule-config.c | 50
The new switch `--init-default-group` updates the submodules which are
configured in `submodule.updateGroup`
Signed-off-by: Stefan Beller
---
Documentation/config.txt| 5
Documentation/git-submodule.txt | 4 ++--
git-submodule.sh| 14
We need the submodule labels in a later patch.
Signed-off-by: Stefan Beller
---
submodule-config.c | 16
submodule-config.h | 2 ++
2 files changed, 18 insertions(+)
diff --git a/submodule-config.c b/submodule-config.c
index b82d1fb..0cdb47e 100644
---
Additionally to taking a pathspec, `module_list_compute` will also take
labels and submodule names, when these are prefixed by '*' and ':'
respectively.
`module_list_compute` is used by other functions in the submodule helper:
* module_list, used by `submodule {deinit, status, sync, foreach}`
*
When adding new submodules, you can specify the labels the submodule
belongs to by giving one or more --label arguments. This will record
each label in the .gitmodules file as a value of the key
"submodule.$NAME.label".
Signed-off-by: Stefan Beller
---
I started from scratch as I think there were some sharp edges in the design.
My thinking shifted from "submodule groups" towards "actually it's just an
enhanced pathspec, called submodulespec".
The meat is found in the last 3 patches.
What is this series about?
==
If you
We could allow more than just alphanumeric and dash characters
for submodule labels. As a precaution we'll first allow only this
subset and later on we can extend it once we have more experience
with them.
Signed-off-by: Stefan Beller
---
builtin/submodule--helper.c | 30
Stefan Beller writes:
> diff --git a/t/t7400-submodule-basic.sh b/t/t7400-submodule-basic.sh
> index 814ee63..0adc4e4 100755
> --- a/t/t7400-submodule-basic.sh
> +++ b/t/t7400-submodule-basic.sh
> @@ -1056,6 +1056,7 @@ test_expect_success 'submodule with UTF-8 name' '
> '
>
Stefan Beller writes:
> +static int submodule_valid_label_name(const char *label)
> +{
> + if (!label || !strlen(label))
> + return 0;
> +
> + if (!isalnum(*label))
> + return 0;
I'd limit this one to isalpha() if I were doing this to make the
On 09.05.16 22:29, Junio C Hamano wrote:
> tbo...@web.de writes:
>
>> +if (stats->stat_bits & earlyout)
>> +break; /* We found what we have been searching for */
>
> Are we sure if our callers are only interested in just one bit at a
> time? Otherwise, if we want
Stefan Beller writes:
> I started from scratch as I think there were some sharp edges in the design.
> My thinking shifted from "submodule groups" towards "actually it's just an
> enhanced pathspec, called submodulespec".
Except for minor things I mentioned separately,
On Wed, Apr 13, 2016 at 9:15 AM, Nguyễn Thái Ngọc Duy wrote:
> Similar to "mv a b/", which is actually "mv a b/a", we extract basename
> of source worktree and create a directory of the same name at
> destination if dst path is a directory.
>
> Signed-off-by: Nguyễn Thái Ngọc
Junio C Hamano writes:
> Stefan Beller writes:
>
>> +static void split_argv_pathspec_groups(int argc, const char **argv,
>> + const char ***pathspec_argv,
>> + struct string_list *group)
On Fri, Apr 22, 2016 at 9:01 AM, Nguyễn Thái Ngọc Duy wrote:
> Similar to the rebase case, we want to detect if "HEAD" in some worktree
> is being bisected because
>
> 1) we do not want to checkout this branch in another worktree, after
>bisect is done it will want to go
On Tue, May 10, 2016 at 7:45 PM, Duy Nguyen wrote:
> If --detach is used, log_warning() can't cover die(),
> warning() or error(), most importantly die() for example because of
> bugs.
A case for redirecting warning() is because watchman-support.c uses
it. But because this
Sorry for any duplicate mails, the list blocked my html mail.
Note to self: Don't use GMail on a tablet.
On Mon, May 9, 2016 at 11:35 PM, Eric Sunshine wrote:
>>
>> Hmph, do we have a broken &&-chain?
>
> I don't know. Unfortunately, Armin didn't provide much information
On Tue, May 10, 2016 at 03:44:32PM +0200, Armin Kunaschik wrote:
> I'm building on a quite current AIX 6.1 where /bin/sh defaults to /bin/ksh
> which is a posix shell (ksh88).
> Using /bin/bash doesn't work because SHELL_PATH is only used in
> git scripts but not in any t* test scripts.
If you
If I do a "git-rebase -i ..." followed by "git reflog expire ..." and
"git gc ..." then I can end up with a repo which has a ref ORIG_HEAD
which points to a non-existing object.
- Is this intended?
- What's the reason to keep a ref which points to a non-existing object?
- Are there any other refs
On Tue, May 10, 2016 at 6:12 AM, David Turner wrote:
> I think that's a SIGPIPE on the first git status. Weird, since I just
> added sigpipe-avoidance code (in v8). Does anyone have any idea why
> the sigchain stuff isn't doing what I think it is?
SIGPIPE code works.
In f924b52 (Windows: add pthread_sigmask() that does nothing,
2016-05-01), we introduced a no-op for Windows. However, this breaks
building Git in Git for Windows' SDK because pthread_sigmask() is
already a no-op there, #define'd in the pthread_signal.h header in
This is the first patch series in preparation for a faster interactive
rebase.
It actually only prepares the test script that I mainly used to develop
the rebase--helper, and the resilience against running with -x proved to
be invaluable in keeping my sanity.
Johannes Schindelin (2):
t3404:
The -x flag (trace commands) is a priceless tool when hunting down bugs
that trigger test failures. It is a worthless tool if the -x flag
*itself* triggers test failures.
So let's change the offending tests so that they are a bit less
stringent and do not stumble over the "+..." lines generated
Signed-off-by: Johannes Schindelin
---
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 d96d0e4..66348f1 100755
--- a/t/t3404-rebase-interactive.sh
+++
So far we haven't needed to identify an existing worktree from command
line. Future commands such as lock or move will need it. There are of
course other options for identifying a worktree, for example by branch
or even by internal id. They may be added later if proved useful.
Signed-off-by:
This provides an API for checking if a worktree is locked. We need to
check this to avoid double locking a worktree, or try to unlock one when
it's not even locked.
Signed-off-by: Nguyễn Thái Ngọc Duy
---
worktree.c | 18 ++
worktree.h | 6 ++
2 files
Signed-off-by: Nguyễn Thái Ngọc Duy
---
Documentation/git-worktree.txt | 12 --
builtin/worktree.c | 41 ++
contrib/completion/git-completion.bash | 5 -
t/t2028-worktree-move.sh (new +x) | 34
Signed-off-by: Nguyễn Thái Ngọc Duy
---
Documentation/git-worktree.txt | 5 +
builtin/worktree.c | 31 +++
contrib/completion/git-completion.bash | 2 +-
t/t2028-worktree-move.sh | 14 ++
4
Main worktree _is_ different. You can lock a linked worktree but not the
main one, for example. Provide an API for checking that.
Signed-off-by: Nguyễn Thái Ngọc Duy
---
worktree.c | 5 +
worktree.h | 5 +
2 files changed, 10 insertions(+)
diff --git a/worktree.c
strbuf is a bit overkill for this function. What we need is call
absolute_path() twice and make sure the second call does not destroy the
result of the first. One buffer allocation is enough.
Signed-off-by: Nguyễn Thái Ngọc Duy
---
worktree.c | 16 +++-
1 file
This is probably not the best order. But it makes it no-brainer to know
where to insert new commands. At some point we might want to reorder at
least the synopsis part again, grouping commonly use subcommands together.
Signed-off-by: Nguyễn Thái Ngọc Duy
---
Signed-off-by: Nguyễn Thái Ngọc Duy
---
builtin/worktree.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/builtin/worktree.c b/builtin/worktree.c
index b53f802..f9dac37 100644
--- a/builtin/worktree.c
+++ b/builtin/worktree.c
@@ -337,7 +337,7 @@ static
Signed-off-by: Nguyễn Thái Ngọc Duy
---
builtin/worktree.c | 2 +-
worktree.c | 2 +-
2 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/builtin/worktree.c b/builtin/worktree.c
index bf80111..aaee0e2 100644
--- a/builtin/worktree.c
+++ b/builtin/worktree.c
Signed-off-by: Nguyễn Thái Ngọc Duy
---
builtin/worktree.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/builtin/worktree.c b/builtin/worktree.c
index aaee0e2..b53f802 100644
--- a/builtin/worktree.c
+++ b/builtin/worktree.c
@@ -262,7 +262,7 @@ static int
Signed-off-by: Nguyễn Thái Ngọc Duy
---
contrib/completion/git-completion.bash | 23 +++
1 file changed, 23 insertions(+)
diff --git a/contrib/completion/git-completion.bash
b/contrib/completion/git-completion.bash
index 3402475..d3ac391 100644
---
The use case is keep some worktree and discard the rest of the worktree
list.
Signed-off-by: Nguyễn Thái Ngọc Duy
---
worktree.c | 14 +++---
worktree.h | 5 +
2 files changed, 16 insertions(+), 3 deletions(-)
diff --git a/worktree.c b/worktree.c
index
On Fri, May 06, 2016 at 05:21:05PM +0700, Duy Nguyen wrote:
> > Similarly, it looks like 'path' doesn't need to be a strbuf at all
> > since the result of absolute_path() should remain valid long enough
> > for fspathcmp(). It could just be:
> >
> > const char *path = absolute_path(...);
> >
Eric Sunshine writes:
> This series modernizes t1500; it takes an entirely different approach
> than [1][2] and is intended to replace that series.
Turns out that it wasn't so painful after all.
The only small niggle I have is on 6/6; my preference would be,
because
> On 07 May 2016, at 00:46, Junio C Hamano wrote:
>
> [...]
>
>
> * jc/doc-lint (2016-05-04) 2 commits
> - Documentation: fix linkgit references
> - ci: validate "gitlink:" in documentation
The first patch "Documentation: fix linkgit references" seems not
to be on the
> On 09 May 2016, at 08:18, Johannes Schindelin
> wrote:
>
> Lars Schneider noticed that the configuration introduced to test the extra
> HTTP headers cannot be used with Apache 2.2 (which is still actively
> maintained, as pointed out by Junio Hamano).
>
> To let
On Tuesday, February 16, 2016 07:33:54 PM Junio C Hamano wrote:
> > ---
>
> I think I mislead you into a slightly wrong direction. While the
> single liner does improve the situation, I think this is merely a
> band-aid upon closer inspection. For example, if you changed your
> "commit
Jeff King writes:
> [+cc Junio as this should be the final version]
Thanks, I think I queued with "do not cat a single file to a pipe"
tweak already.
>> > When commit 618310a taught t6302 to run without the GPG
>>
>> 618310a (t6302: skip only signed tags rather than all tests
On Tue, May 10, 2016 at 1:20 AM, Eric Sunshine wrote:
> Ideally, each test should be responsible for setting up state it needs
> rather than relying upon transient global state. Toward this end, teach
> test_rev_parse() to accept a "-b " option to allow callers to set
>
Hi Junio,
On Fri, 6 May 2016, Junio C Hamano wrote:
> * jc/fsck-nul-in-commit (2016-04-14) 2 commits
> - fsck: detect and warn a commit with embedded NUL
> - fsck_commit_buffer(): do not special case the last validation
>
> "git fsck" learned to catch NUL byte in a commit object as
>
This is the second preparatory patch series for my rebase--helper work
(i.e. moving parts of the interactive rebase into a builtin).
It simply introduces a perf test (and ensures that it runs in my
environment) so as to better determine how much the performance changes,
really.
Johannes
On 10/05/16 12:52, Dennis Kaarsemaker wrote:
> On ma, 2016-05-09 at 15:22 -0700, Junio C Hamano wrote:
>> It passes on one box and fails on another. They both run the same
>> Ubuntu 14.04 derivative, with same ext3 filesystem. The failing one
>> is on a VM.
>
> Same here, except ext4 instead
This developer spent a lot of time trying to speed up the interactive
rebase, in particular on Windows. And will continue to do so.
To make it easier to demonstrate the performance improvement, let's have
a reproducible performance test.
The topic branch we use to test performance was found
This patch makes perf-lib.sh more robust so that it can run correctly
even inside a worktree. For example, it assumed that $GIT_DIR/objects is
the objects directory (which is not the case for worktrees) and it used
the commondir file verbatim, even if it contained a relative path.
Signed-off-by:
In Git for Windows' SDK, Git's source code is always checked out
with symlinks disabled. The reason is that POSIX symlinks have no
accurate equivalent on Windows [*1*]. More precisely, though, it is
not just Git's source code but *all* source code that is checked
out with symlinks disabled:
My use case is an army of build agents that need only limited and
selective access to otherwise private repositories.
The first part already made it into `master`, this is the remainder.
This iteration is based on 'jk/submodule-c-credential' and therefore
converted the original config-sanitizing
Lars Schneider noticed that the configuration introduced to test the
extra HTTP headers cannot be used with Apache 2.2 (which is still
actively maintained, as pointed out by Junio Hamano).
To let the tests pass with Apache 2.2 again, let's substitute the
offending and `expr` by using old school
To test that extra HTTP headers are passed correctly, t5551 verifies that
a fetch succeeds when two required headers are passed, and that the fetch
does not succeed when those headers are not passed.
However, this test would also succeed if the configuration required only
one header. As Apache's
Johannes Schindelin writes:
>> To be honest, I do not quite understand why you call it "ugly hack"
>> at all.
>
> Well, it is convoluted. I would have preferred to say "if this condition
> is not met or that condition is not met, fail". Instead I had to say "If`
>
On Fri, May 6, 2016 at 9:35 AM, Michael Rappazzo wrote:
> Executing `git-rev-parse` with `--git-common-dir`, `--git-path `,
> or `--shared-index-path` from the root of the main worktree results in
> a relative path to the git dir.
>
> When executed from a subdirectory of the
Hi Junio,
On Mon, 9 May 2016, Junio C Hamano wrote:
> Johannes Schindelin writes:
>
> > Okay, I already force-pushed my extra-http-header branch and the next
> > iteration will sport this paragraph.
>
> The new explanation is well written and can and should also
To support this developer's use case of allowing build agents token-based
access to private repositories, we introduced the http.extraheader
feature, allowing extra HTTP headers to be sent along with every HTTP
request.
This patch verifies that we can configure these extra HTTP headers via the
Lars Schneider writes:
>> On 09 May 2016, at 08:18, Johannes Schindelin
>> wrote:
>>
>> Lars Schneider noticed that the configuration introduced to test the extra
>> HTTP headers cannot be used with Apache 2.2 (which is still actively
>>
1 - 100 of 131 matches
Mail list logo