Git v2.0.0

The latest feature release Git v2.0.0 is now available at the
usual places.

We had to delay the final release by a week or so because we found a
few problems in earlier release candidates (request-pull had a
regression that stopped it from showing the “tags/” prefix in
“Please pull tags/frotz” when the user asked to compose a request
for frotz to be pulled; a code path in git-gui to support ancient
versions of Git incorrectly triggered for Git 2.0), which we had to
fix in an extra unplanned release candidate.

Hopefully the next cycle will become shorter, as topics that have
been cooking on the next branch had extra time to mature, so it
all evens out in the end ;-).

The tarballs are found at:

The following public repositories all have a copy of the v2.0.0
tag and the master branch that the tag points at:

url =
url = git://
url =
url = git://
url = git://
url =

Git v2.0 Release Notes

Backward compatibility notes

When “git push [$there]” does not say what to push, we have used the
traditional “matching” semantics so far (all your branches were sent
to the remote as long as there already are branches of the same name
over there). In Git 2.0, the default is now the “simple” semantics,
which pushes:

  • only the current branch to the branch with the same name, and only
    when the current branch is set to integrate with that remote
    branch, if you are pushing to the same remote as you fetch from; or

  • only the current branch to the branch with the same name, if you
    are pushing to a remote that is not where you usually fetch from.

You can use the configuration variable “push.default” to change
this. If you are an old-timer who wants to keep using the
“matching” semantics, you can set the variable to “matching”, for
example. Read the documentation for other possibilities.

When “git add -u” and “git add -A” are run inside a subdirectory
without specifying which paths to add on the command line, they
operate on the entire tree for consistency with “git commit -a” and
other commands (these commands used to operate only on the current
subdirectory). Say “git add -u .” or “git add -A .” if you want to
limit the operation to the current directory.

“git add <path>” is the same as “git add -A <path>” now, so that
“git add dir/” will notice paths you removed from the directory and
record the removal. In older versions of Git, “git add <path>” used
to ignore removals. You can say “git add –ignore-removal <path>” to
add only added or modified paths in <path>, if you really want to.

The “-q” option to “git diff-files”, which does NOT mean “quiet”,
has been removed (it told Git to ignore deletion, which you can do
with “git diff-files –diff-filter=d”).

“git request-pull” lost a few “heuristics” that often led to mistakes.

The default prefix for “git svn” has changed in Git 2.0. For a long
time, “git svn” created its remote-tracking branches directly under
refs/remotes, but it now places them under refs/remotes/origin/ unless
it is told otherwise with its “–prefix” option.

Updates since v1.9 series

UI, Workflows & Features

  • The “multi-mail” post-receive hook (in contrib/) has been updated
    to a more recent version from upstream.

  • The “remote-hg/bzr” remote-helper interfaces (used to be in
    contrib/) are no more. They are now maintained separately as
    third-party plug-ins in their own repositories.

  • “git gc –aggressive” learned “–depth” option and
    “gc.aggressiveDepth” configuration variable to allow use of a less
    insane depth than the built-in default value of 250.

  • “git log” learned the “–show-linear-break” option to show where a
    single strand-of-pearls is broken in its output.

  • The “rev-parse –parseopt” mechanism used by scripted Porcelains to
    parse command-line options and to give help text learned to take
    the argv-help (the placeholder string for an option parameter,
    e.g. “key-id” in “–gpg-sign=<key-id>”).

  • The pattern to find where the function begins in C/C used in
    "diff" and "grep -p" has been updated to improve viewing C


  • “git rebase” learned to interpret a lone “-” as “@{-1}”, the
    branch that we were previously on.

  • “git commit –cleanup=<mode>” learned a new mode, scissors.

  • “git tag –list” output can be sorted using “version sort” with

  • Discard the accumulated “heuristics” to guess from which branch the
    result wants to be pulled from and make sure that what the end user
    specified is not second-guessed by “git request-pull”, to avoid
    mistakes. When you pushed out your master branch to your public
    repository as for-linus, use the new “master:for-linus” syntax to
    denote the branch to be pulled.

  • “git grep” learned to behave in a way similar to native grep when
    “-h” (no header) and “-c” (count) options are given.

  • “git push” via transport-helper interface has been updated to
    allow forced ref updates in a way similar to the natively
    supported transports.

  • The “simple” mode is the default for “git push”.

  • “git add -u” and “git add -A”, when run without any pathspec, is a
    tree-wide operation even when run inside a subdirectory of a
    working tree.

  • “git add <path>” is the same as “git add -A <path>” now.

  • “core.statinfo” configuration variable, which is a
    never-advertised synonym to “core.checkstat”, has been removed.

  • The “-q” option to “git diff-files”, which does NOT mean
    “quiet”, has been removed (it told Git to ignore deletion, which
    you can do with “git diff-files –diff-filter=d”).

  • Server operators can loosen the “tips of refs only” restriction for
    the remote archive service with the uploadarchive.allowUnreachable
    configuration option.

  • The progress indicators from various time-consuming commands have
    been marked for i18n/l10n.

  • “git notes -C <blob>” diagnoses as an error an attempt to use an
    object that is not a blob.

  • “git config” learned to read from the standard input when “-” is
    given as the value to its “–file” parameter (attempting an
    operation to update the configuration in the standard input is
    rejected, of course).

  • Trailing whitespaces in .gitignore files, unless they are quoted
    for fnmatch(3), e.g. “path\ “, are warned and ignored. Strictly
    speaking, this is a backward-incompatible change, but very unlikely
    to bite any sane user and adjusting should be obvious and easy.

  • Many commands that create commits, e.g. “pull” and “rebase”,
    learned to take the “–gpg-sign” option on the command line.

  • “git commit” can be told to always GPG sign the resulting commit
    by setting the “commit.gpgsign” configuration variable to “true”
    (the command-line option “–no-gpg-sign” should override it).

  • “git pull” can be told to only accept fast-forward by setting the
    new “pull.ff” configuration variable.

  • “git reset” learned the “-N” option, which does not reset the index
    fully for paths the index knows about but the tree-ish the command
    resets to does not (these paths are kept as intend-to-add entries).

Performance, Internal Implementation, etc.

  • The compilation options to port to AIX and to MSVC have been

  • We started using wildmatch() in place of fnmatch(3) a few releases
    ago; complete the process and stop using fnmatch(3).

  • Uses of curl’s “multi” interface and “easy” interface do not mix
    well when we attempt to reuse outgoing connections. Teach the RPC
    over HTTP code, used in the smart HTTP transport, not to use the
    “easy” interface.

  • The bitmap-index feature from JGit has been ported, which should
    significantly improve performance when serving objects from a
    repository that uses it.

  • The way “git log –cc” shows a combined diff against multiple
    parents has been optimized.

  • The prefixcmp() and suffixcmp() functions are gone. Use
    starts_with() and ends_with(), and also consider if skip_prefix()
    suits your needs better when using the former.

Also contains various documentation updates and code clean-ups. Many
of them came from flurry of activities as GSoC candidate microproject

Fixes since v1.9 series

Unless otherwise noted, all the fixes since v1.9 in the maintenance
track are contained in this release (see the maintenance releases’
notes for details).

  • “git p4” was broken in 1.9 release to deal with changes in binary
    (merge 749b668 cl/p4-use-diff-tree later to maint).

  • The shell prompt script (in contrib/), when using the PROMPT_COMMAND
    interface, used an unsafe construct when showing the branch name in
    (merge 1e4119c8 rh/prompt-pcmode-avoid-eval-on-refname later to maint).

  • “git rebase” used a POSIX shell construct FreeBSD’s /bin/sh does not
    work well with.
    (merge 8cd6596 km/avoid-non-function-return-in-rebase later to maint).

  • zsh prompt (in contrib/) leaked unnecessary error messages.

  • Bash completion (in contrib/) did not complete the refs and remotes
    correctly given “git pu<TAB>” when “pu” is aliased to “push”.

  • Some more Unicode code points, defined in Unicode 6.3 as having zero
    width, have been taught to our display column counting logic.
    (merge d813ab9 tb/unicode-6.3-zero-width later to maint).

  • Some tests used shell constructs that did not work well on FreeBSD
    (merge ff7a1c6 km/avoid-bs-in-shell-glob later to maint).
    (merge 00764ca km/avoid-cp-a later to maint).

  • “git update-ref –stdin” did not fail a request to create a ref
    when the ref already existed.
    (merge b9d56b5 mh/update-ref-batch-create-fix later to maint).

  • “git diff –no-index -Mq a b” fell into an infinite loop.
    (merge ad1c3fb jc/fix-diff-no-index-diff-opt-parse later to maint).

  • “git fetch –prune”, when the right-hand side of multiple fetch
    refspecs overlap (e.g. storing “refs/heads/” to
    “, while storing “refs/frotz/” to
    “), aggressively thought that lack of
    “refs/heads/fr/otz” on the origin site meant we should remove
    “refs/remotes/origin/fr/otz” from us, without checking their
    “refs/frotz/otz” first.

    Note that such a configuration is inherently unsafe (think what
    should happen when “refs/heads/fr/otz” does appear on the origin
    site), but that is not a reason not to be extra careful.
    (merge e6f6371 cn/fetch-prune-overlapping-destination later to maint).

  • “git status –porcelain –branch” showed its output with labels
    “ahead/behind/gone” translated to the user’s locale.
    (merge 7a76c28 mm/status-porcelain-format-i18n-fix later to maint).

  • A stray environment variable $prefix could have leaked into and
    affected the behaviour of the “subtree” script (in contrib/).

  • When it is not necessary to edit a commit log message (e.g. “git
    commit -m” is given a message without specifying “-e”), we used to
    disable the spawning of the editor by overriding GIT_EDITOR, but
    this means all the uses of the editor, other than to edit the
    commit log message, are also affected.
    (merge b549be0 bp/commit-p-editor later to maint).

  • “git mv” that moves a submodule forgot to adjust the array that
    uses to keep track of which submodules were to be moved to update
    its configuration.
    (merge fb8a4e8 jk/mv-submodules-fix later to maint).

  • Length limit for the pathname used when removing a path in a deep
    subdirectory has been removed to avoid buffer overflows.
    (merge 2f29e0c mh/remove-subtree-long-pathname-fix later to maint).

  • The test helper lib-terminal always run an actual test_expect_*
    when included, which screwed up with the use of skil-all that may
    have to be done later.
    (merge 7e27173 jk/lib-terminal-lazy later to maint).

  • “git index-pack” used a wrong variable to name the keep-file in an
    error message when the file cannot be written or closed.
    (merge de983a0 nd/index-pack-error-message later to maint).

  • “rebase -i” produced a broken insn sheet when the title of a commit
    happened to contain \n (or ended with \c) due to a careless use
    of echo.
    (merge cb1aefd us/printf-not-echo later to maint).

  • There were a few instances of git-foo remaining in the
    documentation that should have been spelled git foo.
    (merge 3c3e6f5 rr/doc-merge-strategies later to maint).

  • Serving objects from a shallow repository needs to write a
    new file to hold the temporary shallow boundaries, but it was not
    cleaned when we exit due to die() or a signal.
    (merge 7839632 jk/shallow-update-fix later to maint).

  • When “git stash pop” stops after failing to apply the stash
    (e.g. due to conflicting changes), the stash is not dropped. State
    that explicitly in the output to let the users know.
    (merge 2d4c993 jc/stash-pop-not-popped later to maint).

  • The labels in “git status” output that describe the nature of
    conflicts (e.g. “both deleted”) were limited to 20 bytes, which was
    too short for some l10n (e.g. fr).
    (merge c7cb333 jn/wt-status later to maint).

  • “git clean -d pathspec” did not use the given pathspec correctly
    and ended up cleaning too much.
    (merge 1f2e108 jk/clean-d-pathspec later to maint).

  • “git difftool” misbehaved when the repository is bound to the
    working tree with the “.git file” mechanism, where a textual file
    “.git” tells us where it is.
    (merge fcfec8b da/difftool-git-files later to maint).

  • “git push” did not pay attention to “branch.*.pushremote” if it is
    defined earlier than “remote.pushdefault”; the order of these two
    variables in the configuration file should not matter, but it did
    by mistake.
    (merge 98b406f jk/remote-pushremote-config-reading later to maint).

  • Code paths that parse timestamps in commit objects have been
    (merge f80d1f9 jk/commit-dates-parsing-fix later to maint).

  • “git diff –external-diff” incorrectly fed the submodule directory
    in the working tree to the external diff driver when it knew that it
    is the same as one of the versions being compared.
    (merge aba4727 tr/diff-submodule-no-reuse-worktree later to maint).

  • “git reset” needs to refresh the index when working in a working
    tree (it can also be used to match the index to the HEAD in an
    otherwise bare repository), but it failed to set up the working
    tree properly, causing GIT_WORK_TREE to be ignored.
    (merge b7756d4 nd/reset-setup-worktree later to maint).

  • “git check-attr” when working on a repository with a working tree
    did not work well when the working tree was specified via the
    “–work-tree” (and obviously with “–git-dir”) option.
    (merge cdbf623 jc/check-attr-honor-working-tree later to maint).

  • “merge-recursive” was broken in 1.7.7 era and stopped working in
    an empty (temporary) working tree, when there are renames
    involved. This has been corrected.
    (merge 6e2068a bk/refresh-missing-ok-in-merge-recursive later to maint.)

  • “git rev-parse” was loose in rejecting command-line arguments
    that do not make sense, e.g. “–default” without the required
    value for that option.
    (merge a43219f ds/rev-parse-required-args later to maint.)

  • “include.path” variable (or any variable that expects a path that
    can use \~username expansion) in the configuration file is not a
    boolean, but the code failed to check it.
    (merge 67beb60 jk/config-path-include-fix later to maint.)

  • Commands that take pathspecs on the command line misbehaved when
    the pathspec is given as an absolute pathname (which is a
    practice not particularly encouraged) that points at a symbolic
    link in the working tree.
    (merge 6127ff6 mw/symlinks later to maint.)

  • “git diff –quiet — pathspec1 pathspec2” sometimes did not return
    the correct status value.
    (merge f34b205 nd/diff-quiet-stat-dirty later to maint.)

  • Attempting to deepen a shallow repository by fetching over smart
    HTTP transport failed in the protocol exchange, when the no-done
    extension was used. The fetching side waited for the list of
    shallow boundary commits after the sending side stopped talking to
    (merge 0232852 nd/http-fetch-shallow-fix later to maint.)

  • Allow “git cmd path/”, when the path is where a submodule is
    bound to the top-level working tree, to match path, despite the
    extra and unnecessary trailing slash (such a slash is often
    given by command-line completion).
    (merge 2e70c01 nd/submodule-pathspec-ending-with-slash later to maint.)

  • Documentation and in-code comments had many instances of mistaken
    use of “nor”, which have been corrected.
    (merge 235e8d5 jl/nor-or-nand-and later to maint).

Junio C Hamano wrote on 28 May 2014