| Junio C Hamano | What's in git.git today
GIT 1.0.6 was silently pushed out. Again, it has one notable
fix among minor test fixes.
- Do not mark tags fetched via --tags flag as mergeable.
This is to fix "git pull --tags"; it tried to merge all of
the tags by mistake --- it should not merge any of them.
I have verified that the patch by Gerrit to fix merge-order test
is correct, but have not got around to apply it yet.
The master branch has the following enhancements in addition to
what is in 1.0.6. I intend to have 1.1 ...
| Dec 28, 4:32 pm 2005 |
| Gerrit Pape | [PATCH] fix skipping merge-order test with NO_OPENSSL=1.
Hi, when building git with NO_OPENSSL=1, the rev-list-merge-order
selftests fail. Moving the check for 'OpenSSL not linked' in the usage
output to after test no 1 and fixing the syntax works for me.
Regards, Gerrit.
* move git-rev-list --merge-order usage check for 'OpenSSL not linked' after
test 1.
Signed-off-by: Gerrit Pape <pape@smarden.org>
---
t/t6001-rev-list-merge-order.sh | 14 +++++++-------
1 files changed, 7 insertions(+), 7 ...
| Dec 28, 4:25 am 2005 |
| Simon Richter | Re: git-inject-tarball
Hi,
Well, the most important thing for such a tool IMO would be that it
wouldn't overwrite any uncommitted changes, so it's not as trivial as it
seems.
Simon
| Dec 28, 8:51 am 2005 |
| Johannes Schindelin | Re: git-inject-tarball
Hi,
Something like this?
-- snip --
#!/bin/sh
TARFILE="$1"
test -z "$TARFILE" && TARFILE=-
git-ls-files -z | xargs -0 rm
tar xvf "$TARFILE" | xargs git add
git-update-index --remove $(git-ls-files)
-- snap --
Notes:
- this updates the index to match the given tarfile, you still
have to commit yourself
- it does not handle compressed tarfiles (you have to do something
like "cat bla.tar.gz | gzip -d | sh git-untar.sh")
- it does not remove empty directories (i.e. if ...
| Dec 28, 4:42 am 2005 |
| Johannes Schindelin | Re: git-inject-tarball
Hi,
Ah, but that's easy:
test -z "$(git-diff-index --name-only HEAD)" || exit 1
(This is just the idea; you should also output an error message.)
Hth,
Dscho
-
| Dec 28, 9:18 am 2005 |
| Adrien Beau | Re: [ANNOUNCE] GIT preformatted documentation available.
Having preformatted documentation is a great step forward, but it
would be even better if we could have it without installing some weird
RPM tool or a local copy of the Git repository -- not all my machines
have that.
Besides, it's not very beginner- or newcomer-friendly. Grab the
source; discover you need a complex toolchain to have manpages; maybe
learn sometime later that "all you had to do" was get an RPM and
explode some parts of it -- nah, it just doesn't feel right.
You're already ...
| Dec 28, 9:51 am 2005 |
| Kalle Valo | Re: [PATCH] debian/ directory
One upside in Junio's debian packages is that they[1] work also in
sarge[2]. If he stops building them, we sarge users will miss them.
Any volunteers to maintain them, for example, through backports.org?
[1] at least git-core and git-doc work
[2] Debian 3.1
--
Kalle Valo
-
| Dec 28, 9:17 am 2005 |
| Gerrit Pape | [PATCH] debian/ directory
Thanks. The most notable change is that debhelper isn't used to do the
packaging stuff, and the build-dependency has been dropped; this is
after my personal preference.
Below is the diff for reference, I think though it might be better if
you don't include the debian/ directory in the upstream package at all.
It may well be that I'm not always available immediately if things need
to be changed in debian/, but do as you wish.
Regards, Gerrit.
* don't use debhelper, use implicit ...
| Dec 28, 4:20 am 2005 |
| Johannes Schindelin |
| Dec 28, 1:38 pm 2005 |
| H. Peter Anvin |
| Dec 28, 9:58 am 2005 |
| Junio C Hamano | Re: [PATCH 2/4] short circuit out of a few places where ...
Yes, that has essentially been the plan (according to the
discussion lead to 7e4a2a848377241b8fb4f624d1151bbf2f8d5814
commit on the list).
After eradicating zero-sized allocations where that change makes
the overall code cleaner (which Johannes and Eric did most of
the heavylifting and I think mostly done), we would apply
something like this, instead of doing x*alloc(size ? size : 1)
at the calling site.
About die(), I think the current code structure is fine. If we
were doing a library, ...
| Dec 27, 10:07 pm 2005 |
