What's in git.git (stable)

* The 'maint' branch has these fixes since v1.5.0.1

 Christian Schlotter (1):
  git-clone: Sync documentation to usage note.

 Jason Riedy (2):
  Add a compat/strtoumax.c for Solaris 8.
  Obey NO_C99_FORMAT in fast-import.c.

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

 Andy Parkins (1):
  Have git-cvsserver call hooks/update before really altering the ref

 Fredrik Kuivinen (2):
  Read the config in rev-list
  Documentation/i18n.txt: it is i18n.commitencoding not core.commitencoding

 Johannes Schindelin (2):
  name-rev: introduce the --refs=<pattern> option
  diff --check: use colour

 Junio C Hamano (9):
  Make gitk work reasonably well on Cygwin.
  gitk: Use show-ref instead of ls-remote
  remotes.not-origin.tagopt
  pretend-sha1: grave bugfix.
  git-merge: minor fix for no_trivial_merge_strategies.
  Do not take mode bits from index after type change.
  Update draft release notes for 1.5.0.1
  Update draft release notes for 1.5.1
  GIT 1.5.0.1

 Mark Levedahl (3):
  gitk - remove trailing whitespace from a few lines.
  Make gitk save and restore the user set window position.
  Make gitk save and restore window pane position on Linux and Cygwin.

 Paul Mackerras (1):
  Change git repo-config to git config

 Shawn O. Pearce (2):
  Attempt to improve git-rebase lead-in description.
  Convert update-index references in docs to add.

 Theodore Ts'o (1):
  Teach git-remote to update existing remotes by fetching from them

 Tommi Kyntola (1):
  git-blame: prevent argument parsing segfault

On the 'maint' front, there aren't that many 'grave fixes' anymore, which is somewhat (happily) surprising after a big release like 1.5.0. Shawn has updated git-gui to 0.6.1. I'll perhaps do 1.5.0.2 next week.

In 'master', I pushed out autoCRLF and git-apply fixes. These are changes people can easily notice breakage and would benefit from wider exposure. Hopefully we can have simple path based text/binary attribute on top of them -- I am hoping it to come from people who actually benefit from autoCRLF ;-).

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

 Fredrik Kuivinen (1):
  Fix 'git commit -a' in a newly initialized repository

 Jason Riedy (1):
  Check for PRIuMAX rather than NO_C99_FORMAT in fast-import.c.

 Johannes Schindelin (1):
  git-diff: fix combined diff

 Martin Koegler (1):
  git-gui: Create new branches from a tag.

 Michael Loeffler (1):
  Use gunzip -c over gzcat in import-tars example.

 Shawn O. Pearce (16):
  git-gui: Refactor 'exec git subcmd' idiom.
  git-gui: Basic version check to ensure git 1.5.0 or later is used.
  git-gui: Permit merging tags into the current branch.
  git-gui: More consistently display the application name.
  git-gui: Print version on the console.
  git-gui: Prefer version file over git-describe.
  git-gui: Expose the browser as a subcommand.
  git-gui: Correct crash when saving options in blame mode.
  git-gui: Use mixed path for docs on Cygwin.
  git-gui: Display all authors of git-gui.
  git-gui: Change summary of git-gui.
  git-gui: Include browser in our usage message.
  git-gui: Remove TODO list.
  git-gui: Don't crash in citool mode on initial commit.
  Document the new core.bare configuration option.
  Include git-gui credits file in dist.

* The 'master' branch has these since the last announcement.

 Alex Riesen (1):
  disable t4016-diff-quote.sh on some filesystems

 Fredrik Kuivinen (2):
  New autoconf test for iconv
  Fix 'git commit -a' in a newly initialized repository

 Jason Riedy (1):
  Check for PRIuMAX rather than NO_C99_FORMAT in fast-import.c.

 Johannes Schindelin (5):
  apply: fix memory leak in prefix_one()
  name-rev: avoid "^0" when unneeded
  git grep: use pager
  Teach diff -B about colours
  git-diff: fix combined diff

 Junio C Hamano (16):
  t0020: add test for auto-crlf
  Teach 'git apply' to look at $GIT_DIR/config
  Teach core.autocrlf to 'git apply'
  Teach 'git apply' to look at $HOME/.gitconfig even outside of a repository
  git-apply: do not lose cwd when run from a subdirectory.
  git-apply: require -p<n> when working in a subdirectory.
  Link 1.5.0.1 documentation from the main page.
  Add prefixcmp()
  Mechanical conversion to use prefixcmp()
  prefixcmp(): fix-up mechanical conversion.
  prefixcmp(): fix-up leftover strncmp().
  t4119: add test for traditional patch and different p_value
  Fix botched "leak fix"
  git-apply: notice "diff --git" patch again
  git-apply: guess correct -p<n> value for non-git patches.
  t4119: test autocomputing -p<n> for traditional diff input.

 Linus Torvalds (2):
  Lazy man's auto-CRLF
  Make AutoCRLF ternary variable.

 Martin Koegler (1):
  git-gui: Create new branches from a tag.

 Martin Waitz (1):
  Support for large files on 32bit systems.

 Michael Loeffler (1):
  Use gunzip -c over gzcat in import-tars example.

 Pavel Roskin (1):
  git-remote: support remotes with a dot in the name

 Shawn O. Pearce (16):
  git-gui: Refactor 'exec git subcmd' idiom.
  git-gui: Basic version check to ensure git 1.5.0 or later is used.
  git-gui: Permit merging tags into the current branch.
  git-gui: More consistently display the application name.
  git-gui: Print version on the console.
  git-gui: Prefer version file over git-describe.
  git-gui: Expose the browser as a subcommand.
  git-gui: Correct crash when saving options in blame mode.
  git-gui: Use mixed path for docs on Cygwin.
  git-gui: Display all authors of git-gui.
  git-gui: Change summary of git-gui.
  git-gui: Include browser in our usage message.
  git-gui: Remove TODO list.
  git-gui: Don't crash in citool mode on initial commit.
  Document the new core.bare configuration option.
  Include git-gui credits file in dist.

 Simon 'corecode' Schubert (1):
  Allow passing of an alternative CVSROOT via -d.

 Theodore Ts'o (2):
  Add config_boolean() method to the Git perl module
  Allow git-remote to update named groups of remotes

I'll do another maintenance release 1.5.0.3 soonish to flush accumulated fixes on 'maint' branch.

On the 'master' front, we have

 * a largish git-svn updates.
 * a handful git diff improvements.
 * a beginning of CRLF mangling.
 * improved behaviour of git-apply run in subdirectories.
 * /etc/gitconfig

Soon we will also have these from 'next'. I think it will be time to do v1.5.1 after they are in:

 * "fake symlink" work for MinGW port
 * built-in revert/cherry-pick
 * "git diff" that compares stdin and /dev/null
 * make "git fetch" quieter when noninteractive.

----------------------------------------------------------------
* The 'maint' branch has these fixes since v1.5.0.2.

 Alexandre Julliard (1):
  git.el: Set the default commit coding system from the repository config.

 Aneesh Kumar K.V (1):
  blameview: Fix the browse behavior in blameview

 Christian Schlotter (1):
  Documentation: Correct minor typo in git-add documentation.

 Eygene Ryabinkin (2):
  http-push.c::lock_remote(): validate all remote refs.
  Another memory overrun in http-push.c

 Gerrit Pape (2):
  git-cvsexportcommit: don't cleanup .msg if not yet committed to cvs.
  Fix quoting in update hook template

 Jim Meyering (1):
  diff --cc: integer overflow given a 2GB-or-larger file

 Johannes Schindelin (3):
  fetch.o depends on the headers, too.
  builtin-archive: use RUN_SETUP
  Document the config variable format.suffix

 Junio C Hamano (4):
  git-apply: do not fix whitespaces on context lines.
  Documentation: git-remote add [-t <branch>] [-m <branch>] [-f] name url
  Start preparing Release Notes for 1.5.0.3
  git-merge: fail correctly when we cannot fast forward.

 Linus Torvalds (2):
  mailinfo: do not get confused with logical lines that are too long.
  git-show: Reject native ref

 Matthias Kestenholz (1):
  Fix git-gc usage note

 Michael Coleman (2):
  Fix minor typos/grammar in user-manual.txt
  builtin-fmt-merge-msg: fix bugs in --file option

 Michael Poole (1):
  Correct ordering in git-cvsimport's option documentation

 Paolo Bonzini (1):
  git-archimport: support empty summaries, put summary on a single line.

 Ramsay Allan Jones (5):
  Fix a "label defined but unreferenced" warning.
  Fix an "implicit function definition" warning.
  Fix some "comparison is always true/false" warnings.
  Fix a "pointer type missmatch" warning.
  Unset NO_C99_FORMAT on Cygwin.

 Sergey Vlasov (3):
  Documentation/build-docdep.perl: Fix dependencies for included asciidoc files
  Documentation/git-quiltimport.txt: Fix labeled list formatting
  Documentation/git-send-email.txt: Fix labeled list formatting

 Shawn O. Pearce (1):
  index-pack: Loop over pread until data loading is complete.

 Theodore Ts'o (1):
  Fix git-show man page formatting in the EXAMPLES section

 Uwe Kleine-König (1):
  Include config.mak in doc/Makefile

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

 Andy Parkins (3):
  cvsserver: Remove trailing "\n" from commithash in checkin function
  cvsserver: Make always-binary mode a config file option
  Sample update hook: typofix and modernization to use "git log"

 Aneesh Kumar K.V (1):
  Documentation: document remote.<name>.tagopt

 Eric Wong (118):
  git-svn: move authentication prompts into their own namespace
  git-svn: cleanup: move process_rm around
  git-svn: cleanup: put SVN workarounds into their own namespace
  git-svn: cleanup: avoid re-use()ing Git.pm in sub-packages
  git-svn: add Git::SVN module (to avoid global variables)
  git-svn: convert 'init' to use Git::SVN
  git-svn: convert multi-init over to using Git::SVN
  git-svn: make multi-init capable of reusing the Ra connection
  git-svn: add a test for show-ignore
  git-svn: convert show-ignore over to Git::SVN
  git-svn: moved the 'log' command into its own namespace
  git-svn: port the 'rebuild' command to use Git::SVN objects
  git-svn: do not let Git.pm warn if we prematurely close pipes
  git-svn: convert the 'commit-diff' command to Git::SVN
  git-svn: get rid of Memoize for now...
  git-svn: fetch/multi-fetch converted over to Git::SVN module
  git-svn: switch dcommit to using Git::SVN code
  git-svn: convert 'set-tree' command to use Git::SVN
  git-svn: remove graft-branches command
  git-svn: add support for metadata in .git/config
  git-svn: fix a regression in dcommit that caused empty log messages
  git-svn: reuse open SVN::Ra connections by URL
  git-svn: enable --minimize to simplify the config and connections
  git-svn: fix --follow-parent to work with Git::SVN
  git-svn: --follow-parent works with svn-remotes multiple branches
  git-svn: disallow ambigious local refspecs
  git-svn: allow --follow-parent on deleted directories
  git-svn: get rid of additional fetch-arguments
  git-svn: allow 'init' to work outside of tests
  git-svn: better error reporting if --follow-parent fails
  git-svn: 'init' attempts to connect to the repository root if possible
  git-svn: --follow-parent now works on sub-directories of larger branches
  git-svn: track writes writes to the index in fetch
  git-svn: add an odd test case that seems to cause segfaults over HTTP
  git-svn: avoid tracking change-less revisions
  git-svn: correctly track revisions made to deleted branches
  git-svn: fix segfaults from accessing svn_log_changed_path_t
  git-svn: fix committing to subdirectories, add tests
  git-svn: avoid an extra svn_ra connection during commits
  git-svn: simplify usage of the SVN::Git::Editor interface
  git-svn: cleanup remove unused function
  git-svn: allow multi-fetch to fetch things chronologically
  git-svn: correctly track diff-less copies with do_switch
  git-svn: correctly handle do_{switch,update} in deep directories
  git-svn: stop using path names as refnames with --follow-parent
  git-svn: cleanup: move editor-specific variables into the editor namespace
  git-svn: just use Digest::MD5 instead of requiring it
  git-svn: reinstate the default SVN error handler after using get_log
  git-svn: don't rely on do_switch + reparenting with svn(+ssh)://
  git-svn: fetch tracks initial change with --follow-parent
  git-svn: remove the 'rebuild' command and make the functionality automatic
  git-svn: fix several fetch bugs related to repeated invocations
  git-svn: reinstate --no-metadata, add --svn-remote=, variable cleanups
  git-svn: gracefully handle --follow-parent failures
  git-svn: make (multi-)fetch safer but slower
  git-svn: avoid a huge memory spike with high-numbered revisions
  git-svn: re-enable repacking flags
  git-svn: do our best to ensure that our ref and rev_db are consistent
  git-svn: avoid redundant get_log calls between invocations
  git-svn: use sys* IO functions for reading rev_db
  git-svn: don't write to the config file from --follow-parent
  git-svn: save paths to tags/branches with for future reuse
  git-svn: migrations default to [svn-remote "git-svn"]
  git-svn: get rid of revisions_eq check for --follow-parent
  git-svn: avoid extra get_log calls when refspecs are added for fetching
  git-svn: just name the default svn-remote "svn" instead of "git-svn"
  git-svn: prepare multi-init for wildcard support
  git-svn: reintroduce using a single get_log() to fetch
  git-svn: run get_log() on a sub-directory if possible
  git-svn: implement auto-discovery of branches/tags
  git-svn: --follow-parent tracks multi-parent paths
  git-svn: remove check_path calls before calling do_update
  git-svn: remove some noisy debugging messages
  git-svn: enable follow-parent functionality by default
  git-svn: fix buggy regular expression usage in several places
  git-svn: correctly handle the -q flag in SVN::Git::Fetcher
  git-svn: correctly handle globs with a right-hand-side path component
  git-svn: remove optimized commit stuff for set-tree
  git-svn: add support for SVN::Mirror/svk using revprops for metadata
  git-svn: add support for per-[svn-remote "..."] options
  git-svn: use private $GIT_DIR/svn/config file more
  git-svn: extra safety for noMetadata and useSvmProps users
  git-svn: use separate, per-repository .rev_db files
  git-svn: write the highest maxRex out for branches and tags
  git-svn: handle multi-init without --trunk, UseSvmProps fixes
  git-svn: make dcommit usable for glob users
  git-svn: include merges when calling rev-list for decommit
  git-svn: usability fixes for the 'git svn log' command
  t910*: s/repo-config/config/g; poke around possible race conditions
  git-svn: hopefully make 'fetch' more user-friendly
  git-svn: allow 'init' to act as multi-init
  git-svn: brown paper bag fixes
  git-svn: simplify the (multi-)init methods of fetching
  git-svn: allow --log-window-size to be specified, default to 100
  git-svn: remember to check for clean indices on globbed refs, too
  git-svn: error checking for invalid [svn-remote "..."] sections
  git-svn: allow dcommit for those who only fetch from SVM with useSvmProps
  git-svn: documentation updates for new functionality
  git-svn: add support for --stat in the log command
  git-svn: checkout files on new fetches
  git-svn: add a 'rebase' command
  git-svn: fix some issues for people migrating from older versions
  git-svn: hide the private git-svn 'config' file as '.metadata'
  git-svn: add 'clone' command, an alias for init + fetch
  git-svn: allow overriding of the SVN repo root in metadata
  git-svn: add support for using svnsync properties
  git-svn: fix useSvmProps, hopefully for the last time
  git-svn: add test for useSvnsyncProps
  git-svn: documentation updates
  git-svn: allow metadata options to be specified with 'init' and 'clone'
  git-svn: give show-ignore HEAD smarts, like dcommit and log
  git-svn: ensure we're at the top-level and can access $GIT_DIR
  git-svn: fix clone when a target directory has been specified
  git-svn: fix reconnections to different paths of svn:// repositories
  git-svn: fix some potential bugs with --follow-parent
  Add test-chmtime: a utility to change mtime on files
  Update tests to use test-chmtime
  git-svn: fix show-ignore when not connected to the repository root

 Johannes Schindelin (18):
  rev-list --max-age, --max-count: support --boundary
  config: read system-wide defaults from /etc/gitconfig
  apply: make --verbose a little more useful
  Teach git-diff-files the new option `--no-index`
  pretty-formats: add 'format:<string>'
  Make tests independent of global config files
  Add git-bundle: move objects and references by archive
  git-bundle: assorted fixes
  git-bundle: avoid fork() in verify_bundle()
  git-bundle: fix 'create --all'
  git-bundle: record commit summary in the prerequisite data
  object name: introduce ':/<oneline prefix>' notation
  Fix typo: do not show name1 when name2 fails
  diff --no-index: also imitate the exit status of diff(1)
  Actually make print_wrapped_text() useful
  show_date(): rename the "relative" parameter to "mode"
  diff: make more cases implicit --no-index
  print_wrapped_text: fix output for negative indent

 Julian Phillips (2):
  git-branch: improve abbreviation of sha1s in verbose mode
  git-branch: document new --no-abbrev option

 Junio C Hamano (13):
  git-status: do not be totally useless in a read-only repository.
  update-index: do not die too early in a read-only repository.
  run_diff_{files,index}(): update calling convention.
  .mailmap maintenance after pulling from git-svn
  bundle: reword missing prerequisite error message
  diff --cached: give more sensible error message when HEAD is yet to be created.
  Documentation: link in 1.5.0.2 material to the top documentation page.
  Make 'cvs ci' lockless in git-cvsserver by using git-update-ref
  index_fd(): use enum object_type instead of type name string.
  index_fd(): pass optional path parameter as hint for blob conversion
  index_fd(): convert blob only if it is a regular file.
  Add recent changes to draft 1.5.1 release notes.
  diff-ni: allow running from a subdirectory.

 Michael Coleman (2):
  git-send-email: abort/usage on bad option
  fix various doc typos

 Nicolas Pitre (7):
  sha1_file.c: cleanup hdr usage
  sha1_file.c: cleanup "offset" usage
  sha1_file.c: don't ignore an error condition in sha1_loose_object_info()
  formalize typename(), and add its reverse type_from_string()
  convert object type handling from a string to a number
  get rid of lookup_object_type()
  make sure enum object_type is signed

 Sam Vilain (3):
  git-svn: make test for SVK mirror path import
  git-svn: don't consider SVN URL usernames significant when comparing
  git-svn: document --username

 Sergey Vlasov (1):
  Documentation/git-svn.txt: Fix formatting errors

 Shawn O. Pearce (1):
  Cleanup check_valid in commit-tree.

Perhaps it is time for 1.5.0.4 this Wednesday (my time) from 'maint'.

On the 'master' front, Johannes and Matthias cleaned up the git-bundle and made it (hopefully) usable shape. Also git-diff got even nicer when used outside the context of git, again thanks to Johannes. I think we are nearing 1.5.1-rc1, which I'll talk about in "What's cooking".

----------------------------------------------------------------

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

 Alexandre Julliard (2):
  git.el: Avoid appending a signoff line that is already present.
  git.el: Retrieve commit log information from .dotest directory.

 Avi Kivity (1):
  git-send-email: Document configuration options

 Brian Gernhardt (1):
  Fix diff-options references in git-diff and git-format-patch

 J. Bruce Fields (13):
  Documentation: mention module option to git-cvsimport
  user-manual: reset to ORIG_HEAD not HEAD to undo merge
  user-manual: ensure generated manual references stylesheet
  user-manual: insert earlier of mention content-addressable architecture
  user-manual: how to replace commits older than most recent
  user-manual: more detailed merge discussion
  glossary: fix overoptimistic automatic linking of defined terms
  user-manual: fix inconsistent example
  user-manual: fix inconsistent use of pull and merge
  user-manual: fix missing colon in git-show example
  user-manual: fix rendering of history diagrams
  user-manual: install user manual stylesheet with other web documents
  git-merge: warn when -m provided on a fast forward

 Jeff King (2):
  Documentation: s/seperator/separator/
  fast-import: grow tree storage more aggressively

 Johannes Schindelin (2):
  Begin SubmittingPatches with a check list
  make t8001 work on Mac OS X again

 Junio C Hamano (2):
  GIT 1.5.0.3
  git-commit: cd to top before showing the final stat

 Matthias Kestenholz (1):
  Adjust reflog filemode in shared repository

 Matthias Lederhofer (1):
  setup_git_directory_gently: fix off-by-one error

 Shawn O. Pearce (13):
  git-gui: Relocate the menu/transport menu code.
  git-gui: Add Reset to the Branch menu.
  git-gui: Don't create empty (same tree as parent) commits.
  git-gui: Remove unnecessary /dev/null redirection.
  fast-import: Avoid infinite loop after reset
  fast-import: Fail if a non-existant commit is used for merge
  git-gui: Make 'make' quieter by default
  Catch write_ref_sha1 failure in receive-pack
  git-gui: Allow committing empty merges
  git-gui: Revert "Don't modify CREDITS-FILE if it hasn't changed."
  git-gui: Revert "git-gui: Display all authors of git-gui."
  git-gui: Allow 'git gui version' outside of a repository
  Don't package the git-gui credits file anymore

 Theodore Ts'o (1):
  Add definition of <commit-ish> to the main git man page.

 Yasushi SHOJI (1):
  glossary: Add definitions for dangling and unreachable objects

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

 Alex Riesen (4):
  git-gui: Support of "make -s" in: do not output anything of the build itself
  More build output cleaning up
  Support of "make -s": do not output anything of the build itself
  Allow "make -w" generate its usual output

 Avi Kivity (1):
  git-send-email: configurable bcc and chain-reply-to

 Frank Lichtenheld (1):
  cvsserver: Use Merged response instead of Update-existing for merged files

 Jakub Narebski (1):
  gitweb: Don't escape attributes in CGI.pm HTML methods

 Jeff King (1):
  New fast-import test case for valid tree sorting

 Jim Meyering (1):
  I like the idea of the new ':/<oneline prefix>' notation, and gave it

 Johannes Schindelin (15):
  fetch & clone: do not output progress when not on a tty
  Fixup no-progress for fetch & clone
  Make git-revert & git-cherry-pick a builtin
  diff: support reading a file from stdin via "-"
  diff --no-index: support /dev/null as filename
  Get rid of the dependency to GNU diff in the tests
  cherry-pick: Suggest a better method to retain authorship
  format-patch: add --inline option and make --attach a true attachment
  bundle: fix wrong check of read_header()'s return value & add tests
  git-bundle: avoid packing objects which are in the prerequisites
  git-bundle: Make thin packs
  git-bundle: handle thin packs in subcommand "unbundle"
  git-bundle: die if a given ref is not included in bundle
  git-bundle: prevent overwriting existing bundles
  git-bundle: only die if pack would be empty, warn if ref is skipped

 Johannes Sixt (3):
  Add core.symlinks to mark filesystems that do not support symbolic links.
  Handle core.symlinks=false case in merge-recursive.
  Tell multi-parent diff about core.symlinks.

 Junio C Hamano (14):
  diff-ni: fix the diff with standard input
  format-patch --attach: not folding some long headers.
  Post 1.5.0.3 cleanup
  fsck: fix broken loose object check.
  unpack_sha1_file(): detect corrupt loose object files.
  fsck: exit with non-zero status upon errors
  git-bundle: fix pack generation.
  revision walker: Fix --boundary when limited
  revision traversal: retire BOUNDARY_SHOW
  git-bundle: various fixups
  revision traversal: SHOWN means shown
  git-bundle: make verify a bit more chatty.
  revision --boundary: fix stupid typo
  revision --boundary: fix uncounted case.

 Li Yang (1):
  gitweb: Change to use explicitly function call cgi->escapHTML()

 Linus Torvalds (1):
  Re-fix get_sha1_oneline()

 Paolo Bonzini (3):
  git-config: document --rename-section, provide --remove-section
  git-archimport: allow remapping branch names
  git-commit: add a --interactive option

 Santi Béjar (2):
  t/t5515-fetch-merge-logic.sh: Added tests for the merge login in git-fetch
  t/t5515-fetch-merge-logic.sh: Add two more tests

 Shawn O. Pearce (31):
  cherry-pick: Bug fix 'cherry picked from' message.
  Make 'make' quieter while building git
  Make 'make' quiet by default
  Display the null SHA-1 as the base for an OBJ_OFS_DELTA.
  Fix mmap leak caused by reading bad indexes.
  Don't build external_grep if its not used
  General const correctness fixes
  Use uint32_t for all packed object counts.
  Use uint32_t for pack-objects counters.
  Use off_t when we really mean a file offset.
  Use off_t in pack-objects/fast-import when we mean an offset
  Cast 64 bit off_t to 32 bit size_t
  Preallocate memory earlier in fast-import
  Move post-update hook to after all other activity
  Don't run post-update hook unless a ref changed
  Refactor run_update_hook to be more useful
  Refactor handling of error_string in receive-pack
  Teach receive-pack to run pre-receive/post-receive hooks
  Use atomic updates to the fast-import mark file
  Allow fast-import frontends to reload the marks table
  Switch to run_command_v_opt in revert
  Remove unused run_command variants
  Start defining a more sophisticated run_command
  Split run_command into two halves (start/finish)
  Teach run_command how to setup a stdin pipe
  Refactor run_command error handling in receive-pack
  Split back out update_hook handling in receive-pack
  Change {pre,post}-receive hooks to use stdin
  Remove unnecessary casts from fast-import
  Simplify closing two fds at once in run-command.c
  Fix t5510-fetch's use of sed

 Xavier Maillard (1):
  contrib/emacs: Use non-interactive function to byte-compile files

Junio C Hamano <junkio@cox.net> writes:

Sorry, but I meant Mark Levedahl here.

Will you be applying your patch for the "checkout fails if something is a file in one branch and a directory in another" problem? I realize it's not a complete solution, but it's at least a significant improvement over the current behavior, and I don't imagine it breaks anything else.

I can resend my test case if you like.

-Steve

Junio C Hamano wrote:

Steven Grimm <koreth@midwinter.com> writes:

If I understand correctly, the patch is already broken. Rather, the patch itself may be correct but it exposes more breakages that currently is hidden only because we do not replace bunch of files in a subdirectory with a single file in the index thanks to the "too strict" check that is bothering you. Lifting that check with the previous patch would make it either fail to notice other local changes (which is worse), or mistakenly say there is a local change when there is none (which is not better but not worse). In particular, I think it is highly suspicious what the current code with the patch does to paths that immediately follow the directory to be removed, although I stopped looking into this issue, as we already said this will not be dealt with in 1.5.1.

In short, I do not think the patch is suitable for production, as the cure seems worse than the disease.

That does not mean anybody is forbidden to look into a fix in the meantime. I am just saying I do not expect it to happen before 1.5.1.

On the 'maint' front, there are a handful gitweb fixes, but there is nothing major. I'll perhaps do a v1.5.0.6 if I have time when I do the v1.5.1-rc2.

I allowed a handful new features to the 'master' side, contrary to the promise of -rc1 which is supposed to be 'no new features, only fixes'. They should be safe changes, though.

 - 'git-diff --quiet' optimizes tree-level comparison by exiting
   early when it finds one difference.  This speeds up path-limited
   form of git-rev-list and git-log significantly.

 - 'git-bisect run <script>' can be used to automate simple
   bisection. 

Notable fixes since -rc1 includes:

 - 'git-clone -q' supresses the "walk" and "got" noises while
   fetching over http.

 - 'git-revert' does not add duplicated abbreviated commit
   object name on the title line anymore.

 - With core.legacyheaders turned off, zero-length loose object
   was incorrectly judged corrupt.

 - 'git-apply' with core.autocrlf generated bogus results.

 - path-limited 'git-bisect' terminated too early before finding
   the guilty commit.

 - the sample update hook does not generate change notification
   e-mails anymore. 

----------------------------------------------------------------
* The 'maint' branch has these fixes since v1.5.0.5.

 Jakub Narebski (4):
  gitweb: Fix "next" link in commit view
  gitweb: Don't escape attributes in CGI.pm HTML methods
  gitweb: Fix not marking signoff lines in "log" view
  gitweb: Add some installation notes in gitweb/INSTALL

 Li Yang (1):
  gitweb: Change to use explicitly function call cgi->escapHTML()

 Michael S. Tsirkin (1):
  fix typo in git-am manpage

 Peter Eriksen (1):
  Documentation/pack-format.txt: Clear up description of types.

* The 'master' branch has these since v1.5.1-rc1.

 Alex Riesen (2):
  Document --quiet option to git-diff
  Use diff* with --exit-code in git-am, git-rebase and git-merge-ours

 Andy Parkins (2):
  update-hook: abort early if the project description is unset
  update-hook: remove e-mail sending hook.

 Brandon Casey (1):
  prefer "git COMMAND" over "git-COMMAND" in gitk

 Chris Wright (1):
  make git clone -q suppress the noise with http fetch

 Christian Couder (6):
  Bisect: implement "git bisect run <cmd>..." to automatically bisect.
  Documentation: bisect: reformat some paragraphs.
  Documentation: bisect: reword one paragraph.
  Documentation: bisect: reformat more paragraphs.
  Documentation: bisect: add some titles to some paragraphs.
  Documentation: bisect: make a comment fit better in the man page.

 Eric Wong (1):
  gitk: bind <F5> key to Update (reread commits)

 James Bowes (1):
  Replace remaining instances of strdup with xstrdup.

 Johannes Schindelin (5):
  xdiff/xutils.c(xdl_hash_record): factor out whitespace handling
  Add a HOWTO for setting up a standalone git daemon
  Provide some technical documentation for shallow clones
  t4118: be nice to non-GNU sed
  git-revert: Revert revert message to old behaviour

 Junio C Hamano (12):
  blame: micro-optimize cmp_suspect()
  blame: cmp_suspect is not "cmp" anymore.
  Teach tree_entry_interesting() that the tree entries are sorted.
  tree-diff: avoid strncmp()
  tree_entry_interesting(): allow it to say "everything is interesting"
  git-rebase: make 'rebase HEAD branch' work as expected.
  git-apply: Do not free the wrong buffer when we convert the data for writeout
  checkout: report where the new HEAD is upon detaching HEAD
  git-bisect: typofix
  git-bisect.sh: properly dq $GIT_DIR
  Fix path-limited "rev-list --bisect" termination condition.
  git-am documentation: describe what is taken from where.

 Linus Torvalds (6):
  Fix loose object uncompression check.
  Don't ever return corrupt objects from "parse_object()"
  Be more careful about zlib return values
  Remove "pathlen" from "struct name_entry"
  Initialize tree descriptors with a helper function rather than by hand.
  Switch over tree descriptors to contain a pre-parsed entry

 Michael S. Tsirkin (1):
  git-merge: Put FETCH_HEAD data in merge commit message

 Nicolas Pitre (10):
  clean up the delta base cache size a bit
  use a LRU eviction policy for the delta base cache
  don't ever allow SHA1 collisions to exist by fetching a pack
  index-pack: use hash_sha1_file()
  index-pack: more validation checks and cleanups
  improve checkout message when asking for same branch
  minor git-prune optimization
  update HEAD reflog when branch pointed to by HEAD is directly modified
  make it more obvious that temporary files are temporary files
  write_sha1_from_fd() should make new objects read-only

 Santi Béjar (1):
  git-fetch: Fix single_force in append_fetch_head

 Shawn O. Pearce (1):
  contrib/continuous: a continuous integration build manager

 Uwe Kleine-König (1):
  Bisect: convert revs given to good and bad to commits