---------------------
* When "git am" is fed an input that has multiple "Content-type: ..."
header, it did not grok charset= attribute correctly.
* Even during a conflicted merge, "git blame $path" always meant to
blame uncommitted changes to the "working tree" version; make it
more useful by showing cleanly merged parts as coming from the other
branch that is being merged.
* "git blame MAKEFILE" run in a history that has "Makefile" but not
"MAKEFILE" should say "No such file MAKEFILE in HEAD", but got
confused on a case insensitive filesystem and failed to do so.
* "git fetch --all", when passed "--no-tags", did not honor the
"--no-tags" option while fetching from individual remotes (the same
issue existed with "--tags", but combination "--all --tags" makes
much less sense than "--all --no-tags").
* "git log/diff/format-patch --stat" showed the "N line(s) added"
comment in user's locale and caused careless submitters to send
patches with such a line in them to projects whose project language
is not their language, mildly irritating others. Localization to
the line has been disabled for now.
* "git log --all-match --grep=A --grep=B" ought to show commits that
mention both A and B, but when these three options are used with
--author or --committer, it showed commits that mention either A or
B (or both) instead.
* The subcommand to remove the definition of a remote in "git remote"
was named "rm" even though all other subcommands were spelled out.
Introduce "git remote remove" to remove confusion, and keep "rm" as
a backward compatible synonym.
Also contains a handful of documentation updates.
UI, Workflows & Features
* Git can be told to normalize pathnames it read from readdir(3) and
all arguments it got from the command line into precomposed UTF-8
(assuming that they come as decomposed UTF-8), in order to work
around issues on Mac OS.
I think there still are other places that need conversion
(e.g. paths that are read from stdin for some commands), but this
should be a good first step in the right direction.
* Per-user $HOME/.gitconfig file can optionally be stored in
$HOME/.config/git/config instead, which is in line with XDG.
* The value of core.attributesfile and core.excludesfile default to
$HOME/.config/git/attributes and $HOME/.config/git/ignore respectively
when these files exist.
* Logic to disambiguate abbreviated object names have been taught to
take advantage of object types that are expected in the context,
e.g. XXXXXX in the "git describe" output v1.2.3-gXXXXXX must be a
commit object, not a blob nor a tree. This will help us prolong
the lifetime of abbreviated object names.
* "git apply" learned to wiggle the base version and perform three-way
merge when a patch does not exactly apply to the version you have.
* Scripted Porcelain writers now have access to the credential API via
the "git credential" plumbing command.
* "git help" used to always default to "man" format even on platforms
where "man" viewer is not widely available.
* "git clone --local $path" started its life as an experiment to
optionally use link/copy when cloning a repository on the disk, but
we didn't deprecate it after we made the option a no-op to always
use the optimization. The command learned "--no-local" option to
turn this off, as a more explicit alternative over use of file://
URL.
* "git fetch" and friends used to say "remote side hung up
unexpectedly" when they failed to get response they expect from the
other side, but one common reason why they don't get expected
response is that the remote repository does not exist or cannot be
read. The error message in this case was updated to give better
hints to the user.
* "git help -w $cmd" can show HTML version of documentation for
"git-$cmd" by setting help.htmlpath to somewhere other than the
default location where the build procedure installs them locally;
the variable can even point at a http:// URL.
* "git rebase [-i] --root $tip" can now be used to rewrite all the
history leading to "$tip" down to the root commit.
* "git rebase -i" learned "-x <cmd>" to insert "exec <cmd>" after
each commit in the resulting history.
* "git status" gives finer classification to various states of paths
in conflicted state and offer advice messages in its output.
* "git submodule" learned to deal with nested submodule structure
where a module is contained within a module whose origin is
specified as a relative URL to its superproject's origin.
* A rather heavy-ish "git completion" script has been split to create
a separate "git prompting" script, to help lazy-autoloading of the
completion part while making prompting part always available.
* "gitweb" pays attention to various forms of credits that are
similar to "Signed-off-by:" lines in the commit objects and
highlights them accordingly.
* The error message from "git push $there :bogo" (and its equivalent
"git push $there --delete bogo") mentioned that we tried and failed
to guess what ref is being deleted based on the LHS of the refspec,
which we don't.
* A handful of files and directories we create had tighter than
necessary permission bits when the user wanted to have group
writability (e.g. by setting "umask 002").
* "commit --amend" used to refuse amending a commit with an empty log
message, with or without "--allow-empty-message".
* "git commit --amend --only --" was meant to allow "Clever" people to
rewrite the commit message without making any change even when they
have already changes for the next commit added to their index, but
it never worked as advertised since it was introduced in 1.3.0 era.
* Even though the index can record pathnames longer than 1<<12 bytes,
in some places we were not comparing them in full, potentially
replacing index entries instead of adding.
* "git show"'s auto-walking behaviour was an unreliable and
unpredictable hack; it now behaves just like "git log" does when it
walks.
* "git diff", "git status" and anything that internally uses the
comparison machinery was utterly broken when the difference
involved a file with "-" as its name. This was due to the way "git
diff --no-index" was incorrectly bolted on to the system, making
any comparison that involves a file "-" at the root level
incorrectly read from the standard input.
* We did not have test to make sure "git rebase" without extra options
filters out an empty commit in the original history.
* "git fast-export" produced an input stream for fast-import without
properly quoting pathnames when they contain SPs in them.
* "git checkout --detach", when you are still on an unborn branch,
should be forbidden, but it wasn't.
* Some implementations of Perl terminates "lines" with CRLF even when
the script is operating on just a sequence of bytes. Make sure to
use "$PERL_PATH", the version of Perl the user told Git to use, in
our tests to avoid unnecessary breakages in tests.
Also contains minor typofixes and documentation updates.
* On Cygwin, the platform pread(2) is not thread safe, just like our
own compat/ emulation, and cannot be used in the index-pack
program. Makefile variable NO_THREAD_SAFE_PREAD can be defined to
avoid use of this function in a threaded program.
* "git add" allows adding a regular file to the path where a
submodule used to exist, but "git update-index" does not allow an
equivalent operation to Porcelain writers.
* "git archive" incorrectly computed the header checksum; the symptom
was observed only when using pathnames with hi-bit set.
* "git blame" did not try to make sure that the abbreviated commit
object names in its output are unique.
* Running "git bundle verify" on a bundle that records a complete
history said "it requires these 0 commits".
* "git clone --single-branch" to clone a single branch did not limit
the cloning to the specified branch.
* "git diff --no-index" did not correctly handle relative paths and
did not correctly give exit codes when run under "--quiet" option.
* "git diff --no-index" did not work with pagers correctly.
* "git diff COPYING HEAD:COPYING" gave a nonsense error message that
claimed that the treeish HEAD did not have COPYING in it.
* When "git log" gets "--simplify-merges/by-decoration" together with
"--first-parent", the combination of these options makes the
simplification logic to use in-core commit objects that haven't
been examined for relevance, either producing incorrect result or
taking too long to produce any output. Teach the simplification
logic to ignore commits that the first-parent traversal logic
ignored when both are in effect to work around the issue.
* "git ls-files --exclude=t -i" did not consider anything under t/ as
excluded, as it did not pay attention to exclusion of leading paths
while walking the index. Other two users of excluded() are also
updated.
* "git request-pull $url dev" when the tip of "dev" branch was tagged
with "ext4-for-linus" used the contents from the tag in the output
but still asked the "dev" branch to be pulled, not the tag.
Also contains minor typofixes and documentation updates.
* The cross links in the HTML version of manual pages were broken.
Also contains minor typofixes and documentation updates.
Changes 1.7.11:
UI, Workflows & Features
* A new mode for push, "simple", which is a cross between "current"
and "upstream", has been introduced. "git push" without any refspec
will push the current branch out to the same name at the remote
repository only when it is set to track the branch with the same
name over there. The plan is to make this mode the new default
value when push.default is not configured.
* A couple of commands learned the "--column" option to produce
columnar output.
* A third-party tool "git subtree" is distributed in contrib/
* A remote helper that acts as a proxy and caches ssl session for the
https:// transport is added to the contrib/ area.
* Error messages given when @{u} is used for a branch without its
upstream configured have been clarified.
* Even with the "-q"uiet option, "checkout" used to report setting up
tracking. Also "branch" learned the "-q"uiet option to squelch
informational message.
* Your build platform may support hardlinks but you may prefer not to
use them, e.g. when installing to DESTDIR to make a tarball and
untarring on a filesystem that has poor support for hardlinks.
There is a Makefile option NO_INSTALL_HARDLINKS for you.
* The smart-http backend used to always override GIT_COMMITTER_*
variables with REMOTE_USER and REMOTE_ADDR, but these variables are
now preserved when set.
* "git am" learned the "--include" option, which is an opposite of
existing the "--exclude" option.
* When "git am -3" needs to fall back to an application of the patch
to a synthesized preimage followed by a 3-way merge, the paths that
needed such treatment are now reported to the end user, so that the
result in them can be eyeballed with extra care.
* The output from "diff/log --stat" used to always allocate 4 columns
to show the number of modified lines, but not anymore.
* "git difftool" learned the "--dir-diff" option to spawn external
diff tools that can compare two directory hierarchies at a time
after populating two temporary directories, instead of running an
instance of the external tool once per a file pair.
* The "fmt-merge-msg" command learned to list the primary contributors
involved in the side topic you are merging in a comment in the merge
commit template.
* "git rebase" learned to optionally keep commits that do not
introduce any change in the original history.
* "git push --recurse-submodules" learned to optionally look into the
histories of submodules bound to the superproject and push them
out.
* A 'snapshot' request to "gitweb" honors If-Modified-Since: header,
based on the commit date.
* "gitweb" learned to highlight the patch it outputs even more.
Fixes since v1.7.10.4
---------------------
* "git fast-export" did not give a readable error message when the
same mark erroneously appeared twice in the --import-marks input.
* "git rebase -p" used to pay attention to rebase.autosquash which
was wrong. "git rebase -p -i" should, but "git rebase -p" by
itself should not.
(ok agc@ to commit during freeze)
===========================
Fixes since v1.7.10.3
---------------------
* The message file for Swedish translation has been updated a bit.
* A name taken from mailmap was copied into an internal buffer
incorrectly and could overun the buffer if it is too long.
* A malformed commit object that has a header line chomped in the
middle could kill git with a NULL pointer dereference.
* An author/committer name that is a single character was mishandled
as an invalid name by mistake.
* The progress indicator for a large "git checkout" was sent to
stderr even if it is not a terminal.
* "git grep -e '$pattern'", unlike the case where the patterns are
read from a file, did not treat individual lines in the given
pattern argument as separate regular expressions as it should.
* When "git rebase" was given a bad commit to replay the history on,
its error message did not correctly give the command line argument
it had trouble parsing.
Also contains minor fixes and documentation updates.
* The message file for German translation has been updated a bit.
* Running "git checkout" on an unborn branch used to corrupt HEAD.
* When checking out another commit from an already detached state, we
used to report all commits that are not reachable from any of the
refs as lossage, but some of them might be reachable from the new
HEAD, and there is no need to warn about them.
* Some time ago, "git clone" lost the progress output for its
"checkout" phase; when run without any "--quiet" option, it should
give progress to the lengthy operation.
* The directory path used in "git diff --no-index", when it recurses
down, was broken with a recent update after v1.7.10.1 release.
* "log -z --pretty=tformat:..." did not terminate each record with
NUL. The fix is not entirely correct when the output also asks for
--patch and/or --stat, though.
* The DWIM behaviour for "log --pretty=format:%gd -g" was somewhat
broken and gave undue precedence to configured log.date, causing
"git stash list" to show "stash@{time stamp string}".
* "git status --porcelain" ignored "--branch" option by mistake. The
output for "git status --branch -z" was also incorrect and did not
terminate the record for the current branch name with NUL as asked.
* When a submodule repository uses alternate object store mechanism,
some commands that were started from the superproject did not
notice it and failed with "No such object" errors. The subcommands
of "git submodule" command that recursed into the submodule in a
separate process were OK; only the ones that cheated and peeked
directly into the submodule's repository from the primary process
were affected.
* The test scaffolding for git-daemon was flaky.
* The test scaffolding for fast-import was flaky.
* The filesystem boundary was not correctly reported when .git directory
discovery stopped at a mount point.
* HTTP transport that requires authentication did not work correctly when
multiple connections are used simultaneously.
* Minor memory leak during unpack_trees (hence "merge" and "checkout"
to check out another branch) has been plugged.
* In the older days, the header "Conflicts:" in "cherry-pick" and "merge"
was separated by a blank line from the list of paths that follow for
readability, but when "merge" was rewritten in C, we lost it by
mistake. Remove the newline from "cherry-pick" to make them match
again.
* The command line parser choked "git cherry-pick $name" when $name can
be both revision name and a pathname, even though $name can never be a
path in the context of the command.
* The "include.path" facility in the configuration mechanism added in
1.7.10 forgot to interpret "~/path" and "~user/path" as it should.
* "git config --rename-section" to rename an existing section into a
bogus one did not check the new name.
* The "diff --no-index" codepath used limited-length buffers, risking
pathnames getting truncated. Update it to use the strbuf API.
* The report from "git fetch" said "new branch" even for a non branch
ref.
* The http-backend (the server side of the smart http transfer) used
to overwrite GIT_COMMITTER_NAME and GIT_COMMITTER_EMAIL with the
value obtained from REMOTE_USER unconditionally, making it
impossible for the server side site-specific customization to use
different identity sources to affect the names logged. It now uses
REMOTE_USER only as a fallback value.
* "log --graph" was not very friendly with "--stat" option and its
output had line breaks at wrong places.
* Octopus merge strategy did not reduce heads that are recorded in the
final commit correctly.
* "git push" over smart-http lost progress output a few releases ago;
this release resurrects it.
* The error and advice messages given by "git push" when it fails due
to non-ff were not very helpful to new users; it has been broken
into three cases, and each is given a separate advice message.
* The insn sheet given by "rebase -i" did not make it clear that the
insn lines can be re-ordered to affect the order of the commits in
the resulting history.
* "git repack" used to write out unreachable objects as loose objects
when repacking, even if such loose objects will immediately pruned
due to its age.
* A contrib script "rerere-train" did not work out of the box unless
user futzed with her $PATH.
* "git rev-parse --show-prefix" used to emit nothing when run at the
top-level of the working tree, but now it gives a blank line.
* The i18n of error message "git stash save" was not properly done.
* "git submodule" used a sed script that some platforms mishandled.
* When using a Perl script on a system where "perl" found on user's
$PATH could be ancient or otherwise broken, we allow builders to
specify the path to a good copy of Perl with $PERL_PATH. The
gitweb test forgot to use that Perl when running its test.
* various "gitk" updates.
- show the path to the top level directory in the window title
- update preference edit dialog
- display file list correctly when directories are given on command line
- make "git-describe" output in the log message into a clickable link
- avoid matching the UNIX timestamp part when searching all fields
- give preference to symbolic font names like sans & monospace
- allow comparing two commits using a mark
- "gitk" honors log.showroot configuration.
* Teams for localizing the messages from the Porcelain layer of
commands are starting to form, thanks to Jiang Xin who volunteered
to be the localization coordinator. Translated messages for
simplified Chinese, Swedish and Portuguese are available.
* The configuration mechanism learned an "include" facility; an
assignment to the include.path pseudo-variable causes the named
file to be included in-place when Git looks up configuration
variables.
* A content filter (clean/smudge) used to be just a way to make the
recorded contents "more useful", and allowed to fail; a filter can
now optionally be marked as "required".
* Options whose names begin with "--no-" (e.g. the "--no-verify"
option of the "git commit" command) can be negated by omitting
"no-" from its name, e.g. "git commit --verify".
* "git am" learned to pass "-b" option to underlying "git mailinfo", so
that a bracketed string other than "PATCH" at the beginning can be kept.
* "git clone" learned "--single-branch" option to limit cloning to a
single branch (surprise!); tags that do not point into the history
of the branch are not fetched.
* "git clone" learned to detach the HEAD in the resulting repository
when the user specifies a tag with "--branch" (e.g., "--branch=v1.0").
Clone also learned to print the usual "detached HEAD" advice in such
a case, similar to "git checkout v1.0".
* When showing a patch while ignoring whitespace changes, the context
lines are taken from the postimage, in order to make it easier to
view the output.
More...
* "git p4" (in contrib/) submit the changes to a wrong place when the
"--use-client-spec" option is set.
* The config.mak.autogen generated by optional autoconf support tried
to link the binary with -lintl even when libintl.h is missing from
the system.
* When the filter driver exits before reading the content before the
main git process writes the contents to be filtered to the pipe to
it, the latter could be killed with SIGPIPE instead of ignoring
such an event as an error.
* "git add --refresh <pathspec>" used to warn about unmerged paths
outside the given pathspec.
* The bulk check-in codepath in "git add" streamed contents that
needs smudge/clean filters without running them, instead of punting
and delegating to the codepath to run filters after slurping
everything to core.
* "git branch --with $that" assumed incorrectly that the user will never
ask the question with nonsense value in $that.
* "git bundle create" produced a corrupt bundle file upon seeing
commits with excessively long subject line.
* When a remote helper exits before reading the blank line from the
main git process to signal the end of commands, the latter could be
killed with SIGPIPE. Instead we should ignore such event as a
non-error.
* The commit log template given with "git merge --edit" did not have
a short instructive text like what "git commit" gives.
* "git rev-list --verify-objects -q" omitted the extra verification
it needs to do over "git rev-list --objects -q" by mistake.
* "gitweb" used to drop warnings in the log file when "heads" view is
accessed in a repository whose HEAD does not point at a valid
branch.
* An invalid regular expression pattern given by an end user made
"gitweb" to return garbled response.
Fixes since v1.7.9.1
--------------------
* Bash completion script (in contrib/) did not like a pattern that
begins with a dash to be passed to __git_ps1 helper function.
* Adaptation of the bash completion script (in contrib/) for zsh
incorrectly listed all subcommands when "git <TAB><TAB>" was given
to ask for list of porcelain subcommands.
* The build procedure for profile-directed optimized binary was not
working very well.
* Some systems need to explicitly link -lcharset to get locale_charset().
* t5541 ignored user-supplied port number used for HTTP server testing.
* The error message emitted when we see an empty loose object was
not phrased correctly.
* The code to ask for password did not fall back to the terminal
input when GIT_ASKPASS is set but does not work (e.g. lack of X
with GUI askpass helper).
* We failed to give the true terminal width to any subcommand when
they are invoked with the pager, i.e. "git -p cmd".
* map_user() was not rewriting its output correctly, which resulted
in the user visible symptom that "git blame -e" sometimes showed
excess '>' at the end of email addresses.
* "git checkout -b" did not allow switching out of an unborn branch.
* When you have both .../foo and .../foo.git, "git clone .../foo" did not
favor the former but the latter.
* "git commit" refused to create a commit when entries added with
"add -N" remained in the index, without telling Git what their content
in the next commit should be. We should have created the commit without
these paths.
* "git diff --stat" said "files", "insertions", and "deletions" even
when it is showing one "file", one "insertion" or one "deletion".
* The output from "git diff --stat" for two paths that have the same
amount of changes showed graph bars of different length due to the
way we handled rounding errors.
* "git grep" did not pay attention to -diff (hence -binary) attribute.
* The transport programs (fetch, push, clone)ignored --no-progress
and showed progress when sending their output to a terminal.
* Sometimes error status detected by a check in an earlier phase of
"git receive-pack" (the other end of "git push") was lost by later
checks, resulting in false indication of success.
* "git rev-list --verify" sometimes skipped verification depending on
the phase of the moon, which dates back to 1.7.8.x series.
* Search box in "gitweb" did not accept non-ASCII characters correctly.
* Search interface of "gitweb" did not show multiple matches in the same file
correctly.
Also contains minor fixes and documentation updates.
* gitk updates accumulated since early 2011.
* git-gui updated to 0.16.0.
* git-p4 (in contrib/) updates.
* Git uses gettext to translate its most common interface messages
into the user's language if translations are available and the
locale is appropriately set. Distributors can drop new PO files
in po/ to add new translations.
* The code to handle username/password for HTTP transactions used in
"git push" & "git fetch" learned to talk "credential API" to
external programs to cache or store them, to allow integration with
platform native keychain mechanisms.
* The input prompts in the terminal use our own getpass() replacement
when possible. HTTP transactions used to ask for the username without
echoing back what was typed, but with this change you will see it as
you type.
* The internals of "revert/cherry-pick" have been tweaked to prepare
building more generic "sequencer" on top of the implementation that
drives them.
* "git rev-parse FETCH_HEAD" after "git fetch" without specifying
what to fetch from the command line will now show the commit that
would be merged if the command were "git pull".
* "git add" learned to stream large files directly into a packfile
instead of writing them into individual loose object files.
* "git checkout -B <current branch> <elsewhere>" is a more intuitive
way to spell "git reset --keep <elsewhere>".
* "git checkout" and "git merge" learned "--no-overwrite-ignore" option
to tell Git that untracked and ignored files are not expendable.
* "git commit --amend" learned "--no-edit" option to say that the
user is amending the tree being recorded, without updating the
commit log message.
* "git commit" and "git reset" re-learned the optimization to prime
the cache-tree information in the index, which makes it faster to
write a tree object out after the index entries are updated.
* "git commit" detects and rejects an attempt to stuff NUL byte in
the commit log message.
* "git commit" learned "-S" to GPG-sign the commit; this can be shown
with the "--show-signature" option to "git log".
* fsck and prune are relatively lengthy operations that still go
silent while making the end-user wait. They learned to give progress
output like other slow operations.
* The set of built-in function-header patterns for various languages
knows MATLAB.
* "git log --format='<format>'" learned new %g[nNeE] specifiers to
show information from the reflog entries when walking the reflog
(i.e. with "-g").
* "git pull" can be used to fetch and merge an annotated/signed tag,
instead of the tip of a topic branch. The GPG signature from the
signed tag is recorded in the resulting merge commit for later
auditing.
* "git log" learned "--show-signature" option to show the signed tag
that was merged that is embedded in the merge commit. It also can
show the signature made on the commit with "git commit -S".
* "git branch --edit-description" can be used to add descriptive text
to explain what a topic branch is about.
* "git fmt-merge-msg" learned to take the branch description into
account when preparing a merge summary that "git merge" records
when merging a local branch.
* "git request-pull" has been updated to convey more information
useful for integrators to decide if a topic is worth merging and
what is pulled is indeed what the requestor asked to pull,
including:
- the tip of the branch being requested to be merged;
- the branch description describing what the topic is about;
- the contents of the annotated tag, when requesting to pull a tag.
* "git pull" learned to notice 'pull.rebase' configuration variable,
which serves as a global fallback for setting 'branch.<name>.rebase'
configuration variable per branch.
* "git tag" learned "--cleanup" option to control how the whitespaces
and empty lines in tag message are cleaned up.
* "gitweb" learned to show side-by-side diff.
* The code to look up attributes for paths reused entries from a wrong
directory when two paths in question are in adjacent directories and
the name of the one directory is a prefix of the other.
* A wildcard that matches deeper hierarchy given to the "diff-index" command,
e.g. "git diff-index HEAD -- '*.txt'", incorrectly reported additions of
matching files even when there is no change.
* When producing a "thin pack" (primarily used in bundles and smart
HTTP transfers) out of a fully packed repository, we unnecessarily
avoided sending recent objects as a delta against objects we know
the other side has.
* "git send-email" did not properly treat sendemail.multiedit as a
boolean (e.g. setting it to "false" did not turn it off).
* Also contains minor fixes and documentation updates.
Fixes since v1.7.7.2
--------------------
* Adjust the "quick-install-doc" procedures as preformatted
html/manpage are no longer in the source repository.
* The logic to optimize the locality of the data in a pack introduced in
1.7.7 was grossly inefficient.
* The logic to filter out forked projects in the project list in
"gitweb" was broken for some time.
* "git branch -m/-M" advertised to update RENAME_REF ref in the
commit log message that introduced the feature but not anywhere in
the documentation, and never did update such a ref anyway. This
undocumented misfeature that did not exist has been excised.
Fixes since v1.7.7.3
--------------------
* A few header dependencies were missing from the Makefile.
* Some newer parts of the code used C99 __VA_ARGS__ while we still
try to cater to older compilers.
* "git name-rev --all" tried to name all _objects_, naturally failing to
describe many blobs and trees, instead of showing only commits as
advertised in its documentation.
Fixes since v1.7.7.4
--------------------
* After fetching from a remote that has very long refname, the reporting
output could have corrupted by overrunning a static buffer.
* "git checkout" and "git merge" treated in-tree .gitignore and exclude
file in $GIT_DIR/info/ directory inconsistently when deciding which
untracked files are ignored and expendable.
Fixes since v1.7.7.5
--------------------
* The code to look up attributes for paths reused entries from a wrong
directory when two paths in question are in adjacent directories and
the name of the one directory is a prefix of the other.
* A wildcard that matches deeper hierarchy given to the "diff-index" command,
e.g. "git diff-index HEAD -- '*.txt'", incorrectly reported additions of
matching files even when there is no change.
* When producing a "thin pack" (primarily used in bundles and smart
HTTP transfers) out of a fully packed repository, we unnecessarily
avoided sending recent objects as a delta against objects we know
the other side has.
unexpected that man pages, which most people would want, are split,
while tk and python dependencies, expected to be the first to be
jettisoned on small systems, are in the base package.
* We used to drop error messages from libcurl on certain kinds of
errors.
* Error report from smart HTTP transport, when the connection was
broken in the middle of a transfer, showed a useless message on
a corrupt packet.
* "git fetch --prune" was unsafe when used with refspecs from the
command line.
* The attribute mechanism did not use case insensitive match when
core.ignorecase was set.
* "git bisect" did not notice when it failed to update the working tree
to the next commit to be tested.
* "git config --bool --get-regexp" failed to separate the variable name
and its value "true" when the variable is defined without "= true".
* "git remote rename $a $b" were not careful to match the remote name
against $a (i.e. source side of the remote nickname).
* "git mergetool" did not use its arguments as pathspec, but as a path to
the file that may not even have any conflict.
* "git diff --[num]stat" used to use the number of lines of context
different from the default, potentially giving different results from
"git diff | diffstat" and confusing the users.
* "git pull" and "git rebase" did not work well even when GIT_WORK_TREE is
set correctly with GIT_DIR if the current directory is outside the working
tree.
* "git send-email" did not honor the configured hostname when restarting
the HELO/EHLO exchange after switching TLS on.
* "gitweb" used to produce a non-working link while showing the contents
of a blob, when JavaScript actions are enabled.
* The scripting part of the codebase is getting prepared for i18n/l10n.
* Interix, Cygwin and Minix ports got updated.
* Various updates to git-p4 (in contrib/), fast-import, and git-svn.
* Gitweb learned to read from /etc/gitweb-common.conf when it exists,
before reading from gitweb_config.perl or from /etc/gitweb.conf
* Various codepaths that invoked zlib deflate/inflate assumed that these
functions can compress or uncompress more than 4GB data in one call on
platforms with 64-bit long, which has been corrected.
* Git now recognizes loose objects written by other implementations that
use a non-standard window size for zlib deflation (e.g. Agit running on
Android with 4kb window). We used to reject anything that was not
deflated with 32kb window.
* Interaction between the use of pager and coloring of the output has
been improved, especially when a command that is not built-in was
involved.
* "git am" learned to pass the "--exclude=<path>" option through to underlying
"git apply".
* You can now feed many empty lines before feeding an mbox file to "git am".
* "git archive" can be told to pass the output to gzip compression and
produce "archive.tar.gz".
* "git bisect" can be used in a bare repository (provided that the test
you perform per each iteration does not need a working tree, of course).
* The length of abbreviated object names in "git branch -v" output
now honors the core.abbrev configuration variable.
* "git check-attr" can take relative paths from the command line.
* "git check-attr" learned an "--all" option to list the attributes for a
given path.
* "git checkout" (both the code to update the files upon checking out a
different branch and the code to checkout a specific set of files) learned
to stream the data from object store when possible, without having to
read the entire contents of a file into memory first. An earlier round
of this code that is not in any released version had a large leak but
now it has been plugged.
* "git clone" can now take a "--config key=value" option to set the
repository configuration options that affect the initial checkout.
* "git commit <paths>..." now lets you feed relative pathspecs that
refer to outside your current subdirectory.
* "git diff --stat" learned a --stat-count option to limit the output of
a diffstat report.
* "git diff" learned a "--histogram" option to use a different diff
generation machinery stolen from jgit, which might give better performance.
* "git diff" had a weird worst case behaviour that can be triggered
when comparing files with potentially many places that could match.
* "git fetch", "git push" and friends no longer show connection
errors for addresses that couldn't be connected to when at least one
address succeeds (this is arguably a regression but a deliberate one).
* "git grep" learned "--break" and "--heading" options, to let users mimic
the output format of "ack".
* "git grep" learned a "-W" option that shows wider context using the same
logic used by "git diff" to determine the hunk header.
* Invoking the low-level "git http-fetch" without "-a" option (which
git itself never did---normal users should not have to worry about
this) is now deprecated.
* The "--decorate" option to "git log" and its family learned to
highlight grafted and replaced commits.
* "git rebase master topci" no longer spews usage hints after giving
the "fatal: no such branch: topci" error message.
* The recursive merge strategy implementation got a fairly large
fix for many corner cases that may rarely happen in real world
projects (it has been verified that none of the 16000+ merges in
the Linux kernel history back to v2.6.12 is affected with the
corner case bugs this update fixes).
* "git stash" learned an "--include-untracked option".
* "git submodule update" used to stop at the first error updating a
submodule; it now goes on to update other submodules that can be
updated, and reports the ones with errors at the end.
* "git push" can be told with the "--recurse-submodules=check" option to
refuse pushing of the supermodule, if any of its submodules'
commits hasn't been pushed out to their remotes.
* "git upload-pack" and "git receive-pack" learned to pretend that only a
subset of the refs exist in a repository. This may help a site to
put many tiny repositories into one repository (this would not be
useful for larger repositories as repacking would be problematic).
* "git verify-pack" has been rewritten to use the "index-pack" machinery
that is more efficient in reading objects in packfiles.
* test scripts for gitweb tried to run even when CGI-related perl modules
are not installed; they now exit early when the latter are unavailable.
to kernel.org issues.
1.7.6.3:
* "git -c var=value subcmd" misparsed the custom configuration when
value contained an equal sign.
* "git fetch" had a major performance regression, wasting many
needless cycles in a repository where there is no submodules
present. This was especially bad, when there were many refs.
* "git reflog $refname" did not default to the "show" subcommand as
the documentation advertised the command to do.
* "git reset" did not leave meaningful log message in the reflog.
* "git status --ignored" did not show ignored items when there is no
untracked items.
* "git tag --contains $commit" was unnecessarily inefficient.
Also contains minor fixes and documentation updates.
1.7.6.2:
Junio C Hamano (3):
whitespace: have SP on both sides of an assignment "="
Revert "Merge branch 'cb/maint-quiet-push' into maint"
Git 1.7.6.2
Pang Yan Han (1):
update-ref: whitespace fix
Thomas Rast (1):
Documentation: clarify effects of -- <path> arguments
* Various codepaths that invoked zlib deflate/inflate assumed that these
functions can compress or uncompress more than 4GB data in one call on
platforms with 64-bit long, which has been corrected.
* "git unexecutable" reported that "unexecutable" was not found, even
though the actual error was that "unexecutable" was found but did
not have a proper she-bang line to be executed.
* Error exits from $PAGER were silently ignored.
* "git checkout -b <branch>" was confused when attempting to create a
branch whose name ends with "-g" followed by hexadecimal digits,
and refused to work.
* "git checkout -b <branch>" sometimes wrote a bogus reflog entry,
causing later "git checkout -" to fail.
* "git diff --cc" learned to correctly ignore binary files.
* "git diff -c/--cc" mishandled a deletion that resolves a conflict, and
looked in the working tree instead.
* "git fast-export" forgot to quote pathnames with unsafe characters
in its output.
* "git fetch" over smart-http transport used to abort when the
repository was updated between the initial connection and the
subsequent object transfer.
* "git fetch" did not recurse into submodules in subdirectories.
* "git ls-tree" did not error out when asked to show a corrupt tree.
* "git pull" without any argument left an extra whitespace after the
command name in its reflog.
* "git push --quiet" was not really quiet.
* "git rebase -i -p" incorrectly dropped commits from side branches.
* "git reset [<commit>] paths..." did not reset the index entry correctly
for unmerged paths.
* "git submodule add" did not allow a relative repository path when
the superproject did not have any default remote url.
* "git submodule foreach" failed to correctly give the standard input to
the user-supplied command it invoked.
* submodules that the user has never showed interest in by running
"git submodule init" was incorrectly marked as interesting by "git
submodule sync".
* "git submodule update --quiet" was not really quiet.
* "git tag -l <glob>..." did not take multiple glob patterns from the
command line.
* Various git-svn updates.
* Updates the way content tags are handled in gitweb. Also adds
a UI to choose common timezone for displaying the dates.
* Similar to branch names, tagnames that begin with "-" are now
disallowed.
* Clean-up of the C part of i18n (but not l10n---please wait)
continues.
* The scripting part of the codebase is getting prepared for i18n/l10n.
* Pushing and pulling from a repository with large number of refs that
point to identical commits are optimized by not listing the same commit
during the common ancestor negotiation exchange with the other side.
* Adding a file larger than core.bigfilethreshold (defaults to 1/2 Gig)
using "git add" will send the contents straight to a packfile without
having to hold it and its compressed representation both at the same
time in memory.
* Processes spawned by "[alias] <name> = !process" in the configuration
can inspect GIT_PREFIX environment variable to learn where in the
working tree the original command was invoked.
* A magic pathspec ":/" tells a command that limits its operation to
the current directory when ran from a subdirectory to work on the
entire working tree. In general, ":/path/to/file" would be relative
to the root of the working tree hierarchy.
After "git reset --hard; edit Makefile; cd t/", "git add -u" would
be a no-op, but "git add -u :/" would add the updated contents of
the Makefile at the top level. If you want to name a path in the
current subdirectory whose unusual name begins with ":/", you can
name it by "./:/that/path" or by "\:/that/path".
* "git blame" learned "--abbrev[=<n>]" option to control the minimum
number of hexdigits shown for commit object names.
* "git blame" learned "--line-porcelain" that is less efficient but is
easier to parse.
* Aborting "git commit --interactive" discards updates to the index
made during the interactive session.
* More...
* The single-key mode of "git add -p" was easily fooled into thinking
that it was told to add everthing ('a') when up-arrow was pressed by
mistake.
* Setting a git command that uses custom configuration via "-c var=val"
as an alias caused a crash due to a realloc(3) failure.
* "git diff -C -C" used to disable the rename detection entirely when
there are too many copy candidate paths in the tree; now it falls
back to "-C" when doing so would keep the copy candidate paths
under the rename detection limit.
* "git rerere" did not diagnose a corrupt MERGE_RR file in some cases.
* "git add -p" did not work correctly when a hunk is split and then
one of them was given to the editor.
* "git add -u" did not resolve a conflict where our history deleted and
their history modified the same file, and the working tree resolved to
keep a file.
* "git cvsimport" did not know that CVSNT stores its password file in a
location different from the traditional CVS.
* "git diff-files" did not show the mode information from the working
tree side of an unmerged path correctly.
* "git diff -M --cached" used to use unmerged path as a possible rename
source candidate, which made no sense.
* The option name parser in "git fast-import" used prefix matches for
some options where it shouldn't, and accepted non-existent options,
e.g. "--relative-marksmith" or "--forceps".
* "git format-patch" did not quote RFC822 special characters in the
email address (e.g From: Junio C. Hamano <jch@example.com>, not
From: "Junio C. Hamano" <jch@example.com>).
* "git format-patch" when run with "--quiet" option used to produce a
nonsense result that consists of alternating empty output.
* In "git merge", per-branch branch.<name>.mergeoptions configuration
variables did not override the fallback default merge.<option>
configuration variables such as merge.ff, merge.log, etc.
* "git merge-one-file" did not honor GIT_WORK_TREE settings when
handling a "both sides added, differently" conflict.
* "git mergetool" did not handle conflicted submoudules gracefully.
* "git-p4" (in contrib) used a wrong base image while merge a file that
was added on both branches differently.
* "git rebase -i -p" failed to preserve the history when there is a
redundant merge created with the --no-ff option.
* When an object "$tree:$path" does not exist, if $path does exist in the
subtree of $tree that corresponds to the subdirectory the user is in,
git now suggests using "$tree:./$path" in addition to the advice to use
the full path from the root of the working tree.
* The "--date=relative" output format used to say "X years, 12 months"
when it should have said "X+1 years".
* The smart-HTTP transfer was broken in 1.7.5 when the client needs
to issue a small POST (which uses content-length) and then a large
POST (which uses chunked) back to back.
* "git clean" used to fail on an empty directory that is not readable,
even though rmdir(2) could remove such a directory. Now we attempt it
as the last resort.
* The "--dirstat" option of "diff" family of commands used to totally
ignore a change that only rearranged lines within a file. Such a
change now counts as at least a minimum but non zero change.
* The "--dirstat" option of "diff" family of commands used to use the
pathname in the original, instead of the pathname in the result,
when renames are involved.
* "git pack-object" did not take core.bigfilethreashold into account
(unlike fast-import); now it does.
* "git reflog" ignored options like "--format=.." on the command line.
* "git stash apply" used to refuse to work if there was any change in
the working tree, even when the change did not overlap with the change
the stash recorded.
* "git stash apply @{99999}" was not diagnosed as an error, even when you
did not have that many stash entries.
* An error message from "git send-email" to diagnose a broken SMTP
connection configuration lacked a space between "hello=<smtp-domain>"
and "port=<smtp-server-port>".
* Various MinGW portability fixes.
* Various git-p4 enhancements (in contrib).
* Various vcs-svn, git-svn and gitk enhancements and fixes.
* Various git-gui updates (0.14.0).
* Update to more modern HP-UX port.
* The codebase is getting prepared for i18n/l10n; no translated
strings nor translation mechanism in the code yet, but the strings
are being marked for l10n.
* The bash completion script can now complete symmetric difference
for "git diff" command, e.g. "git diff ...bra<TAB>".
* The default minimum length of abbreviated and unique object names
can now be configured by setting the core.abbrev configuration
variable.
* "git apply -v" reports offset lines when the patch does not apply at
the exact location recorded in the diff output.
* "git config" used to be also known as "git repo-config", but the old
name is now officially deprecated.
* "git checkout --detach <commit>" is a more user friendly synonym for
"git checkout <commit>^0".
* "git checkout" performed on detached HEAD gives a warning and
advice when the commit being left behind will become unreachable from
any branch or tag.
* "git cherry-pick" and "git revert" can be told to use a custom merge
strategy, similar to "git rebase".
* "git cherry-pick" remembers which commit failed to apply when it is
stopped by conflicts, making it unnecessary to use "commit -c $commit"
to conclude it.
* "git cvsimport" bails out immediately when the cvs server cannot be
reached, without spewing unnecessary error messages that complain about
the server response it never got.
* "git fetch" vs "git upload-pack" transfer learned 'no-done'
protocol extension to save one round-trip after the content
negotiation is done. This saves one HTTP RPC, reducing the overall
latency for a trivial fetch.
* "git fetch" can be told to recursively fetch submodules on-demand.
* "git grep -f <filename>" learned to treat "-" as "read from the
standard input stream".
* "git grep --no-index" did not honor pathspecs correctly, returning
paths outside the specified area.
* "git init" learned the --separate-git-dir option to allow the git
directory for a new repository created elsewhere and linked via the
gitdir mechanism. This is primarily to help submodule support later
to switch between a branch of superproject that has the submodule
and another that does not.
* "git log" type commands now understand globbing pathspecs. You
can say "git log -- '*.txt'" for example.
* "git log" family of commands learned --cherry and --cherry-mark
options that can be used to view two diverged branches while omitting
or highlighting equivalent changes that appear on both sides of a
symmetric difference (e.g. "log --cherry A...B").
* A lazy "git merge" that didn't say what to merge used to be an error.
When run on a branch that has an upstream defined, however, the command
now merges from the configured upstream.
* "git mergetool" learned how to drive "beyond compare 3" as well.
* "git rerere forget" without pathspec used to forget all the saved
conflicts that relate to the current merge; it now requires you to
give it pathspecs.
* "git rev-list --objects $revs -- $pathspec" now limits the objects listed
in its output properly with the pathspec, in preparation for narrow
clones.
* "git push" with no parameters gives better advice messages when
"tracking" is used as the push.default semantics or there is no remote
configured yet.
* A possible value to the "push.default" configuration variable,
'tracking', gained a synonym that more naturally describes what it
does, 'upstream'.
* "git rerere" learned a new subcommand "remaining" that is similar to
"status" and lists the paths that had conflicts which are known to
rerere, but excludes the paths that have already been marked as
resolved in the index from its output. "git mergetool" has been
updated to use this facility.
* Compilation of sha1_file.c on BSD platforms were broken due to our
recent use of getrlimit() without including <sys/resource.h>.
* "git config" did not diagnose incorrect configuration variable names.
* "git format-patch" did not wrap a long subject line that resulted from
rfc2047 encoding.
* "git instaweb" should work better again with plackup.
* "git log --max-count=4 -Sfoobar" now shows 4 commits that changes the
number of occurrences of string "foobar"; it used to scan only for 4
commits and then emitted only matching ones.
* "git log --first-parent --boundary $c^..$c" segfaulted on a merge.
* "git pull" into an empty branch should have behaved as if
fast-forwarding from emptiness to the version being pulled, with
the usual protection against overwriting untracked files.
* "git submodule" that is run while a merge in the superproject is in
conflicted state tried to process each conflicted submodule up to
three times.
* "git status" spent all the effort to notice racily-clean index entries
but didn't update the index file to help later operations go faster in
some cases.
* "git apply" used to confuse lines updated by previous hunks as lines
that existed before when applying a hunk, contributing misapplication
of patches with offsets.
* "git branch --track" (and "git checkout --track --branch") used to
allow setting up a random non-branch that does not make sense to follow
as the "upstream". The command correctly diagnoses it as an error.
* "git checkout $other_branch" silently removed untracked symbolic links
in the working tree that are in the way in order to check out paths
under it from the named branch.
* "git cvsimport" did not bail out immediately when the cvs server cannot
be reached, spewing unnecessary error messages that complain about the
server response that it never got.
* "git diff --quiet" did not work very well with the "--diff-filter" option.
* "git grep -n" lacked a long-hand synonym --line-number.
* "git stash apply" reported the result of its operation by running
"git status" from the top-level of the working tree; it should (and
now does) run it from the user's working directory.
* On Windows platform, the codepath to spawn a new child process forgot
to first flush the output buffer.
* "git bundle" did not use OFS_DELTA encoding, making its output a few
per-cent larger than necessarily.
* The option to tell "git clone" to recurse into the submodules was
misspelled with an underscore "--recurse_submodules".
* "git diff --cached HEAD" before the first commit does what an end user
would expect (namely, show what would be committed without further "git
add").
* "git fast-import" didn't accept the command to ask for "notes" feature
to be present in its input stream, even though it was capable of the
feature.
* "git fsck" gave up scanning loose object files in directories with
garbage files.
* The xfuncname pattern used by "git diff" and "git grep" to show the
last notable line in context were broken for python and ruby for a long
time.
* "git merge" into an unborn branch removed an untracked file "foo" from
the working tree when merged branch had "foo" (this fix was already in
1.7.3.3 but was omitted from the release notes by mistake).
* "git status -s" did not quote unprintable characters in paths as
documented.
* "git am --abort" used to always reset to the commit at the beginning of
the last "am" invocation that has stopped, losing any unrelated commits
that may have been made since then. Now it refrains from doing so and
instead issues a warning.
* "git blame" incorrectly reused bogusly cached result of textconv
filter for files from the working tree.
* "git commit" used to abort after the user edited the log message
when the committer information was not correctly set up. It now
aborts before starting the editor.
* "git commit --date=invalid" used to silently ignore the incorrectly
specified date; it is now diagnosed as an error.
* "git rebase --skip" to skip the last commit in a series used to fail
to run post-rewrite hook and to copy notes from old commits that have
successfully been rebased so far. Now it do (backmerge ef88ad2).
* "gitweb" tried to show a wrong feed logo when none was specified.
Git v1.7.3.4 Release Notes
==========================
Fixes since v1.7.3.3
--------------------
* Smart HTTP transport used to incorrectly retry redirected POST
request with GET request.
* "git apply" did not correctly handle patches that only change modes
if told to apply while stripping leading paths with -p option.
* "git apply" can deal with patches with timezone formatted with a
colon between the hours and minutes part (e.g. "-08:00" instead of
"-0800").
* "git checkout" removed an untracked file "foo" from the working
tree when switching to a branch that contains a tracked path
"foo/bar". Prevent this, just like the case where the conflicting
path were "foo" (c752e7f..7980872d).
* "git cherry-pick" or "git revert" refused to work when a path that
would be modified by the operation was stat-dirty without a real
difference in the contents of the file.
* "git diff --check" reported an incorrect line number for added
blank lines at the end of file.
* "git imap-send" failed to build under NO_OPENSSL.
* Setting log.decorate configuration variable to "0" or "1" to mean
"false" or "true" did not work.
* "git push" over dumb HTTP protocol did not work against WebDAV
servers that did not terminate a collection name with a slash.
* "git tag -v" did not work with GPG signatures in rfc1991 mode.
* The post-receive-email sample hook was accidentally broken in 1.7.3.3
update.
* "gitweb" can sometimes be tricked into parrotting a filename argument
given in a request without properly quoting.
Other minor fixes and documentation updates are also included.
Git v1.7.3.3 Release Notes
==========================
In addition to the usual fixes, this release also includes support for
the new "add.ignoreErrors" name given to the existing "add.ignore-errors"
configuration variable.
The next version, Git 1.7.4, and future versions, will support both
old and incorrect name and the new corrected name, but without this
backport, users who want to use the new name "add.ignoreErrors" in
their repositories cannot use older versions of Git.
Fixes since v1.7.3.2
--------------------
* "git apply" segfaulted when a bogus input is fed to it.
* Running "git cherry-pick --ff" on a root commit segfaulted.
* "diff", "blame" and friends incorrectly applied textconv filters to
symlinks.
* Highlighting of whitespace breakage in "diff" output was showing
incorrect amount of whitespaces when blank-at-eol is set and the line
consisted only of whitespaces and a TAB.
* "diff" was overly inefficient when trying to find the line to use for
the function header (i.e. equivalent to --show-c-function of GNU diff).
* "git imap-send" depends on libcrypto but our build rule relied on the
linker to implicitly link it via libssl, which was wrong.
* "git merge-file" can be called from within a subdirectory now.
* "git repack -f" expanded and recompressed non-delta objects in the
existing pack, which was wasteful. Use new "-F" option if you really
want to (e.g. when changing the pack.compression level).
* "git rev-list --format="...%x00..." incorrectly chopped its output
at NUL.
* "git send-email" did not correctly remove duplicate mail addresses from
the Cc: header that appear on the To: header.
* The completion script (in contrib/completion) ignored lightweight tags
in __git_ps1().
* "git-blame" mode (in contrib/emacs) didn't say (require 'format-spec)
even though it depends on it; it didn't work with Emacs 22 or older
unless Gnus is used.
* "git-p4" (in contrib/) did not correctly handle deleted files.
Other minor fixes and documentation updates are also included.
This is primarily to push out many documentation fixes accumulated since
the 1.7.3.1 release.
Changes 1.7.3.1:
* "git stash show stash@{$n}" was accidentally broken in 1.7.3 ("git
stash show" without any argument still worked, though).
* "git stash branch $branch stash@{$n}" was accidentally broken in
1.7.3 and started dropping the named stash even when branch creation
failed.
Changes 1.7.3:
* git-gui, now at version 0.13.0, got various updates and a new
maintainer, Pat Thoyts.
* Gitweb allows its configuration to change per each request; it used to
read the configuration once upon startup.
* When git finds a corrupt object, it now reports the file that contains
it.
* "git checkout -B <it>" is a shorter way to say "git branch -f <it>"
followed by "git checkout <it>".
* When "git checkout" or "git merge" refuse to proceed in order to
protect local modification to your working tree, they used to stop
after showing just one path that might be lost. They now show all,
in a format that is easier to read.
* "git clean" learned "-e" ("--exclude") option.
* Hunk headers produced for C# files by "git diff" and friends show more
relevant context than before.
* diff.ignoresubmodules configuration variable can be used to squelch the
differences in submodules reported when running commands (e.g. "diff",
"status", etc.) at the superproject level.
* http.useragent configuration can be used to lie who you are to your
restrictive firewall.
* "git rebase --strategy <s>" learned "-X" option to pass extra options
that are understood by the chosen merge strategy.
* "git rebase -i" learned "exec" that you can insert into the insn sheet
to run a command between its steps.
* "git rebase" between branches that have many binary changes that do
not conflict should be faster.
* "git rebase -i" peeks into rebase.autosquash configuration and acts as
if you gave --autosquash from the command line.
This fixes http://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2010-2542.
Changes since 1.7.0.5:
* "git diff --stat" used "int" to count the size of differences,
which could result in overflowing.
* "git rev-list --abbrev-commit" defaulted to 40-byte abbreviations, unlike
newer tools in the git toolset.
* "make NO_CURL=NoThanks install" was broken.
* An overlong line after ".gitdir: " in a git file caused out of bounds
access to an array on the stack.
* "git config --path conf.var" to attempt to expand a variable conf.var
that uses "~/" short-hand segfaulted when $HOME environment variable
was not set.
ChangeLogs:
Updates since v1.6.6
--------------------
(subsystems)
* "git fast-import" updates; adds "option" and "feature" to detect the
mismatch between fast-import and the frontends that produce the input
stream.
* "git svn" support of subversion "merge tickets" and miscellaneous fixes.
* "gitk" and "git gui" translation updates.
* "gitweb" updates (code clean-up, load checking etc.)
(portability)
* Some more MSVC portability patches for msysgit port.
* Minimum Pthreads emulation for msysgit port.
(performance)
* More performance improvement patches for msysgit port.
(usability, bells and whistles)
* More commands learned "--quiet" and "--[no-]progress" options.
* Various commands given by the end user (e.g. diff.type.textconv,
and GIT_EDITOR) can be specified with command line arguments. E.g. it
is now possible to say "[diff "utf8doc"] textconv = nkf -w".
* "sparse checkout" feature allows only part of the work tree to be
checked out.
* HTTP transfer can use authentication scheme other than basic
(i.e./e.g. digest).
* Switching from a version of superproject that used to have a submodule
to another version of superproject that no longer has it did not remove
the submodule directory when it should (namely, when you are not
interested in the submodule at all and didn't clone/checkout).
* A new attribute conflict-marker-size can be used to change the size of
the conflict markers from the default 7; this is useful when tracked
contents (e.g. git-merge documentation) have strings that resemble the
conflict markers.
* A new syntax "<branch>@{upstream}" can be used on the command line to
substitute the name of the "upstream" of the branch. Missing branch
defaults to the current branch, so "git fetch && git merge @{upstream}"
will be equivalent to "git pull".
* "git am --resolved" has a synonym "git am --continue".
* "git branch --set-upstream" can be used to update the (surprise!) upstream,
i.e. where the branch is supposed to pull and merge from (or rebase onto).
* "git checkout A...B" is a way to detach HEAD at the merge base between
A and B.
* "git checkout -m path" to reset the work tree file back into the
conflicted state works even when you already ran "git add path" and
resolved the conflicts.
* "git commit --date='<date>'" can be used to override the author date
just like "git commit --author='<name> <email>'" can be used to
override the author identity.
* "git commit --no-status" can be used to omit the listing of the index
and the work tree status in the editor used to prepare the log message.
* "git commit" warns a bit more aggressively until you configure user.email,
whose default value almost always is not (and fundamentally cannot be)
what you want.
* "git difftool" has been extended to make it easier to integrate it
with gitk.
* "git fetch --all" can now be used in place of "git remote update".
* "git grep" does not rely on external grep anymore. It can use more than
one thread to accelerate the operation.
* "git grep" learned "--quiet" option.
* "git log" and friends learned "--glob=heads/*" syntax that is a more
flexible way to complement "--branches/--tags/--remotes".
* "git merge" learned to pass options specific to strategy-backends. E.g.
- "git merge -Xsubtree=path/to/directory" can be used to tell the subtree
strategy how much to shift the trees explicitly.
- "git merge -Xtheirs" can be used to auto-merge as much as possible,
while discarding your own changes and taking merged version in
conflicted regions.
* "git push" learned "git push origin --delete branch", a syntactic sugar
for "git push origin :branch".
* "git push" learned "git push --set-upstream origin forker:forkee" that
lets you configure your "forker" branch to later pull from "forkee"
branch at "origin".
* "git rebase --onto A...B" means the history is replayed on top of the
merge base between A and B.
* "git rebase -i" learned new action "fixup" that squashes the change
but does not affect existing log message.
* "git rebase -i" also learned --autosquash option that is useful
together with the new "fixup" action.
* "git remote" learned set-url subcommand that updates (surprise!) url
for an existing remote nickname.
* "git rerere" learned "forget path" subcommand. Together with "git
checkout -m path" it will be useful when you recorded a wrong
resolution.
* Use of "git reset --merge" has become easier when resetting away a
conflicted mess left in the work tree.
* "git rerere" had rerere.autoupdate configuration but there was no way
to countermand it from the command line; --no-rerere-autoupdate option
given to "merge", "revert", etc. fixes this.
* "git status" learned "-s(hort)" output format.
(developers)
* The infrastructure to build foreign SCM interface has been updated.
* Many more commands are now built-in.
* THREADED_DELTA_SEARCH is no more. If you build with threads, delta
compression will always take advantage of it.
Fixes since v1.6.6
------------------
* "git branch -d branch" used to refuse deleting the branch even when
the branch is fully merged to its upstream branch if it is not merged
to the current branch. It now deletes it in such a case.
* "fiter-branch" command incorrectly said --prune-empty and --filter-commit
were incompatible; the latter should be read as --commit-filter.
* When using "git status" or asking "git diff" to compare the work tree
with something, they used to consider that a checked-out submodule with
uncommitted changes is not modified; this could cause people to forget
committing these changes in the submodule before committing in the
superproject. They now consider such a change as a modification and
"git diff" will append a "-dirty" to the work tree side when generating
patch output or when used with the --submodule option.
Fixes since v1.7.0
------------------
* In a freshly created repository "rev-parse HEAD^0" complained that
it is dangling symref, even though "rev-parse HEAD" didn't.
* "git show :no-such-name" tried to access the index without bounds
check, leading to a potential segfault.
* Message from "git cherry-pick" was harder to read and use than necessary
when it stopped due to conflicting changes.
* We referred to ".git/refs/" throughout the documentation when we
meant to talk about abstract notion of "ref namespace". Because
people's repositories often have packed refs these days, this was
confusing.
* "git diff --output=/path/that/cannot/be/written" did not correctly
error out.
* "git grep -e -pattern-that-begin-with-dash paths..." could not be
spelled as "git grep -- -pattern-that-begin-with-dash paths..." which
would be a GNU way to use "--" as "end of options".
* "git grep" compiled with threading support tried to access an
uninitialized mutex on boxes with a single CPU.
* "git stash pop -q --index" failed because the unnecessary --index
option was propagated to "git stash drop" that is internally run at the
end.
Fixes since v1.7.0.1
--------------------
* GIT_PAGER was not honored consistently by some scripted Porcelains, most
notably "git am".
* updating working tree files after telling git to add them to the
index and while it is still working created garbage object files in
the repository without diagnosing it as an error.
* "git bisect -- pathspec..." did not diagnose an error condition properly when
the simplification with given pathspec made the history empty.
* "git rev-list --cherry-pick A...B" now has an obvious optimization when the
histories haven't diverged (i.e. when one end is an ancestor of the other).
* "git diff --quiet -w" did not work as expected.
* "git fast-import" didn't work with a large input, as it lacked support
for producing the pack index in v2 format.
* "git imap-send" didn't use CRLF line endings over the imap protocol
when storing its payload to the draft box, violating RFC 3501.
* "git log --format='%w(x,y,z)%b'" and friends that rewrap message
has been optimized for utf-8 payload.
* Error messages generated on the receiving end did not come back to "git
push".
* "git status" in 1.7.0 lacked the optimization we used to have in 1.6.X
* "gitweb" did not diagnose parsing errors properly while reading tis configuration
file.
Fixes since v1.7.0.2
--------------------
* Object files are created in a more ACL friendly way in repositories
where group permission is ACL controlled.
* "git add -i" didn't handle a deleted path very well.
* "git blame" padded line numbers with one extra SP when the total number
of lines was one less than multiple of ten due to an off-by-one error.
* "git fetch --all/--multi" used to discard information for remotes that
are fetched earlier.
* "git log --author=me --grep=it" tried to find commits that have "it"
or are written by "me", instead of the ones that have "it" _and_ are
written by "me".
* "git log -g branch" misbehaved when there was no entries in the reflog
for the named branch.
* "git mailinfo" (hence "git am") incorrectly removed initial indent from
paragraphs.
* "git prune" and "git reflog" (hence "git gc" as well) didn't honor
an instruction never to expire by setting gc.reflogexpire to never.
* "git push" misbehaved when branch.<name>.merge was configured without
matching branch.<name>.remote.
Fixes since v1.7.0.3
--------------------
* Optimized ntohl/htonl on big-endian machines were broken.
* Color values given to "color.<cmd>.<slot>" configuration can now have
more than one attributes (e.g. "bold ul").
* "git add -u nonexistent-path" did not complain.
* "git apply --whitespace=fix" didn't work well when an early patch in
a patch series adds trailing blank lines and a later one depended on
such a block of blank lines at the end.
* "git fast-export" didn't check error status and stop when marks file
cannot be opened.
* "git format-patch --ignore-if-in-upstream" gave unwarranted errors
when the range was empty, instead of silently finishing.
* "git remote prune" did not detect remote tracking refs that became
dangling correctly.
Fixes since v1.7.0.4
--------------------
* "git daemon" failed to compile on platforms without sockaddr_storage type.
* Output from "git rev-list --pretty=oneline" was unparsable when a
commit did not have any message, which is abnormal but possible in a
repository converted from foreign scm.
* "git stash show <commit-that-is-not-a-stash>" gave an error message
that was not so useful. Reworded the message to "<it> is not a
stash".
* Python scripts in contrib/ area now start with "#!/usr/bin/env python"
to honor user's PATH.
* "git imap-send" used to mistake any line that begins with "From " as a
message separator in format-patch output.
* Smart http server backend failed to report an internal server error and
infinitely looped instead after output pipe was closed.
Changes:
* recursive merge didn't correctly diagnose its own programming errors,
and instead caused the caller to segfault.
* The new "smart http" aware clients probed the web servers to see if
they support smart http, but did not fall back to dumb http transport
correctly with some servers.
* Time based reflog syntax e.g. "@{yesterday}" didn't diagnose a misspelled
time specification and instead assumed "@{now}".
* "git archive HEAD -- no-such-directory" produced an empty archive
without complaining.
* "git blame -L start,end -- file" misbehaved when given a start that is
larger than the number of lines in the file.
* "git checkout -m" didn't correctly call custom merge backend supplied
by the end user.
* "git config -f <file>" misbehaved when run from a subdirectory.
* "git cvsserver" didn't like having regex metacharacters (e.g. '+') in
CVSROOT environment.
* "git fast-import" did not correctly handle large blobs that may
bust the pack size limit.
* "git gui" is supposed to work even when launched from inside a .git
directory.
* "git gui" misbehaved when applying a hunk that ends with deletion.
* "git imap-send" did not honor imap.preformattedHTML as documented.
* "git log" family incorrectly showed the commit notes unconditionally by
mistake, which was especially irritating when running "git log --oneline".
* "git status" shouldn't require an write access to the repository.
* Other minor documentation updates are included.
Changes:
* "git blame" did not work well when commit lacked the author name.
* "git branch -a name" wasn't diagnosed as an error.
* "git count-objects" did not handle packfiles that are bigger than 4G on
platforms with 32-bit off_t.
* "git checkout -m other" while on a branch that does not have any commit
segfaulted, instead of failing.
* "git fast-import" choked when fed a tag that do not point at a
commit.
* "git grep" finding from work tree files could have fed garbage to
the underlying regexec(3).
* "git grep -L" didn't show empty files (they should never match, and
they should always appear in -L output as unmatching).
* "git rebase -i" did not abort cleanly if it failed to launch the editor.
* "git reset --hard" did not work correctly when GIT_WORK_TREE environment
variable is used to point at the root of the true work tree.
* http-backend was not listed in the command list in the documentation.
* Building on FreeBSD (both 7 and 8) needs OLD_ICONV set in the Makefile
* "git checkout -m some-branch" while on an unborn branch crashed.
Git v1.6.6 Release Notes
========================
Notes on behaviour change
-------------------------
* In this release, "git fsck" defaults to "git fsck --full" and
checks packfiles, and because of this it will take much longer to
complete than before. If you prefer a quicker check only on loose
objects (the old default), you can say "git fsck --no-full". This
has been supported by 1.5.4 and newer versions of git, so it is
safe to write it in your script even if you use slightly older git
on some of your machines.
Preparing yourselves for compatibility issues in 1.7.0
------------------------------------------------------
In git 1.7.0, which is planned to be the release after 1.6.6, there will
be a handful of behaviour changes that will break backward compatibility.
These changes were discussed long time ago and existing behaviours have
been identified as more problematic to the userbase than keeping them for
the sake of backward compatibility.
When necessary, a transition strategy for existing users has been designed
not to force them running around setting configuration variables and
updating their scripts in order to either keep the traditional behaviour
or adjust to the new behaviour, on the day their sysadmin decides to install
the new version of git. When we switched from "git-foo" to "git foo" in
1.6.0, even though the change had been advertised and the transition
guide had been provided for a very long time, the users procrastinated
during the entire transtion period, and ended up panicking on the day
their sysadmins updated their git installation. We are trying to avoid
repeating that unpleasantness in the 1.7.0 release.
For changes decided to be in 1.7.0, commands that will be affected
have been much louder to strongly discourage such procrastination, and
they continue to be in this release. If you have been using recent
versions of git, you would have seen warnings issued when you used
features whose behaviour will change, with a clear instruction on how
to keep the existing behaviour if you want to. You hopefully are
already well prepared.
Of course, we have also been giving "this and that will change in
1.7.0; prepare yourselves" warnings in the release notes and
announcement messages for the past few releases. Let's see how well
users will fare this time.
* "git push" into a branch that is currently checked out (i.e. pointed by
HEAD in a repository that is not bare) will be refused by default.
Similarly, "git push $there :$killed" to delete the branch $killed
in a remote repository $there, when $killed branch is the current
branch pointed at by its HEAD, will be refused by default.
Setting the configuration variables receive.denyCurrentBranch and
receive.denyDeleteCurrent to 'ignore' in the receiving repository
can be used to override these safety features. Versions of git
since 1.6.2 have issued a loud warning when you tried to do these
operations without setting the configuration, so repositories of
people who still need to be able to perform such a push should
already have been future proofed.
Please refer to:
http://git.or.cz/gitwiki/GitFaq#non-barehttp://thread.gmane.org/gmane.comp.version-control.git/107758/focus=108007
for more details on the reason why this change is needed and the
transition process that already took place so far.
* "git send-email" will not make deep threads by default when sending a
patch series with more than two messages. All messages will be sent
as a reply to the first message, i.e. cover letter. Git 1.6.6 (this
release) will issue a warning about the upcoming default change, when
it uses the traditional "deep threading" behaviour as the built-in
default. To squelch the warning but still use the "deep threading"
behaviour, give --chain-reply-to option or set sendemail.chainreplyto
to true.
It has been possible to configure send-email to send "shallow thread"
by setting sendemail.chainreplyto configuration variable to false.
The only thing 1.7.0 release will do is to change the default when
you haven't configured that variable.
* "git status" will not be "git commit --dry-run". This change does not
affect you if you run the command without pathspec.
Nobody sane found the current behaviour of "git status Makefile" useful
nor meaningful, and it confused users. "git commit --dry-run" has been
provided as a way to get the current behaviour of this command since
1.6.5.
* "git diff" traditionally treated various "ignore whitespace" options
only as a way to filter the patch output. "git diff --exit-code -b"
exited with non-zero status even if all changes were about changing the
ammount of whitespace and nothing else. and "git diff -b" showed the
"diff --git" header line for such a change without patch text.
In 1.7.0, the "ignore whitespaces" will affect the semantics of the
diff operation itself. A change that does not affect anything but
whitespaces will be reported with zero exit status when run with
--exit-code, and there will not be "diff --git" header for such a
change.
Updates since v1.6.5
--------------------
(subsystems)
* various gitk updates including use of themed widgets under Tk 8.5,
Japanese translation, a fix to a bug when running "gui blame" from
a subdirectory, etc.
* various git-gui updates including new translations, wm states fixes,
Tk bug workaround after quitting, improved heuristics to trigger gc,
etc.
* various git-svn updates.
* "git fetch" over http learned a new mode that is different from the
traditional "dumb commit walker".
(portability)
* imap-send can be built on mingw port.
(performance)
* "git diff -B" has smaller memory footprint.
(usability, bells and whistles)
* The object replace mechanism can be bypassed with --no-replace-objects
global option given to the "git" program.
* In configuration files, a few variables that name paths can begin with ~/
and ~username/ and they are expanded as expected.
* "git subcmd -h" now shows short usage help for many more subcommands.
* "git bisect reset" can reset to an arbitrary commit.
* "git checkout frotz" when there is no local branch "frotz" but there
is only one remote tracking branch "frotz" is taken as a request to
start the named branch at the corresponding remote tracking branch.
* "git commit -c/-C/--amend" can be told with a new "--reset-author" option
to ignore authorship information in the commit it is taking the message
from.
* "git describe" can be told to add "-dirty" suffix with "--dirty" option.
* "git diff" learned --submodule option to show a list of one-line logs
instead of differences between the commit object names.
* "git diff" learned to honor diff.color.func configuration to paint
function name hint printed on the hunk header "@@ -j,k +l,m @@" line
in the specified color.
* "git fetch" learned --all and --multiple options, to run fetch from
many repositories, and --prune option to remove remote tracking
branches that went stale. These make "git remote update" and "git
remote prune" less necessary (there is no plan to remove "remote
update" nor "remote prune", though).
* "git fsck" by default checks the packfiles (i.e. "--full" is the
default); you can turn it off with "git fsck --no-full".
* "git grep" can use -F (fixed strings) and -i (ignore case) together.
* import-tars contributed fast-import frontend learned more types of
compressed tarballs.
* "git instaweb" knows how to talk with mod_cgid to apache2.
* "git log --decorate" shows the location of HEAD as well.
* "git log" and "git rev-list" learned to take revs and pathspecs from
the standard input with the new "--stdin" option.
* "--pretty=format" option to "log" family of commands learned:
. to wrap text with the "%w()" specifier.
. to show reflog information with "%g[sdD]" specifier.
* "git notes" command to annotate existing commits.
* "git merge" (and "git pull") learned --ff-only option to make it fail
if the merge does not result in a fast-forward.
* "git mergetool" learned to use p4merge.
* "git rebase -i" learned "reword" that acts like "edit" but immediately
starts an editor to tweak the log message without returning control to
the shell, which is done by "edit" to give an opportunity to tweak the
contents.
* "git send-email" can be told with "--envelope-sender=auto" to use the
same address as "From:" address as the envelope sender address.
* "git send-email" will issue a warning when it defaults to the
--chain-reply-to behaviour without being told by the user and
instructs to prepare for the change of the default in 1.7.0 release.
* In "git submodule add <repository> <path>", <path> is now optional and
inferred from <repository> the same way "git clone <repository>" does.
* "git svn" learned to read SVN 1.5+ and SVK merge tickets.
* "git svn" learned to recreate empty directories tracked only by SVN.
* "gitweb" can optionally render its "blame" output incrementally (this
requires JavaScript on the client side).
* Author names shown in gitweb output are links to search commits by the
author.
Fixes since v1.6.5
------------------
All of the fixes in v1.6.5.X maintenance series are included in this
release, unless otherwise noted.
GIT v1.6.5 Release Notes
========================
In git 1.7.0, which was planned to be the release after 1.6.5, "git
push" into a branch that is currently checked out will be refused by
default.
You can choose what should happen upon such a push by setting the
configuration variable receive.denyCurrentBranch in the receiving
repository.
Also, "git push $there :$killed" to delete the branch $killed in a remote
repository $there, when $killed branch is the current branch pointed at by
its HEAD, will be refused by default.
You can choose what should happen upon such a push by setting the
configuration variable receive.denyDeleteCurrent in the receiving
repository.
To ease the transition plan, the receiving repository of such a
push running this release will issue a big warning when the
configuration variable is missing. Please refer to:
http://git.or.cz/gitwiki/GitFaq#non-barehttp://thread.gmane.org/gmane.comp.version-control.git/107758/focus=108007
for more details on the reason why this change is needed and the
transition plan.
Updates since v1.6.4
--------------------
(subsystems)
* various updates to gitk, git-svn and gitweb.
(portability)
* more improvements on mingw port.
* mingw will also give FRSX as the default value for the LESS
environment variable when the user does not have one.
* initial support to compile git on Windows with MSVC.
(performance)
* On major platforms, the system can be compiled to use with Linus's
block-sha1 implementation of the SHA-1 hash algorithm, which
outperforms the default fallback implementation we borrowed from
Mozilla.
* Unnecessary inefficiency in deepening of a shallow repository has
been removed.
* "git clone" does not grab objects that it does not need (i.e.
referenced only from refs outside refs/heads and refs/tags
hierarchy) anymore.
* The "git" main binary used to link with libcurl, which then dragged
in a large number of external libraries. When using basic plumbing
commands in scripts, this unnecessarily slowed things down. We now
implement http/https/ftp transfer as a separate executable as we
used to.
* "git clone" run locally hardlinks or copies the files in .git/ to
newly created repository. It used to give new mtime to copied files,
but this delayed garbage collection to trigger unnecessarily in the
cloned repository. We now preserve mtime for these files to avoid
this issue.
(usability, bells and whistles)
* Human writable date format to various options, e.g. --since=yesterday,
master@{2000.09.17}, are taught to infer some omitted input properly.
* A few programs gave verbose "advice" messages to help uninitiated
people when issuing error messages. An infrastructure to allow
users to squelch them has been introduced, and a few such messages
can be silenced now.
* refs/replace/ hierarchy is designed to be usable as a replacement
of the "grafts" mechanism, with the added advantage that it can be
transferred across repositories.
* "git am" learned to optionally ignore whitespace differences.
* "git am" handles input e-mail files that has CRLF line endings sensibly.
* "git am" learned "--scissors" option to allow you to discard early part
of an incoming e-mail.
* "git archive -o output.zip" works without being told what format to
use with an explicit "--format=zip".option.
* "git checkout", "git reset" and "git stash" learned to pick and
choose to use selected changes you made, similar to "git add -p".
* "git clone" learned a "-b" option to pick a HEAD to check out
different from the remote's default branch.
* "git clone" learned --recursive option.
* "git clone" from a local repository on a different filesystem used to
copy individual object files without preserving the old timestamp, giving
them extra lifetime in the new repository until they gc'ed.
* "git commit --dry-run $args" is a new recommended way to ask "what would
happen if I try to commit with these arguments."
* "git commit --dry-run" and "git status" shows conflicted paths in a
separate section to make them easier to spot during a merge.
* "git cvsimport" now supports password-protected pserver access even
when the password is not taken from ~/.cvspass file.
* "git fast-export" learned --no-data option that can be useful when
reordering commits and trees without touching the contents of
blobs.
* "git fast-import" has a pair of new front-end in contrib/ area.
* "git init" learned to mkdir/chdir into a directory when given an
extra argument (i.e. "git init this").
* "git instaweb" optionally can use mongoose as the web server.
* "git log --decorate" can optionally be told with --decorate=full to
give the reference name in full.
* "git merge" issued an unnecessarily scary message when it detected
that the merge may have to touch the path that the user has local
uncommitted changes to. The message has been reworded to make it
clear that the command aborted, without doing any harm.
* "git push" can be told to be --quiet.
* "git push" pays attention to url.$base.pushInsteadOf and uses a URL
that is derived from the URL used for fetching.
* informational output from "git reset" that lists the locally modified
paths is made consistent with that of "git checkout $another_branch".
* "git submodule" learned to give submodule name to scripts run with
"foreach" subcommand.
* various subcommands to "git submodule" learned --recursive option.
* "git submodule summary" learned --files option to compare the work
tree vs the commit bound at submodule path, instead of comparing
the index.
* "git upload-pack", which is the server side support for "git clone" and
"git fetch", can call a new post-upload-pack hook for statistics purposes.
(developers)
* With GIT_TEST_OPTS="--root=/p/a/t/h", tests can be run outside the
source directory; using tmpfs may give faster turnaround.
* With NO_PERL_MAKEMAKER set, DESTDIR= is now honoured, so you can
build for one location, and install into another location to tar it
up.
Fixes since v1.6.4
------------------
All of the fixes in v1.6.4.X maintenance series are included in this
release, unless otherwise noted.
Major changes:
* git-svn updates, including a new --authors-prog option to map author
names by invoking an external program, 'git svn reset' to unwind
'git svn fetch', support for more than one branches, documenting
of the useful --minimize-url feature, new "git svn gc" command, etc.
* We feed iconv with "UTF-8" instead of "utf8"; the former is
understood more widely. Similarly updated test scripts to use
encoding names more widely understood (e.g. use "ISO8859-1" instead
of "ISO-8859-1").
* Various portability fixes/workarounds for different vintages of
SunOS, IRIX, and Windows.
* Git-over-ssh transport on Windows supports PuTTY plink and TortoisePlink.
* Many repeated use of lstat() are optimized out in "checkout" codepath.
* git-status (and underlying git-diff-index --cached) are optimized
to take advantage of cache-tree information in the index.
This version fixes the remote DoS problem in
http://secunia.com/advisories/35437/.
Major changes between 1.6.2 and 1.6.3:
* various git-svn updates.
* git-gui updates, including an update to Russian translation, and a
fix to an infinite loop when showing an empty diff.
* gitk updates, including an update to Russian translation and
improved Windows support.
* many uses of lstat(2) in the codepath for "git checkout" have been
optimized out.
* usuability improvements.
* bug fixes.
Major changes in 1.6.1:
When some commands (e.g. "git log", "git diff") spawn pager
internally, we used to make the pager the parent process of the git
command that produces output. This meant that the exit status of the
whole thing comes from the pager, not the underlying git command. We
swapped the order of the processes around and you will see the exit
code from the command from now on.
Major changes in 1.6.2:
With the next major release, "git push" into a branch that is
currently checked out will be refused by default. You can choose what
should happen upon such a push by setting the configuration variable
receive.denyCurrentBranch in the receiving repository.
To ease the transition plan, the receiving repository of such a push
running this release will issue a big warning when the configuration
variable is missing.
For a similar reason, "git push $there :$killed" to delete the branch
$killed in a remote repository $there, if $killed branch is the
current branch pointed at by its HEAD, gets a large warning. You can
choose what should happen upon such a push by setting the
configuration variable receive.denyDeleteCurrent in the receiving
repository.
security issues and to prepare for the outstanding upgrade of gitweb.
Changes since git 1.6.0.2:
- Fix output line format for zip archive generation.
- Fix continuation of git rebase -i in case of modified files (conflict).
- Honor the pre-rebase hook for git rebase -i.
- Fix inconsistent behavior of git diff --quiet and diff --exit-code.
- Use multiple element hunk headers from git diff correctly.
- Portability fix for the git diff hunk header regexps.
- Fix git index-pack inside .git/objects/pack directory.
- Fix git index-pack in thin-pack mode.
- Some git stash apply fixes.
- Fixed format string vulnerabilities (e.g. in git remote).
- Fix behavior of git checkout -b <existingbranch>.
- Make git checkout -q actually quiet.
- In quiet remote operation, we don't need the remote side to send along
all the statistics we discard.
- Don't do cross-directory renames when creating packs.
- git prune now also removes stale temp files under .git/objects/pack.
- Have git merge prune empty directories.
- Have git merge allocate enough memory for the structure itself when
enumerating the parents of the resulting commit.
- Stop git blame -C -C from segfaulting if it encounters a submodule
reference.
- If only file times were changed, don't have git rm claim local
modifications.
- If set-tree fails, don't let git svn output Perl compile errors.
- Fix various NULL pointer crashes (e.g. in apply, reset, update-index).
- Remove bash completion support for various non-space tools.
- Don't have git add -p claim no changes if binary files have been
modified.
- Fix git archive in bare repositories.
- Have git diff display the number of new or deleted files for the case
where there have been too many of them to allow rename detection.
- Fix git push --mirror to not try to send the stash.
- If the remote end is up to date, still update the tracking reference upon
git push.
- Make git pull $there $branch:$current_branch work on unborn branches.
- Have git reset --hard remove new paths after giving up resolving a
conflicting merge.
- Fix git send-email fd leak in directory scanning.
- Make submodule directories appear as tracked in git status.
- Have git svn invoke "git command" rather than "git-command".
- Have git update-ref -d honor --no-deref.
- Fix various memory leaks.
- Fix git checkout segfault when HEAD points to a deleted branch.
- Allow switching out of the current branch with git checkout on an
un-checked-out state.
- Have git fast-export export all tags.
- Make git ls-files --with-tree=<tree> work with options other than -c.
- If the first object in git pack-objects exceeds --max-pack-size, don't
stuff even more objects in.
- Stop git-p4 from replacing multilike keywords. (They don't exist.)
- Make git tag complain about mutually incompatible options.
- Fix performance for git internal diff on large files with repeating
chunks.
- Don't let git repack grab objects out of packs marked with .keep into
new packs.
- Fix git fsck deep recursion.
- Fix git fast-export and fast-import on old style annotated tags without
tagger information.
- Have git mergetool honor the "--".
- Fix segfault of git show <tag> where <tag> points at a nonexistent object.
- Fix handling of lists of mail addresses for git send-email.
- Fix branch ancestry logic in git svn if the history fetching process
was interrupted.
change ${.CURDIR}/../... to ../../devel/scmgit/...; makes
a lot more sense.
Fixes since v1.6.0.1
--------------------
* Installation on platforms that needs .exe suffix to git-* programs were
broken in 1.6.0.1.
* Installation on filesystems without symbolic links support did nto
work well.
* In-tree documentations and test scripts now use "git foo" form to set a
better example, instead of the "git-foo" form (which is an acceptable
form if you have "PATH=$(git --exec-path):$PATH" in your script)
* Many commands did not use the correct working tree location when used
with GIT_WORK_TREE environment settings.
* Some systems needs to use compatibility fnmach and regex libraries
independent from each other; the compat/ area has been reorganized to
allow this.
* "git apply --unidiff-zero" incorrectly applied a -U0 patch that inserts
a new line before the second line.
* "git blame -c" did not exactly work like "git annotate" when range
boundaries are involved.
* "git checkout file" when file is still unmerged checked out contents from
a random high order stage, which was confusing.
* "git clone $there $here/" with extra trailing slashes after explicit
local directory name $here did not work as expected.
* "git diff" on tracked contents with CRLF line endings did not drive "less"
intelligently when showing added or removed lines.
* "git diff --dirstat -M" did not add changes in subdirectories up
correctly for renamed paths.
* "git diff --cumulative" did not imply "--dirstat".
* "git for-each-ref refs/heads/" did not work as expected.
* "git gui" allowed users to feed patch without any context to be applied.
* "git gui" botched parsing "diff" output when a line that begins with two
dashes and a space gets removed or a line that begins with two pluses
and a space gets added.
* "git gui" translation updates and i18n fixes.
* "git index-pack" is more careful against disk corruption while completing
a thin pack.
* "git log -i --grep=pattern" did not ignore case; neither "git log -E
--grep=pattern" triggered extended regexp.
* "git log --pretty="%ad" --date=short" did not use short format when
showing the timestamp.
* "git log --author=author" match incorrectly matched with the
timestamp part of "author " line in commit objects.
* "git log -F --author=author" did not work at all.
* Build procedure for "git shell" that used stub versions of some
functions and globals was not understood by linkers on some platforms.
* "git stash" was fooled by a stat-dirty but otherwise unmodified paths
and refused to work until the user refreshed the index.
* "git svn" was broken on Perl before 5.8 with recent fixes to reduce
use of temporary files.
* "git verify-pack -v" did not work correctly when given more than one
packfile.
Also contains many documentation updates.
asau
adam
cheusov
gdt
apb
wiz
tnn
minskim
morr
rillig
tonnerre
bjs