Documentation / RelNotes / 2.14.0.txton commit notes: don't access hash of NULL object_id pointer (3ea6b85)
   1Git 2.14 Release Notes
   2======================
   3
   4Backward compatibility notes.
   5
   6 * Use of an empty string as a pathspec element that is used for
   7   'everything matches' is still warned and Git asks users to use a
   8   more explicit '.' for that instead.  The hope is that existing
   9   users will not mind this change, and eventually the warning can be
  10   turned into a hard error, upgrading the deprecation into removal of
  11   this (mis)feature.  That is not scheduled to happen in the upcoming
  12   release (yet).
  13
  14 * Git now avoids blindly falling back to ".git" when the setup
  15   sequence said we are _not_ in Git repository.  A corner case that
  16   happens to work right now may be broken by a call to die("BUG").
  17   We've tried hard to locate such cases and fixed them, but there
  18   might still be cases that need to be addressed--bug reports are
  19   greatly appreciated.
  20
  21
  22Updates since v2.13
  23-------------------
  24
  25UI, Workflows & Features
  26
  27 * The colors in which "git status --short --branch" showed the names
  28   of the current branch and its remote-tracking branch are now
  29   configurable.
  30
  31 * "git clone" learned the "--no-tags" option not to fetch all tags
  32   initially, and also set up the tagopt not to follow any tags in
  33   subsequent fetches.
  34
  35 * "git archive --format=zip" learned to use zip64 extension when
  36   necessary to go beyond the 4GB limit.
  37   (merge 867e40ff3a rs/large-zip later to maint).
  38
  39 * "git reset" learned "--recurse-submodules" option.
  40
  41 * "git diff --submodule=diff" now recurses into nested submodules.
  42   (merge 5a5221427c jk/diff-submodule-diff-inline later to maint).
  43
  44 * "git repack" learned to accept the --threads=<n> option and pass it
  45   to pack-objects.
  46
  47 * "git send-email" learned to run sendemail-validate hook to inspect
  48   and reject a message before sending it out.
  49   (merge 6489660b4b jt/send-email-validate-hook later to maint).
  50
  51 * There is no good reason why "git fetch $there $sha1" should fail
  52   when the $sha1 names an object at the tip of an advertised ref,
  53   even when the other side hasn't enabled allowTipSHA1InWant.
  54
  55 * The recently introduced "[includeIf "gitdir:$dir"] path=..."
  56   mechansim has further been taught to take symlinks into account.
  57   The directory "$dir" specified in "gitdir:$dir" may be a symlink to
  58   a real location, not something that $(getcwd) may return.  In such
  59   a case, a realpath of "$dir" is compared with the real path of the
  60   current repository to determine if the contents from the named path
  61   should be included.
  62
  63
  64Performance, Internal Implementation, Development Support etc.
  65
  66 * The default packed-git limit value has been raised on larger
  67   platforms to save "git fetch" from a (recoverable) failure while
  68   "gc" is running in parallel.
  69
  70 * Code to update the cache-tree has been tightened so that we won't
  71   accidentally write out any 0{40} entry in the tree object.
  72   (merge a96d3cc3f6 jk/no-null-sha1-in-cache-tree later to maint).
  73
  74 * Attempt to allow us notice "fishy" situation where we fail to
  75   remove the temporary directory used during the test.
  76
  77 * Travis CI gained a task to format the documentation with both
  78   AsciiDoc and AsciiDoctor.
  79   (merge 505ad91304 ls/travis-doc-asciidoctor later to maint).
  80
  81 * Some platforms have ulong that is smaller than time_t, and our
  82   historical use of ulong for timestamp would mean they cannot
  83   represent some timestamp that the platform allows.  Invent a
  84   separate and dedicated timestamp_t (so that we can distingiuish
  85   timestamps and a vanilla ulongs, which along is already a good
  86   move), and then declare uintmax_t is the type to be used as the
  87   timestamp_t.
  88
  89 * We can trigger Windows auto-build tester (credits: Dscho &
  90   Microsoft) from our existing Travis CI tester now.
  91
  92 * Conversion from uchar[20] to struct object_id continues.
  93
  94 * Simplify parse_pathspec() codepath and stop it from looking at the
  95   default in-core index.
  96   (merge 08de9151a8 bw/pathspec-sans-the-index later to maint).
  97
  98 * Add perf-test for wildmatch.
  99   (merge 62ca75a6b9 ab/perf-wildmatch later to maint).
 100
 101 * Code from "conversion using external process" codepath has been
 102   extracted to a separate sub-process.[ch] module.
 103   (merge 4f2a2e9f0e bp/sub-process-convert-filter later to maint).
 104
 105 * When "git checkout", "git merge", etc. manipulates the in-core
 106   index, various pieces of information in the index extensions are
 107   discarded from the original state, as it is usually not the case
 108   that they are kept up-to-date and in-sync with the operation on the
 109   main index.  The untracked cache extension is copied across these
 110   operations now, which would speed up "git status" (as long as the
 111   cache is properly invalidated).
 112
 113
 114Also contains various documentation updates and code clean-ups.
 115
 116
 117Fixes since v2.13
 118-----------------
 119
 120Unless otherwise noted, all the fixes since v2.13 in the maintenance
 121track are contained in this release (see the maintenance releases'
 122notes for details).
 123
 124 * "git gc" did not interact well with "git worktree"-managed
 125   per-worktree refs.
 126
 127 * "git cherry-pick" and other uses of the sequencer machinery
 128   mishandled a trailer block whose last line is an incomplete line.
 129   This has been fixed so that an additional sign-off etc. are added
 130   after completing the existing incomplete line.
 131   (merge 44dc738a39 jt/use-trailer-api-in-commands later to maint).
 132
 133 * The codepath in "git am" that is used when running "git rebase"
 134   leaked memory held for the log message of the commits being rebased.
 135   (merge 721f5f1e35 jk/am-leakfix later to maint).
 136
 137 * "git clone --config var=val" is a way to populate the
 138   per-repository configuration file of the new repository, but it did
 139   not work well when val is an empty string.  This has been fixed.
 140   (merge db4eca1fea jn/clone-add-empty-config-from-command-line later to maint).
 141
 142 * Setting "log.decorate=false" in the configuration file did not take
 143   effect in v2.13, which has been corrected.
 144   (merge c74271aae7 ah/log-decorate-default-to-auto later to maint).
 145
 146 * A few codepaths in "checkout" and "am" working on an unborn branch
 147   tried to access an uninitialized piece of memory.
 148   (merge 57e0ef0e0e rs/checkout-am-fix-unborn later to maint).
 149
 150 * The Web interface to gmane news archive is long gone, even though
 151   the articles are still accessible via NTTP.  Replace the links with
 152   ones to public-inbox.org.  Because their message identification is
 153   based on the actual message-id, it is likely that it will be easier
 154   to migrate away from it if/when necessary.
 155   (merge 5840eb9d14 ab/doc-replace-gmane-links later to maint).
 156
 157 * The receive-pack program now makes sure that the push certificate
 158   records the same set of push options used for pushing.
 159   (merge cbaf82cc6b jt/push-options-doc later to maint).
 160
 161 * Tests have been updated to pass under GETTEXT_POISON (a mechanism
 162   to ensure that output strings that should not be translated are
 163   not translated by mistake), and TravisCI is told to run them.
 164   (merge b8e188f6f5 ab/fix-poison-tests later to maint).
 165
 166 * "git checkout --recurse-submodules" did not quite work with a
 167   submodule that itself has submodules.
 168   (merge 218c883783 sb/checkout-recurse-submodules later to maint).
 169
 170 * Plug some leaks and updates internal API used to implement the
 171   split index feature to make it easier to avoid such a leak in the
 172   future.
 173   (merge de950c5773 nd/split-index-unshare later to maint).
 174
 175 * "pack-objects" can stream a slice of an existing packfile out when
 176   the pack bitmap can tell that the reachable objects are all needed
 177   in the output, without inspecting individual objects.  This
 178   strategy however would not work well when "--local" and other
 179   options are in use, and need to be disabled.
 180   (merge da5a1f8100 jk/disable-pack-reuse-when-broken later to maint).
 181
 182 * Fix memory leaks pointed out by Coverity (and people).
 183   (merge 443a12f37b js/plug-leaks later to maint).
 184
 185 * "git read-tree -m" (no tree-ish) gave a nonsense suggestion "use
 186   --empty if you want to clear the index".  With "-m", such a request
 187   will still fail anyway, as you'd need to name at least one tree-ish
 188   to be merged.
 189   (merge b9b10d3681 jc/read-tree-empty-with-m later to maint).
 190
 191 * Make sure our tests would pass when the sources are checked out
 192   with "platform native" line ending convention by default on
 193   Windows.  Some "text" files out tests use and the test scripts
 194   themselves that are meant to be run with /bin/sh, ought to be
 195   checked out with eol=LF even on Windows.
 196   (merge 2779f66505 js/eol-on-ourselves later to maint).
 197
 198 * Introduce the BUG() macro to improve die("BUG: ...").
 199   (merge 3d7dd2d3b6 jk/bug-to-abort later to maint).
 200
 201 * Clarify documentation for include.path and includeIf.<condition>.path
 202   configuration variables.
 203   (merge ce933ebd5a jk/doc-config-include later to maint).
 204
 205 * Git sometimes gives an advice in a rhetorical question that does
 206   not require an answer, which can confuse new users and non native
 207   speakers.  Attempt to rephrase them.
 208   (merge 6963893943 ja/do-not-ask-needless-questions later to maint).
 209
 210 * A few http:// links that are redirected to https:// in the
 211   documentation have been updated to https:// links.
 212   (merge 5e68729fd9 jk/update-links-in-docs later to maint).
 213
 214 * "git for-each-ref --format=..." with %(HEAD) in the format used to
 215   resolve the HEAD symref as many times as it had processed refs,
 216   which was wasteful, and "git branch" shared the same problem.
 217   (merge 613a0e52ea kn/ref-filter-branch-list later to maint).
 218
 219 * Regression fix to topic recently merged to 'master'.
 220   (merge d096d7f1ef pw/rebase-i-regression-fix later to maint).
 221
 222 * The shell completion script (in contrib/) learned "git stash" has
 223   a new "push" subcommand.
 224   (merge 3851e4483f tg/stash-push-fixup later to maint).
 225
 226 * "git interpret-trailers", when used as GIT_EDITOR for "git commit
 227   -v", looked for and appended to a trailer block at the very end,
 228   i.e. at the end of the "diff" output.  The command has been
 229   corrected to pay attention to the cut-mark line "commit -v" adds to
 230   the buffer---the real trailer block should appear just before it.
 231   (merge d76650b8d1 bm/interpret-trailers-cut-line-is-eom later to maint).
 232
 233 * A test allowed both "git push" and "git receive-pack" on the other
 234   end write their traces into the same file.  This is OK on platforms
 235   that allows atomically appending to a file opened with O_APPEND,
 236   but on other platforms led to a mangled output, causing
 237   intermittent test failures.  This has been fixed by disabling
 238   traces from "receive-pack" in the test.
 239   (merge 71406ed4d6 jk/alternate-ref-optim later to maint).
 240
 241 * Tag objects, which are not reachable from any ref, that point at
 242   missing objects were mishandled by "git gc" and friends (they
 243   should silently be ignored instead)
 244   (merge a3ba6bf10a jk/ignore-broken-tags-when-ignoring-missing-links later to maint).
 245
 246 * "git describe --contains" penalized light-weight tags so much that
 247   they were almost never considered.  Instead, give them about the
 248   same chance to be considered as an annotated tag that is the same
 249   age as the underlying commit would.
 250   (merge ef1e74065c jc/name-rev-lw-tag later to maint).
 251
 252 * The "run-command" API implementation has been made more robust
 253   against dead-locking in a threaded environment.
 254   (merge e3f43ce765 bw/forking-and-threading later to maint).
 255
 256 * A recent update to t5545-push-options.sh started skipping all the
 257   tests in the script when a web server testing is disabled or
 258   unavailable, not just the ones that require a web server.  Non HTTP
 259   tests have been salvaged to always run in this script.
 260   (merge 2e397e4ddf jc/skip-test-in-the-middle later to maint).
 261
 262 * "git send-email" now uses Net::SMTP::SSL, which is obsolete, only
 263   when needed.  Recent versions of Net::SMTP can do TLS natively.
 264   (merge 0ead000c3a dk/send-email-avoid-net-smtp-ssl-when-able later to maint).
 265
 266 * Other minor doc, test and build updates and code cleanups.
 267   (merge 515360f9e9 jn/credential-doc-on-clear later to maint).
 268   (merge 0e6d899fee ab/aix-needs-compat-regex later to maint).
 269   (merge e294e8959f jc/apply-fix-mismerge later to maint).
 270   (merge 7f1b225153 bw/submodule-with-bs-path later to maint).
 271   (merge c8f7c8b704 tb/dedup-crlf-tests later to maint).
 272   (merge 449456ad47 sg/core-filemode-doc-typofix later to maint).
 273   (merge ba4dce784e km/log-showsignature-doc later to maint).
 274   (merge c5a9157393 jh/memihash-opt later to maint).