Junio C Hamano wrote:
Two announcements for their version 0.2 on the list archive are not
quite enough to advertise them to their users.
Signed-off-by: Junio C Hamano gits...@pobox.com
---
* I am inclined to queue this for 2.0, and we would also need an
update to the release notes as
On Fri, May 16, 2014 at 3:26 AM, Junio C Hamano gits...@pobox.com wrote:
Jakub Narębski jna...@gmail.com writes:
On Thu, May 15, 2014 at 9:38 PM, Junio C Hamano gits...@pobox.com wrote:
Jakub Narębski jna...@gmail.com writes:
Writing test for this would not be easy, and require some HTML
On Thu, May 15, 2014 at 06:31:21PM -0700, Jeremiah Mahler wrote:
Added feature that allows a signature file to be used with format-patch.
$ git format-patch --signature-file ~/.signature -1
Now signatures with newlines and other special characters can be
easily included.
I think this
Junio C Hamano wrote:
Felipe Contreras felipe.contre...@gmail.com writes:
== contrib vs. core ==
This is the only point relevant to contrib vs. core:
- We may be painted in a hard place if remote-hg or remote-bzr take
us to a position where the Git as a whole is blocked while
Hello,
On Thu, May 15, 2014 at 03:22:26PM -0400, Jeff King wrote:
As the person who is proposing the patch for git.git, I would hope
Stepan would follow up on such review and confirm whether or not it is
still needed.
well, I try to. (I verified that less -I works in msysGit before
On Thu, May 15, 2014 at 07:17:35AM -0700, Elia Pinto wrote:
Even though POSIX.1 lists -a/-o as options to test, they are
marked Obsolescent XSI. Scripts using these expressions
should be converted as follow:
test $1 -a $2
should be written as:
test $1 test $2
Likewise
test $1
On Fri, May 16, 2014 at 10:19:57AM +0200, Stepan Kasal wrote:
On Thu, May 15, 2014 at 03:22:26PM -0400, Jeff King wrote:
As the person who is proposing the patch for git.git, I would hope
Stepan would follow up on such review and confirm whether or not it is
still needed.
well, I try
The patch returned by
http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/patch/?id=462fb2af9788a82a534f8184abfde31574e1cfa0
is truncated. The page which refers to that patch, at
On Thu, May 15, 2014 at 03:56:29PM -0700, Junio C Hamano wrote:
Two announcements for their version 0.2 on the list archive are not
quite enough to advertise them to their users.
I do not think this README nor a mention in the release notes will get
their attention either, and many people (and
On Fri, May 16, 2014 at 06:04:57PM +0930, David Newall wrote:
The patch returned by
http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/patch/?id=462fb2af9788a82a534f8184abfde31574e1cfa0
is truncated. The page which refers to that patch, at
On Fri, May 16, 2014 at 10:41 AM, Jeff King p...@peff.net wrote:
But that being said, this is Felipe's code. While we have a legal right
to distribute it in v2.0, if he would really prefer it out for v2.0, I
would respect that.
My understanding is that Felipe would prefer to keep it _in_ the
On Fri, May 16, 2014 at 10:55:54AM +0200, Paolo Ciarrocchi wrote:
On Fri, May 16, 2014 at 10:41 AM, Jeff King p...@peff.net wrote:
But that being said, this is Felipe's code. While we have a legal right
to distribute it in v2.0, if he would really prefer it out for v2.0, I
would respect
On Fri, May 16, 2014 at 03:08:51AM -0500, Felipe Contreras wrote:
Junio C Hamano wrote:
Felipe Contreras felipe.contre...@gmail.com writes:
== contrib vs. core ==
This is the only point relevant to contrib vs. core:
- We may be painted in a hard place if remote-hg or
Paolo Ciarrocchi wrote:
On Fri, May 16, 2014 at 10:41 AM, Jeff King p...@peff.net wrote:
But that being said, this is Felipe's code. While we have a legal right
to distribute it in v2.0, if he would really prefer it out for v2.0, I
would respect that.
My understanding is that Felipe
Hi,
On Fri, May 16, 2014 at 04:29:58AM -0400, Jeff King wrote:
[..] I hope I did not sound like blame Stepan, he was screwed up.
no, you did not, it was ok.
From msysgit's perspective, they may or may not want to revert the patch
that they already have. That is a _separate_ issue, and I
On Fri, May 16, 2014 at 04:51:16AM -0400, Jeff King wrote:
On Fri, May 16, 2014 at 06:04:57PM +0930, David Newall wrote:
The patch returned by
http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/patch/?id=462fb2af9788a82a534f8184abfde31574e1cfa0
is truncated. The page which
Jeff King wrote:
On Thu, May 15, 2014 at 03:56:29PM -0700, Junio C Hamano wrote:
Two announcements for their version 0.2 on the list archive are not
quite enough to advertise them to their users.
I do not think this README nor a mention in the release notes will get
their attention
On Fri, May 16, 2014 at 12:46 AM, Jeff Sipek jef...@josefsipek.net wrote:
On Tue, May 13, 2014 at 10:30:48PM +0200, Per Cederqvist wrote:
If you run something like guilt header '.*' the command would crash,
because the grep comand that tries to ensure that the patch exist
would detect a match,
On Fri, May 16, 2014 at 2:06 AM, Philip Oakley philipoak...@iee.org wrote:
From: John Fisher fishook2...@gmail.com
I assert based on one piece of evidence ( a post from a facebook dev) that
I now have the worlds biggest and slowest git
repository, and I am not a happy guy. I used to have the
Hi,
I have stumbled on a weird bug. At work, we use redmine as an issue
tracker and its task are marked by a number starting with #. When I
commit some work and write #1234 in the message, it works. However,
later on when I remember that I forgot to add some files and amend the
commit, vim
On Fri, May 16, 2014 at 5:18 PM, Michal Stasa michal.st...@gmail.com wrote:
Hi,
I have stumbled on a weird bug. At work, we use redmine as an issue
tracker and its task are marked by a number starting with #. When I
commit some work and write #1234 in the message, it works. However,
later on
William Giokas wrote:
On Fri, May 16, 2014 at 03:08:51AM -0500, Felipe Contreras wrote:
This is a red herring. Ignore the fact that it will never happen (which
it won't), the next point remains a FACT, and you conveniently ignore
it.
It may not block git being released, but as we can see
Michal Stasa michal.st...@gmail.com writes:
I have stumbled on a weird bug. At work, we use redmine as an issue
tracker and its task are marked by a number starting with #. When I
commit some work and write #1234 in the message, it works. However,
later on when I remember that I forgot to add
It sounds like a good workaround but I think there could be a problem.
When vim opens there is the message on the first line and two lines
below is a commented text which uses # as comment char. Does the char
change when you change the comment char?
Michal Staša
Santhos.net
www.santhos.net
+420
If we need to use core.commentChar and it's already in the prepared
message, find another char among a small subset. This should stop
surprises because git strips some lines unexpectedly. Of course if
candicate characters happen to be all out, this change does not help.
Signed-off-by: Nguyễn Thái
On Fri, May 16, 2014 at 05:21:36AM -0500, Felipe Contreras wrote:
How exactly would it be better?
If you concede that the Git release wouldn't be affected, then assuming
a hypothetical future where git-remote-hg is bundled, and we have a
Mercurial API breakage, we would have:
Git v2.5
William Giokas wrote:
On Fri, May 16, 2014 at 05:21:36AM -0500, Felipe Contreras wrote:
How exactly would it be better?
If you concede that the Git release wouldn't be affected, then assuming
a hypothetical future where git-remote-hg is bundled, and we have a
Mercurial API breakage, we
On Fri, May 16, 2014 at 11:51:37AM +0200, Per Cederqvist wrote:
On Fri, May 16, 2014 at 12:46 AM, Jeff Sipek jef...@josefsipek.net wrote:
On Tue, May 13, 2014 at 10:30:48PM +0200, Per Cederqvist wrote:
If you run something like guilt header '.*' the command would crash,
because the grep
Pretty format string %(N,[ml]trunc)%s truncates subject to a given
length with an appropriate padding. This works for non-ASCII texts when
i18n.logOutputEncoding is UTF-8 only (independently of a printed commit
message encoding) but does not work when i18n.logOutputEncoding is NOT
UTF-8.
Pretty format string %(N,[ml]trunc)%s truncates subject to a given
length with an appropriate padding. This works for non-ASCII texts when
i18n.logOutputEncoding is UTF-8 only (independently of a printed commit
message encoding) but does not work when i18n.logOutputEncoding is NOT
UTF-8.
There
Pretty format string %(N,[ml]trunc)%s truncates subject to a given
length with an appropriate padding. This works for non-ASCII texts when
i18n.logOutputEncoding is UTF-8 only (independently of a printed commit
message encoding) but does not work when i18n.logOutputEncoding is NOT
UTF-8.
In
Stepan Kasal wrote:
From: Johannes Schindelin johannes.schinde...@gmx.de
Date: Mon, 9 Apr 2012 13:04:35 -0500
After importing anything with fast-import, we should always let the
garbage collector do its job, since the objects are written to disk
inefficiently.
This brings down an initial
Johannes Schindelin wrote:
Hi,
On Thu, 15 May 2014, Stepan Kasal wrote:
From: Sverre Rabbelier srabbel...@gmail.com
Date: Sat, 28 Aug 2010 20:49:01 -0500
[PT: ensure we add an additional element to the argv array]
IIRC we had problems with refs vs file names.
Actually I tried to
And I thought I was the only one using this :)
diff --git a/pretty.c b/pretty.c
index 6e266dd..7eb43c1 100644
--- a/pretty.c
+++ b/pretty.c
@@ -1500,16 +1500,19 @@ void format_commit_message(const struct commit
*commit,
const struct pretty_print_context
core.commentChar starts with '#' as in default but if it's already in
the prepared message, find another one among a small subset. This
should stop surprises because git strips some lines unexpectedly.
Signed-off-by: Nguyễn Thái Ngọc Duy pclo...@gmail.com
---
Documentation/config.txt | 3 +++
We don't support comment _strings_ (at least not yet). And multi-byte
character encoding could also be misinterpreted.
The test with two commas is deleted because it violates this. It's
added with the patch that introduces core.commentChar in eff80a9
(Allow custom comment char - 2013-01-16). It's
Explicitly set guilt.diffstat to its default value. Without this, the
027 test (and possibly others) fail if guilt.diffstat is set to true
in ~/.gitconfig.
Signed-off-by: Per Cederqvist ced...@opera.com
Signed-off-by: Josef 'Jeff' Sipek jef...@josefsipek.net
---
regression/scaffold | 1 +
1
Signed-off-by: Per Cederqvist ced...@opera.com
Signed-off-by: Josef 'Jeff' Sipek jef...@josefsipek.net
---
guilt-delete | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/guilt-delete b/guilt-delete
index 3e394f8..967ac10 100755
--- a/guilt-delete
+++ b/guilt-delete
@@ -49,7
This is version 3 of the patch series I sent back on 21 Mar 2014. I
have addressed all feedback to date, updated the coding style, and
added signed-off-by lines from Jeff Sipek for those commits that I
have received an explicit approval from him (but only if I have not
made any other change to
Signed-off-by: Per Cederqvist ced...@opera.com
Signed-off-by: Josef 'Jeff' Sipek jef...@josefsipek.net
---
guilt-import-commit | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/guilt-import-commit b/guilt-import-commit
index 20dcee2..f14647c 100755
---
A patch file consists of:
(1) the description
(2) optional diffstat
(3) the patches
When extracting the patch, we only want part 3. The do_get_patch used
to give us part 2 and part 3. That made for instance this series of
operations fail if guilt.diffstat is true:
$ guilt new empty-1
Ensure that the file really is deleted.
Signed-off-by: Per Cederqvist ced...@opera.com
Signed-off-by: Josef 'Jeff' Sipek jef...@josefsipek.net
---
regression/t-026.out | 15 +++
regression/t-026.sh | 5 -
2 files changed, 19 insertions(+), 1 deletion(-)
diff --git
The argument parser arbitrarily refused to accept more than 4
arguments. That made it impossible to run guilt new -f -s -m msg
patch. Removed the checks for the number of arguments from the
guilt new parser -- the other checks that are already there are
enough to catch all errors.
Give a better
Test that we can combine any combination of patches with empty and
non-empty messages, both with and without guilt.diffstat. (All
patches are empty.)
Signed-off-by: Per Cederqvist ced...@opera.com
Signed-off-by: Josef 'Jeff' Sipek jef...@josefsipek.net
---
regression/t-035.out | 467
Test that empty patches are handled correctly, both with and without
the guilt.diffstat configuration option.
Signed-off-by: Per Cederqvist ced...@opera.com
---
regression/t-020.out | 269 +++
regression/t-020.sh | 60
2 files
The cmd and shouldfail functions checked the exit status of the
replace_path function instead of the actual command that was running.
(The $? construct checks the exit status of the last command in a
pipeline, not the first command.)
Updated t-032.sh, which used shouldfail instead of cmd in one
The shouldfail function already redirects stderr to stdout, so there
is no need to do the same in t-028.sh and t-021.sh.
Signed-off-by: Per Cederqvist ced...@opera.com
Signed-off-by: Josef 'Jeff' Sipek jef...@josefsipek.net
---
regression/t-021.sh | 2 +-
regression/t-025.sh | 2 +-
There were two problems with the old code:
- Since set -e is in effect (that is set in scaffold) the run-test
script exited immediately if a t-*.sh script failed. This is not
nice, as we want the error report that test_failed prints.
- The code ran cd - between running the t-*.sh script
Nguyễn Thái Ngọc Duy wrote:
We don't support comment _strings_ (at least not yet). And multi-byte
character encoding could also be misinterpreted.
The test with two commas is deleted because it violates this. It's
added with the patch that introduces core.commentChar in eff80a9
(Allow
The valid_patchname now lets git check-ref-format do its job instead
of trying (and failing) to implement the same rules. See
git-check-ref-format(1) for a list of the rules.
Refer to the git-check-ref-format(1) man page in the error messages
produced when valid_patchname indicates that the name
If you run something like guilt header '.*' the command would crash,
because the grep comand that tries to ensure that the patch exist
would detect a match, but the later code expected the match to be
exact.
Fixed by comparing exact strings.
And as a creeping feature guilt header will now try to
Try harder to create patch names that adhere to the rules in
git-check-ref-format(1) when deriving a patch name from the commit
message. Verify that the derived name using git check-ref-format,
and as a final fallback simply use the patch name x (to ensure that
the code is future-proof in case
The 'echo $s' construct sometimes processes escape sequences. (This
happens, for instance, under Ubuntu 14.04 when /bin/sh is actually
dash.) We don't want that to happen when we are importing commits, so
use 'printf %s $s' instead.
(The -E option of bash that explicitly disables backslash
This fix relies on the fact that git branch names can not contain :.
Signed-off-by: Per Cederqvist ced...@opera.com
Signed-off-by: Josef 'Jeff' Sipek jef...@josefsipek.net
---
guilt-graph | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/guilt-graph b/guilt-graph
index
Give an error message if no patches are applied. Added a test case
that never terminates unless this fix is applied.
Signed-off-by: Per Cederqvist ced...@opera.com
Signed-off-by: Josef 'Jeff' Sipek jef...@josefsipek.net
---
guilt-graph | 9 +++--
regression/t-033.out | 3 +++
git branch names can contain commas. Check that guilt graph works
even in that case.
Signed-off-by: Per Cederqvist ced...@opera.com
Signed-off-by: Josef 'Jeff' Sipek jef...@josefsipek.net
---
regression/t-033.out | 65
regression/t-033.sh |
Use --no-decorate in the call to git log that tries to read the commit
message to produce patch names. Otherwise, if the user has set
log.decorate to short or full, the patch name will be less useful.
Modify the t-034.sh test case to demonstrate that this is needed.
Signed-off-by: Per
Quote quotes with a backslash in the guilt graph output. Otherwise,
the dot file could contain syntax errors.
Added a test case.
Signed-off-by: Per Cederqvist ced...@opera.com
Signed-off-by: Josef 'Jeff' Sipek jef...@josefsipek.net
---
guilt-graph | 2 ++
regression/t-033.out | 22
Only one invocation of disp or _disp actually needed backslash
processing. In quite a few instances, it was wrong to do backslash
processing, as the message contained data derived from the user.
Created the new function disp_e that should be used when backslash
processing is required, and
Signed-off-by: Per Cederqvist ced...@opera.com
Signed-off-by: Josef 'Jeff' Sipek jef...@josefsipek.net
---
guilt-rebase | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/guilt-rebase b/guilt-rebase
index fd28e48..a1714a0 100755
--- a/guilt-rebase
+++ b/guilt-rebase
@@ -66,7
Signed-off-by: Per Cederqvist ced...@opera.com
Signed-off-by: Josef 'Jeff' Sipek jef...@josefsipek.net
---
guilt-patchbomb | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/guilt-patchbomb b/guilt-patchbomb
index 1231418..164b10c 100755
--- a/guilt-patchbomb
+++
This makes it easier to script operations on the entire queue, for
example run the test suite on each patch in the queue:
guilt pop -a;while guilt push; do make test||break; done
This brings guilt push in line with the push operation in Mercurial
Queues (hg qpush), which fails when there are
Fix remove_topic() in t-061.sh so that it doesn't print a git hash.
Signed-off-by: Per Cederqvist ced...@opera.com
Signed-off-by: Josef 'Jeff' Sipek jef...@josefsipek.net
---
regression/t-061.out | 1 -
regression/t-061.sh | 2 +-
2 files changed, 1 insertion(+), 2 deletions(-)
diff --git
Signed-off-by: Per Cederqvist ced...@opera.com
---
regression/t-061.sh | 6 ++
1 file changed, 2 insertions(+), 4 deletions(-)
diff --git a/regression/t-061.sh b/regression/t-061.sh
index db26e12..bda50c7 100755
--- a/regression/t-061.sh
+++ b/regression/t-061.sh
@@ -15,8 +15,7 @@
This is analogous to how guilt push now fails when there are no more
patches to push. Like push, the --all argument still succeeds even
if there was no need to pop anything.
Updated the test suite.
Signed-off-by: Per Cederqvist ced...@opera.com
Signed-off-by: Josef 'Jeff' Sipek
When the option is true, Guilt does not create a new Git branch when
patches are applied. This way, you can switch between Guilt 0.35 and
the current version of Guilt with no issues.
By default, the option is false, so that all users will immediately
get the added protection against pushing a
Signed-off-by: Per Cederqvist ced...@opera.com
Signed-off-by: Josef 'Jeff' Sipek jef...@josefsipek.net
---
.dir-locals.el | 3 +++
Documentation/Contributing | 15 +++
2 files changed, 18 insertions(+)
create mode 100644 .dir-locals.el
diff --git a/.dir-locals.el
Fedex Delivery Post Open Attachments File and Contact:
fedex-courierservic...@fengv.commailto:fedex-courierservic...@fengv.com
FEDEX.docx
Description: FEDEX.docx
Signed-off-by: Per Cederqvist ced...@opera.com
Signed-off-by: Josef 'Jeff' Sipek jef...@josefsipek.net
---
regression/t-028.out | 7 +++
regression/t-028.sh | 4
2 files changed, 11 insertions(+)
diff --git a/regression/t-028.out b/regression/t-028.out
index 1564c09..ea72a3a 100644
---
Signed-off-by: Josef 'Jeff' Sipek jef...@josefsipek.net
On Fri, May 16, 2014 at 04:45:55PM +0200, Per Cederqvist wrote:
Test that empty patches are handled correctly, both with and without
the guilt.diffstat configuration option.
Signed-off-by: Per Cederqvist ced...@opera.com
---
On Thu, May 15, 2014 at 5:20 PM, Jonathan Nieder jrnie...@gmail.com wrote:
Ronnie Sahlberg wrote:
Change ref_transaction_commit so that it does not free the transaction.
Instead require that a caller will end a transaction by either calling
ref_transaction_rollback or ref_transaction_free.
On Fri, May 16, 2014 at 8:02 AM, Ronnie Sahlberg sahlb...@google.com wrote:
On Thu, May 15, 2014 at 5:20 PM, Jonathan Nieder jrnie...@gmail.com wrote:
Ronnie Sahlberg wrote:
Change ref_transaction_commit so that it does not free the transaction.
Instead require that a caller will end a
Signed-off-by: Josef 'Jeff' Sipek jef...@josefsipek.net
On Fri, May 16, 2014 at 04:45:59PM +0200, Per Cederqvist wrote:
If you run something like guilt header '.*' the command would crash,
because the grep comand that tries to ensure that the patch exist
would detect a match, but the later
Signed-off-by: Josef 'Jeff' Sipek jef...@josefsipek.net
On Fri, May 16, 2014 at 04:46:16PM +0200, Per Cederqvist wrote:
When the option is true, Guilt does not create a new Git branch when
patches are applied. This way, you can switch between Guilt 0.35 and
the current version of Guilt with
On Fri, May 16, 2014 at 04:46:01PM +0200, Per Cederqvist wrote:
The valid_patchname now lets git check-ref-format do its job instead
of trying (and failing) to implement the same rules. See
git-check-ref-format(1) for a list of the rules.
Refer to the git-check-ref-format(1) man page in the
In this reroll (against v1) remarks of Nguyễn are respected.
Comments style changed from C++ to C. variable declaration moved back to
the beginning of a function.
Also, added tests for the same case for git rev-list
(see t6006-rev-list-format.sh)
Alexey Shumkin (2):
t4205, t6006: Add failing
Pretty format string %(N,[ml]trunc)%s truncates subject to a given
length with an appropriate padding. This works for non-ASCII texts when
i18n.logOutputEncoding is UTF-8 only (independently of a printed commit
message encoding) but does not work when i18n.logOutputEncoding is NOT
UTF-8.
There
Pretty format string %(N,[ml]trunc)%s truncates subject to a given
length with an appropriate padding. This works for non-ASCII texts when
i18n.logOutputEncoding is UTF-8 only (independently of a printed commit
message encoding) but does not work when i18n.logOutputEncoding is NOT
UTF-8.
In
On Fri, May 16, 2014 at 04:45:56PM +0200, Per Cederqvist wrote:
The cmd and shouldfail functions checked the exit status of the
replace_path function instead of the actual command that was running.
(The $? construct checks the exit status of the last command in a
pipeline, not the first
Nguyễn Thái Ngọc Duy wrote:
--- a/config.c
+++ b/config.c
@@ -826,8 +826,12 @@ static int git_default_core_config(const char *var,
const char *value)
if (!strcmp(var, core.commentchar)) {
const char *comment;
int ret = git_config_string(comment, var,
Nguyễn Thái Ngọc Duy wrote:
core.commentChar starts with '#' as in default but if it's already in
the prepared message, find another one among a small subset. This
should stop surprises because git strips some lines unexpectedly.
Probably worth mentioning this only kicks in if someone
Add an err argument to delete_loose_ref so that we can pass a descriptive
error string back to the caller. Pass the err argument from transaction
commit to this function so that transaction users will have a nice error
string if the transaction failed due to delete_loose_ref.
Add a new function
This patch series can also be found at
https://github.com/rsahlberg/git/tree/ref-transactions
This patch series is based on next and expands on the transaction API. It
converts all ref updates, inside refs.c as well as external, to use the
transaction API for updates. This makes most of the ref
Change to use ref transactions for all updates to refs.
Signed-off-by: Ronnie Sahlberg sahlb...@google.com
---
sequencer.c | 24
1 file changed, 16 insertions(+), 8 deletions(-)
diff --git a/sequencer.c b/sequencer.c
index 0a80c58..b047fb8 100644
--- a/sequencer.c
+++
Change create_branch to use a ref transaction when creating the new branch.
ref_transaction_create will check that the ref does not already exist and fail
otherwise meaning that we no longer need to keep a lock on the ref during the
setup_tracking. This simplifies the code since we can now do the
Change ref_transaction_commit so that it does not free the transaction.
Instead require that a caller will end a transaction by calling
ref_transaction_free.
By having the transaction object remaining valid after _commit returns allows
us to write much nicer code and still be able to call
Change the update_ref helper function to use a ref transaction internally.
Signed-off-by: Ronnie Sahlberg sahlb...@google.com
---
refs.c | 25 +
1 file changed, 21 insertions(+), 4 deletions(-)
diff --git a/refs.c b/refs.c
index a588194..78312b5 100644
--- a/refs.c
+++
Change update_branch() to use ref transactions for updates.
Signed-off-by: Ronnie Sahlberg sahlb...@google.com
---
fast-import.c | 25 +++--
1 file changed, 15 insertions(+), 10 deletions(-)
diff --git a/fast-import.c b/fast-import.c
index 6707a66..60d4538 100644
---
Change delete_ref to use a ref transaction for the deletion. At the same time
since we no longer have any callers of repack_without_ref we can now delete
this function.
Signed-off-by: Ronnie Sahlberg sahlb...@google.com
---
refs.c | 31 ++-
1 file changed, 10
Change tag.c to use ref transactions for all ref updates.
Signed-off-by: Ronnie Sahlberg sahlb...@google.com
---
builtin/tag.c | 14 --
1 file changed, 8 insertions(+), 6 deletions(-)
diff --git a/builtin/tag.c b/builtin/tag.c
index c6e8a71..b05f9a5 100644
--- a/builtin/tag.c
+++
Add a new flag REF_ISPACKONLY that we can use in ref_transaction_delete.
This flag indicates that the ref does not exist as a loose ref andf only as
a packed ref. If this is the case we then change the commit code so that
we skip taking out a lock file and we skip calling delete_ref_loose.
Check
Change s_update_ref to use a ref transaction for the ref update.
Signed-off-by: Ronnie Sahlberg sahlb...@google.com
---
builtin/fetch.c | 17 +
1 file changed, 9 insertions(+), 8 deletions(-)
diff --git a/builtin/fetch.c b/builtin/fetch.c
index a93c893..8cf70cd 100644
---
Signed-off-by: Ronnie Sahlberg sahlb...@google.com
---
fast-import.c | 17 -
1 file changed, 12 insertions(+), 5 deletions(-)
diff --git a/fast-import.c b/fast-import.c
index cb3f5af..9c76c73 100644
--- a/fast-import.c
+++ b/fast-import.c
@@ -1735,15 +1735,22 @@ static void
Signed-off-by: Ronnie Sahlberg sahlb...@google.com
---
refs.c | 2 --
1 file changed, 2 deletions(-)
diff --git a/refs.c b/refs.c
index 71e3059..bf84306 100644
--- a/refs.c
+++ b/refs.c
@@ -260,8 +260,6 @@ struct ref_entry {
};
static void read_loose_refs(const char *dirname, struct ref_dir
Allow passing a list of refs to skip checking to name_conflict_fn.
There are some conditions where we want to allow a temporary conflict and skip
checking those refs. For example if we have a transaction that
1, guarantees that m is a packed refs and there is no loose ref for m
2, the transaction
Since we only call update_ref_write from a single place and we only call it
with onerr==QUIET_ON_ERR we can just as well get rid of it and just call
write_ref_sha1 directly.
Signed-off-by: Ronnie Sahlberg sahlb...@google.com
---
refs.c | 35 +--
1 file changed, 9
Change rename_ref to use a single transaction to perform the ref rename.
Signed-off-by: Ronnie Sahlberg sahlb...@google.com
---
refs.c | 73 ++
1 file changed, 20 insertions(+), 53 deletions(-)
diff --git a/refs.c b/refs.c
index
This means that most loose refs will no longer be present after the rename
which triggered a test failure since it assumes the file for an unrelated
ref would still be present after the rename.
Signed-off-by: Ronnie Sahlberg sahlb...@google.com
---
refs.c| 3 +++
t/t3200-branch.sh |
lock_ref_sha1 was only called from one place in refc.c and only provided
a check that the refname was sane before adding back the initial refs/
part of the ref path name, the initial refs/ that this caller had already
stripped off before calling lock_ref_sha1.
Signed-off-by: Ronnie Sahlberg
Move the check for check_refname_format from lock_any_ref_for_update
to lock_ref_sha1_basic. At some later stage we will get rid of
lock_any_ref_for_update completely.
This leaves lock_any_ref_for_updates as a no-op wrapper which could be removed.
But this wrapper is also called from an external
1 - 100 of 155 matches
Mail list logo