git/Documentation/RelNotes/1.7.7.1.txt
Jonathan Nieder 7fe74f52f9 RelNotes/1.7.7.1: setgid bit patch is about fixing "git init" via Makefile setting
The change was actually about "git init -s" which sets the setgid bit on
SysV-style systems to allow shared access to a repository, and can provoke
errors on BSD-style systems, depending on how permissive the filesystem in
use wants to be.

More to the point, the patch was just taking a fix that arrived for
FreeBSD in v1.5.5 days and making it also apply to machines using an
(obscure) GNU userland/FreeBSD kernel mixture.

Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-10-23 21:16:31 -07:00

61 lines
2.4 KiB
Plaintext

Git v1.7.7.1 Release Notes
==========================
Fixes since v1.7.7
------------------
* On some BSD systems, adding +s bit on directories is detrimental
(it is not necessary on BSD to begin with). "git init --shared"
has been updated to take this into account without extra makefile
settings on platforms the Makefile knows about.
* After incorrectly written third-party tools store a tag object in
HEAD, git diagnosed it as a repository corruption and refused to
proceed in order to avoid spreading the damage. We now gracefully
recover from such a situation by pretending as if the commit that
is pointed at by the tag were in HEAD.
* "git apply --whitespace=error" did not bother to report the exact
line number in the patch that introduced new blank lines at the end
of the file.
* "git apply --index" did not check corrupted patch.
* "git checkout $tree $directory/" resurrected paths locally removed or
modified only in the working tree in $directory/ that did not appear
in $directory of the given $tree. They should have been kept intact.
* "git diff $tree $path" used to apply the pathspec at the output stage,
reading the whole tree, wasting resources.
* The code to check for updated submodules during a "git fetch" of the
superproject had an unnecessary quadratic loop.
* "git fetch" from a large bundle did not enable the progress output.
* When "git fsck --lost-and-found" found that an empty blob object in the
object store is unreachable, it incorrectly reported an error after
writing the lost blob out successfully.
* "git filter-branch" did not refresh the index before checking that the
working tree was clean.
* "git grep $tree" when run with multiple threads had an unsafe access to
the object database that should have been protected with mutex.
* The "--ancestry-path" option to "git log" and friends misbehaved in a
history with complex criss-cross merges and showed an uninteresting
side history as well.
* Test t1304 assumed LOGNAME is always set, which may not be true on
some systems.
* Tests with --valgrind failed to find "mergetool" scriptlets.
* "git patch-id" miscomputed the patch-id in a patch that has a line longer
than 1kB.
* When an "exec" insn failed after modifying the index and/or the working
tree during "rebase -i", we now check and warn that the changes need to
be cleaned up.