Bug#637840: pu: package git/1:1.7.2.5-3
Package: release.debian.org
User: release.debian.org@packages.debian.org
Usertags: pu
Tags: squeeze
Hi,
A few updates have collected on git's debian-stable branch.
1. Upstream's maint-1.7.2 branch gets very few changes, but there have
been a few since 1.7.2.5 was released:
- fix off-by-one bug that makes git read past the end of a buffer
when extracting the first line from an empty commit message (and
include an extra line when the first line is blank)
- two typo fixes and one clarification in documentation
- fast-import: accept no-op "feature notes" command for frontends
use to declare they require an importer able to write notes.
Of those, the fast-import change probably seems iffy (since it does
not fix a critical bug) but I would prefer to include it to match
upstream.
2. git-daemon/run: use SO_REUSEADDR to allow restarting the service
with connections active or recently closed (thanks to Daniel Kahn
Gillmor). Bug#609405 has details.
3. git-daemon-run.postrm purge: always terminate logging process more
aggressively so the logging user can be removed and the package
purged when a connection is active (also thanks to dkg, Bug#627314).
4. Fixes a server-side deadlock when performing a shallow clones
that people had been running into on git.sv.gnu.org (see [1], [2]).
The motivation is mainly that fourth change (preventing shallow clones
from hanging).
These changes have been in sid for at least two months, wheezy for a
month and a half. debdiff attached, or see
http://repo.or.cz/w/git/debian/jrn.git/commitdiff/debian-stable?hp=debian-1.7.2.5-2
Diffstat:
debian/diff/0034-revert-fix-off-by-one-read-when-searching-the-end-of-.diff | 71 +++++++
debian/diff/0035-revert-refactor-code-to-find-commit-subject-in-find_c.diff | 95 +++++++++
debian/diff/0036-revert-rename-variables-related-to-subject-in-get_mes.diff | 57 +++++
debian/diff/0037-bisect-use-find_commit_subject-instead-of-custom-code.diff | 48 ++++
debian/diff/0038-merge-recursive-use-find_commit_subject-instead-of-cu.diff | 42 ++++
debian/diff/0039-blame-use-find_commit_subject-instead-of-custom-code.diff | 59 +++++
debian/diff/0040-Documentation-git-archive-spell-worktree-attributes-c.diff | 38 +++
debian/diff/0041-Documentation-githooks-post-rewrite-copy-notes-never-.diff | 42 ++++
debian/diff/0042-fast-import-clarify-documentation-of-feature-command.diff | 79 +++++++
debian/diff/0043-fast-import-introduce-feature-notes-command.diff | 85 ++++++++
debian/diff/0044-upload-pack-start-pack-objects-before-async-rev-list.diff | 99 ++++++++++
debian/changelog | 21 ++
debian/git-daemon-run.postrm | 3
debian/git-daemon/run | 3
14 files changed, 741 insertions(+), 1 deletion(-)
Effect of patches:
Documentation/git-archive.txt | 2 +-
Documentation/git-fast-import.txt | 47 +++++++++++++++++++-----------------
Documentation/githooks.txt | 4 ---
bisect.c | 13 ++++------
builtin/blame.c | 22 +++++-----------
builtin/revert.c | 20 ++++-----------
commit.c | 19 +++++++++++++++
commit.h | 3 ++
fast-import.c | 2 +
merge-recursive.c | 14 +++--------
t/t3505-cherry-pick-empty.sh | 20 +++++++++++++++-
t/t9301-fast-import-notes.sh | 1 +
upload-pack.c | 23 ++++++++---------
13 files changed, 102 insertions(+), 88 deletions(-)
What do you think? Would this be reasonable for upload to stable, or
would it be better to leave out the less important upstream changes
(fast-import.c and Documentation/)?
Sorry to take so long to get to this.
Jonathan
[1] http://thread.gmane.org/gmane.comp.version-control.git/172042
[2] http://thread.gmane.org/gmane.comp.version-control.git/170789
Reply to: