On 18.03.2015 10:42, Jeff King wrote:
On Wed, Mar 18, 2015 at 09:11:48AM +0100, Ephrim Khong wrote:
I have a non-bare repository /home/a set up with an alternate to the bare
repository /b. Running git gc on /home/a produces below's error
[...]
git --version
git version 2.3.0
Try v
I have a non-bare repository /home/a set up with an alternate to the
bare repository /b. Running git gc on /home/a produces below's error
error: unable to open
/b/objects/56/b969ffdf64343777a069260f41761dc0551bfa/00: Not a directory
The referenced file
/b/objects/56/b969ffdf64343777a06
Without having looked into this and nd/multiple-work-trees, but with
"make multiple checkouts aware of each other" in mind: Could this
mechanism be re-used to make alternates aware of each other, to mitigate
the dangers of having git gc on an alternate remove objects that are
used by a refere
On 15.07.2014 21:48, Junio C Hamano wrote:
Ephrim Khong writes:
+test_expect_success setup '
+ GIT_OBJECT_DIRECTORY=.git//../.git/objects &&
+ export GIT_OBJECT_DIRECTORY &&
Do you need this artificially strange environment settings for the
problem to mani
On 15.07.2014 21:26, Junio C Hamano wrote:
+ strbuf_addstr(&objdirbuf, absolute_path(get_object_directory()));
+ normalize_path_copy(objdirbuf.buf, objdirbuf.buf);
This is somewhat a strange usage of a strbuf.
There might be a more elegant way, but I tried to mimic the local codin
When adding alternate object directories, we try not to add the
directory of the current repository to avoid cycles. Unfortunately,
that test was broken, since it compared an absolute with a relative
path.
Signed-off-by: Ephrim Khong
---
Since v2: Added Johannes' comments.
sha1_f
When adding alternate object directories, we try not to add the
directory of the current repository to avoid cycles. Unfortunately,
that test was broken, since it compared an absolute with a relative
path.
Signed-off-by: Ephrim Khong
---
As proposed by Duy, v2 of the patch normalizes the object
When adding alternate object directories, we try not to add the
directory of the current repository to avoid cycles. Unfortunately,
that test was broken, since it compared an absolute with a relative
path.
Signed-off-by: Ephrim Khong
---
My first patch, so be harsh. I'm not sure abou
Am 11.07.14 18:01, schrieb Junio C Hamano:
Ephrim Khong writes:
git seems to have issues with alternates when cycles are present (repo
A has B/objects as alternates, B has A/objects as alternates).
Yeah, don't do that. A thinks "eh, the other guy must have it" and
B think
Hi,
git seems to have issues with alternates when cycles are present (repo A
has B/objects as alternates, B has A/objects as alternates). In such
cases, gc and repack might delete objects that are present in only one
of the alternates, leading to data loss.
I understand that this is no big u
Am 20.03.14 20:54, schrieb Jeff King:
On Thu, Feb 20, 2014 at 08:35:33AM +0100, Ephrim Khong wrote:
Hi, git log seems to omit merge commits that delete a file if --follow or
--diff-filter=D is given. Below is a testcase. I'm not sure if it is desired
behaviour for --diff-filter=D, but
Hi, git log seems to omit merge commits that delete a file if --follow
or --diff-filter=D is given. Below is a testcase. I'm not sure if it is
desired behaviour for --diff-filter=D, but it's probably not correct
that --follow _removes_ the merge commit from the log output.
Thanks - Eph
--
git
Hi,
for files that contain windows line endings in a repository with
core.autocrlf=input, git blame will show lines as "Not Committed Yet",
even though they were not modified.
Example:
--
git init
git config core.autocrlf false
echo "foo" > a
unix2dos a
git add a
git commit -m "initial commi
13 matches
Mail list logo
