Git 2.14 Release Notes
======================

Backward compatibility notes and other notable changes.

* Use of an empty string as a pathspec element that is used for
  'everything matches' is still warned and Git asks users to use a
  more explicit '.' for that instead.  The hope is that existing
  users will not mind this change, and eventually the warning can be
  turned into a hard error, upgrading the deprecation into removal of
  this (mis)feature.  That is not scheduled to happen in the upcoming
  release (yet).

* Git now avoids blindly falling back to ".git" when the setup
  sequence said we are _not_ in Git repository.  A corner case that
  happens to work right now may be broken by a call to die("BUG").
  We've tried hard to locate such cases and fixed them, but there
  might still be cases that need to be addressed--bug reports are
  greatly appreciated.

* The experiment to improve the hunk-boundary selection of textual
  diff output has finished, and the "indent heuristics" has now
  become the default.

* Git can now be built with PCRE v2 instead of v1 of the PCRE
  library. Replace USE_LIBPCRE=YesPlease with USE_LIBPCRE2=YesPlease
  in existing build scripts to build against the new version.  As the
  upstream PCRE maintainer has abandoned v1 maintenance for all but
  the most critical bug fixes, use of v2 is recommended.

Updates since v2.13
-------------------

UI, Workflows & Features

* The colors in which "git status --short --branch" showed the names
  of the current branch and its remote-tracking branch are now
  configurable.

* "git clone" learned the "--no-tags" option not to fetch all tags
  initially, and also set up the tagopt not to follow any tags in
  subsequent fetches.

* "git archive --format=zip" learned to use zip64 extension when
  necessary to go beyond the 4GB limit.

* "git reset" learned "--recurse-submodules" option.

* "git diff --submodule=diff" now recurses into nested submodules.

* "git repack" learned to accept the --threads=<n> option and pass it
  to pack-objects.

* "git send-email" learned to run sendemail-validate hook to inspect
  and reject a message before sending it out.

* There is no good reason why "git fetch $there $sha1" should fail
  when the $sha1 names an object at the tip of an advertised ref,
  even when the other side hasn't enabled allowTipSHA1InWant.

* The "[includeIf "gitdir:$dir"] path=..." mechanism introduced in
  2.13.0 would canonicalize the path of the gitdir being matched,
  and did not match e.g. "gitdir:~/work/*" against a repo in
  "~/work/main" if "~/work" was a symlink to "/mnt/storage/work".
  Now we match both the resolved canonical path and what "pwd" would
  show. The include will happen if either one matches.

* The "indent" heuristics is now the default in "diff". The
  diff.indentHeuristic configuration variable can be set to "false"
  for those who do not want it.

* Many commands learned to pay attention to submodule.recurse
  configuration.

* The convention for a command line is to follow "git cmdname
  --options" with revisions followed by an optional "--"
  disambiguator and then finally pathspecs.  When "--" is not there,
  we make sure early ones are all interpretable as revs (and do not
  look like paths) and later ones are the other way around.  A
  pathspec with "magic" (e.g. ":/p/a/t/h" that matches p/a/t/h from
  the top-level of the working tree, no matter what subdirectory you
  are working from) are conservatively judged as "not a path", which
  required disambiguation more often.  The command line parser
  learned to say "it's a pathspec" a bit more often when the syntax
  looks like so.

* Update "perl-compatible regular expression" support to enable JIT
  and also allow linking with the newer PCRE v2 library.

* "filter-branch" learned a pseudo filter "--setup" that can be used
  to define common functions/variables that can be used by other
  filters.

* Using "git add d/i/r" when d/i/r is the top of the working tree of
  a separate repository would create a gitlink in the index, which
  would appear as a not-quite-initialized submodule to others.  We
  learned to give warnings when this happens.

* "git status" learned to optionally give how many stash entries there
  are in its output.

* "git status" has long shown essentially the same message as "git
  commit"; the message it gives while preparing for the root commit,
  i.e. "Initial commit", was hard to understand for some new users.
  Now it says "No commits yet" to stress more on the current status
  (rather than the commit the user is preparing for, which is more in
  line with the focus of "git commit").

* "git send-email" now has --batch-size and --relogin-delay options
    which can be used to overcome limitations on SMTP servers that
    restrict on how many of e-mails can be sent in a single session.

* An old message shown in the commit log template was removed, as it
  has outlived its usefulness.

* "git pull --rebase --recurse-submodules" learns to rebase the
  branch in the submodules to an updated base.

* "git log" learned -P as a synonym for --perl-regexp, "git grep"
  already had such a synonym.

* "git log" didn't understand --regexp-ignore-case when combined with
  --perl-regexp. This has been fixed.

Performance, Internal Implementation, Development Support etc.

* The default packed-git limit value has been raised on larger
  platforms to save "git fetch" from a (recoverable) failure while
  "gc" is running in parallel.

* Code to update the cache-tree has been tightened so that we won't
  accidentally write out any 0{40} entry in the tree object.

* Attempt to allow us notice "fishy" situation where we fail to
  remove the temporary directory used during the test.

* Travis CI gained a task to format the documentation with both
  AsciiDoc and AsciiDoctor.

* Some platforms have ulong that is smaller than time_t, and our
  historical use of ulong for timestamp would mean they cannot
  represent some timestamp that the platform allows.  Invent a
  separate and dedicated timestamp_t (so that we can distingiuish
  timestamps and a vanilla ulongs, which along is already a good
  move), and then declare uintmax_t is the type to be used as the
  timestamp_t.

* We can trigger Windows auto-build tester (credits: Dscho &
  Microsoft) from our existing Travis CI tester now.

* Conversion from uchar[20] to struct object_id continues.

* Simplify parse_pathspec() codepath and stop it from looking at the
  default in-core index.

* Add perf-test for wildmatch.

* Code from "conversion using external process" codepath has been
  extracted to a separate sub-process.[ch] module.

* When "git checkout", "git merge", etc. manipulates the in-core
  index, various pieces of information in the index extensions are
  discarded from the original state, as it is usually not the case
  that they are kept up-to-date and in-sync with the operation on the
  main index.  The untracked cache extension is copied across these
  operations now, which would speed up "git status" (as long as the
  cache is properly invalidated).

* The internal implementation of "git grep" has seen some clean-up.

* Update the C style recommendation for notes for translators, as
  recent versions of gettext tools can work with our style of
  multi-line comments.

* The implementation of "ref" API around the "packed refs" have been
  cleaned up, in preparation for further changes.

* The internal logic used in "git blame" has been libified to make it
  easier to use by cgit.

* Our code often opens a path to an optional file, to work on its
  contents when we can successfully open it.  We can ignore a failure
  to open if such an optional file does not exist, but we do want to
  report a failure in opening for other reasons (e.g. we got an I/O
  error, or the file is there, but we lack the permission to open).

  The exact errors we need to ignore are ENOENT (obviously) and
  ENOTDIR (less obvious).  Instead of repeating comparison of errno
  with these two constants, introduce a helper function to do so.

* We often try to open a file for reading whose existence is
  optional, and silently ignore errors from open/fopen; report such
  errors if they are not due to missing files.

* When an existing repository is used for t/perf testing, we first
  create bit-for-bit copy of it, which may grab a transient state of
  the repository and freeze it into the repository used for testing,
  which then may cause Git operations to fail.  Single out "the index
  being locked" case and forcibly drop the lock from the copy.

* Three instances of the same helper function have been consolidated
  to one.

* "fast-import" uses a default pack chain depth that is consistent
  with other parts of the system.

* A new test to show the interaction between the pattern [^a-z]
  (which matches '/') and a slash in a path has been added.  The
  pattern should not match the slash with "pathmatch", but should
  with "wildmatch".

* The 'diff-highlight' program (in contrib/) has been restructured
  for easier reuse by an external project 'diff-so-fancy'.

* A common pattern to free a piece of memory and assign NULL to the
  pointer that used to point at it has been replaced with a new
  FREE_AND_NULL() macro.

* Traditionally, the default die() routine had a code to prevent it
  from getting called multiple times, which interacted badly when a
  threaded program used it (one downside is that the real error may
  be hidden and instead the only error message given to the user may
  end up being "die recursion detected", which is not very useful).

* Introduce a "repository" object to eventually make it easier to
  work in multiple repositories (the primary focus is to work with
  the superproject and its submodules) in a single process.

* Optimize "what are the object names already taken in an alternate
  object database?" query that is used to derive the length of prefix
  an object name is uniquely abbreviated to.

* The hashmap API has been updated so that data to customize the
  behaviour of the comparison function can be specified at the time a
  hashmap is initialized.

* The "collision detecting" SHA-1 implementation shipped with 2.13 is
  now integrated into git.git as a submodule (the first submodule to
  ship with git.git). Clone git.git with --recurse-submodules to get
  it. For now a non-submodule copy of the same code is also shipped
  as part of the tree.

* A recent update made it easier to use "-fsanitize=" option while
  compiling but supported only one sanitize option.  Allow more than
  one to be combined, joined with a comma, like "make SANITIZE=foo,bar".

* Use "p4 -G" to make "p4 changes" output more Python-friendly
  to parse.

* We started using "%" PRItime, imitating "%" PRIuMAX and friends, as
  a way to format the internal timestamp value, but this does not
  play well with gettext(1) i18n framework, and causes "make pot"
  that is run by the l10n coordinator to create a broken po/git.pot
  file.  This is a possible workaround for that problem.

* It turns out that Cygwin also needs the fopen() wrapper that
  returns failure when a directory is opened for reading.

(adam)