What's in git.git (Sep 2008, #01; Sat, 06)

Quite a many minor fixes appeared on 'maint'. 1.6.0.2 needs to happen soon. Any favorite fixes (not feature enhancements) not on the list here?

On the 'master' front, gitk, git-gui and git-svn subsystems had moderate amount of updates. Among other things:

 - From gitk you can ask git-gui to show blame, and from git-gui's blame
   you can ask gitk to show the commit and history;

 - "log --simplify-merges";

 - "commit --author=$name" does not require "A U Thor <author@email.xz>"
   in full; abbreviated request will look for a matching one in the
   existing history;

 - "pack-objects" is safer against disc corruption;

* The 'maint' branch has these fixes since the last announcement.

Alex Riesen (1):
  Fix use of hardlinks in "make install"

Alexander Gavrilov (1):
  git-gui: Fix string escaping in po2msg.sh

Alexandre Bourget (2):
  git-gui: Update french translation
  git-gui: update all remaining translations to French.

Andreas Färber (1):
  Makefile: always provide a fallback when hardlinks fail

Ask Bjørn Hansen (1):
  Document sendemail.envelopesender configuration

Björn Steinbrink (1):
  for-each-ref: Allow a trailing slash in the patterns

Clemens Buchacher (2):
  git gui: show diffs with a minimum of 1 context line
  clone: fix creation of explicitly named target directory

Gustaf Hendeby (1):
  Document clarification: gitmodules, gitattributes

Heikki Orsila (2):
  Start conforming code to "git subcmd" style
  Improve documentation for --dirstat diff option

Jeff King (2):
  pretty=format: respect date format options
  checkout: fix message when leaving detached HEAD

Johan Herland (1):
  Bring local clone's origin URL in line with that of a remote clone

Jonas Fonseca (1):
  Fix passwd(5) ref and reflect that commit doens't use commit-tree

Junio C Hamano (11):
  shell: do not play duplicated definition games to shrink the executable
  Fix example in git-name-rev documentation
  git-apply: Loosen "match_beginning" logic
  gitattributes: -crlf is not binary
  'git foo' program identifies itself without dash in die() messages
  Start 1.6.0.2 maintenance cycle
  diff --cumulative is a sub-option of --dirstat
  "blame -c" should be compatible with "annotate"
  Mention the fact that 'git annotate' is only for backward compatibility.
  stash: refresh the index before deciding if the work tree is dirty
  Update draft release notes for 1.6.0.2

Linus Torvalds (1):
  Fix '--dirstat' with cross-directory renaming

Miklos Vajna (1):
  Makefile: add merge_recursive.h to LIB_H

Nanako Shiraishi (4):
  tests: use "git xyzzy" form (t0000 - t3599)
  tests: use "git xyzzy" form (t3600 - t6999)
  tests: use "git xyzzy" form (t7000 - t7199)
  tests: use "git xyzzy" form (t7200 - t9001)

Nguyễn Thái Ngọc Duy (5):
  diff*: fix worktree setup
  grep: fix worktree setup
  read-tree: setup worktree if merge is required
  update-index: fix worktree setup
  setup_git_directory(): fix move to worktree toplevel directory

Nicolas Pitre (2):
  discard revindex data when pack list changes
  improve handling of sideband message display

Paolo Bonzini (1):
  make git-shell paranoid about closed stdin/stdout/stderr

Paolo Ciarrocchi (1):
  tutorial: gentler illustration of Alice/Bob workflow using gitk

Petr Baudis (1):
  bash completion: Hide more plumbing commands

SZEDER Gábor (3):
  Documentation: fix reference to a for-each-ref option
  Documentation: fix disappeared lines in 'git stash' manpage
  Documentation: minor cleanup in a use case in 'git stash' manual

Shawn O. Pearce (2):
  pack-objects: Allow missing base objects when creating thin packs
  git-gui: Fix diff parsing for lines starting with "--" or "++"

Teemu Likonen (1):
  config.txt: Add missing colons after option name

Yann Dirson (1):
  Document gitk --argscmd flag.

* The 'master' branch has these since the last announcement
  in addition to the above.

Alex Riesen (5):
  Remove calculation of the longest command name from where it is not used
  Remove useless memset of static command name lists in builtin-merge.c
  Make main_cmds and other_cmds local to builtin-help.c
  Allow git help work without PATH set
  list_commands: only call git_exec_path if it is needed

Alexander Gavrilov (14):
  git-gui: Support starting gitk from Gui Blame
  git-gui: Support passing blame to a parent commit.
  git-gui: Better positioning in Blame Parent Commit
  git-gui: Allow specifying an initial line for git gui blame.
  gitk: Add option to specify the default commit on command line
  gitk: Add menu item for calling git gui blame
  git-gui: Support resolving conflicts via the diff context menu.
  git-gui: Support calling merge tools.
  git-gui: Don't allow staging files with conflicts.
  git-gui: Support more merge tools.
  git-gui: Support conflict states _U & UT.
  git-gui: Reimplement and enhance auto-selection of diffs.
  git-gui: Make F5 reselect a diff, if an untracked file is selected.
  git-gui: Show special diffs for complex conflict cases.

Brandon Casey (1):
  dir.c: Avoid c99 array initialization

Christian Stimming (1):
  git-gui: Mark forgotten strings for translation.

David Soria Parra (1):
  cast pid_t's to uintmax_t to improve portability

Eric Wong (2):
  git-svn: check error code of send_txstream
  t9124: clean up chdir usage

Florian Weimer (3):
  git-svn: extract base blob in generate_diff
  git-svn: Introduce SVN::Git::Editor::_chg_file_get_blob
  git-svn: Send deltas during commits

Giuseppe Bilotta (1):
  gitweb: ref markers link to named shortlogs

Gustaf Hendeby (1):
  git-gui: Teach git gui about file type changes

Junio C Hamano (11):
  revision.c: whitespace fix
  revision traversal: show full history with merge simplification
  Topo-sort before --simplify-merges
  revision --simplify-merges: use decoration instead of commit->util field
  revision --simplify-merges: do not leave commits unprocessed
  revision --simplify-merges: make it a no-op without pathspec
  daemon.c: minor style fixup
  git commit --author=$name: look $name up in existing commits
  Update draft release notes for 1.6.1
  log --author/--committer: really match only with name part
  Update draft release notes for 1.6.1

Matthias Kestenholz (1):
  bash-completion: Add all submodule subcommands to the completion list

Nicolas Pitre (5):
  pack-objects: improve returned information from write_one()
  improve reliability of fixup_pack_header_footer()
  pack-objects: use fixup_pack_header_footer()'s validation mode
  index-pack: use fixup_pack_header_footer()'s validation mode
  fixup_pack_header_footer(): use nicely aligned buffer sizes

Paul Talacko (1):
  git-svn: set auto_props when renaming files

Romain Francoise (1):
  builtin-help: fallback to GIT_MAN_VIEWER before man

Stephen R. van den Berg (4):
  git-daemon: call logerror() instead of error()
  git-daemon: use LOG_PID, simplify logging code
  git-daemon: Simplify dead-children reaping logic
  git-daemon: rewrite kindergarden, new option --max-connections

Thomas Rast (13):
  filter-branch: Extend test to show rewriting bug
  filter-branch: fix ref rewriting with --subdirectory-filter
  filter-branch: use --simplify-merges
  Documentation: rev-list-options: move --simplify-merges documentation
  git svn: catch lack of upstream info for dcommit earlier
  git svn info: tests: let 'init' test run with SVN 1.5
  git svn info: tests: do not use set -e
  git svn info: tests: use test_cmp instead of git-diff
  git svn info: tests: fix ptouch argument order in setup
  git svn info: make info relative to the current directory
  git svn info: always quote URLs in 'info' output
  git-svn: match SVN 1.5 behaviour of info' on unknown item
  git-svn: fix 'info' tests for unknown items

Tor Arvid Lund (1):
  git-p4: Fix checkout bug when using --import-local.

On Sat, 6 Sep 2008, Junio C Hamano wrote:

I would really like to see this graduate to master before 1.6.0.2

* jc/setlinebuf-setvbuf (Wed Sep 3 20:33:29 2008 -0700) 1 commit
  + daemon.c: avoid setlinebuf()

-- Boyd Gerber <gerberb@zenez.com> ZENEZ 1042 East Fort Union #135, Midvale Utah 84047

Boyd Lynn Gerber <gerberb@zenez.com> writes:

I do not think the setlinebuf() change affected 'maint' to begin with; it was part of the series to clean-up daemon and to make maximum simultanous number of connections configurable. Are you having trouble building and/ or running 1.6.0.1 on your box?

On Sun, 7 Sep 2008, Junio C Hamano wrote:

No, it does not work. I have the git-shell problems, that I reported earlier with all the missing/undefined symbols. Same problems with 1.6.0 release. That is why I was hoping to have a master that would be released that is clean and able to build. I am currently running on all my systems the master with this commit. It is the only way for me to run any git 1.6.x.x on any of the 12 platforms. I have been getting master and then adding this to compile. I have it running on 4 platforms daily doing an pull and rebuilding git. I have testing now working on 2 platforms on next.

Thanks,

-- Boyd Gerber <gerberb@zenez.com> ZENEZ 1042 East Fort Union #135, Midvale Utah 84047

Boyd Lynn Gerber <gerberb@zenez.com> writes:

Then it does not have anything to do with the setlinebuf-setvbuf topic at all, does it?

I think we already have 6ffaecc (shell: do not play duplicated definition games to shrink the executable, 2008-08-19), which was cherry-picked from 4cfc24a (shell: do not play duplicated definition games to shrink the executable, 2008-08-19) on 'master'. Do you still have problem building 'maint' on your boxes?

On Sun, 7 Sep 2008, Junio C Hamano wrote:

No, maint does build. I just thought your were going to release 1.6.0.2 from master. I think the next commit for set-linebug-setvbug needs to be in it for release.

Thanks,

-- Boyd Gerber <gerberb@zenez.com> ZENEZ 1042 East Fort Union #135, Midvale Utah 84047

Boyd Lynn Gerber <gerberb@zenez.com> writes:

Thanks for clarifying, so as far as you are concerned, 'maint', which is the only thing that matters for 1.6.0.2, is in a good shape ;-)

On Sun, 7 Sep 2008, Junio C Hamano wrote:

Yes, I guess I miss-understood your maintainer notes. I re-read them before posting. I thought I read that you branch off master for the point releases. I guess that master would be only for 1.6.1 not 1.6.0.2. Sorry for my miss-understanding.

Thanks,

-- Boyd Gerber <gerberb@zenez.com> ZENEZ 1042 East Fort Union #135, Midvale Utah 84047