Documentation / RelNotes / 1.8.4.txton commit Merge branch 'cm/remote-mediawiki-perlcritique' (39abbd3)
   1Git v1.8.4 Release Notes
   2========================
   3
   4Updates since v1.8.3
   5--------------------
   6
   7Foreign interfaces, subsystems and ports.
   8
   9 * Git-gui has been updated to its 0.18.0 version.
  10
  11 * MediaWiki remote helper (in contrib/) has been updated to use the
  12   credential helper interface from Git.pm.
  13
  14 * Update build for Cygwin 1.[57].  Torsten Bögershausen reports that
  15   this is fine with Cygwin 1.7 ($gmane/225824) so let's try moving it
  16   ahead.
  17
  18 * The credential helper to talk to keychain on OS X (in contrib/) has
  19   been updated to kick in not just when talking http/https but also
  20   imap(s) and smtp.
  21
  22 * Remote transport helper has been updated to report errors and
  23   maintain ref hierarchy used to keep track of its own state better.
  24
  25 * With "export" remote-helper protocol, (1) a push that tries to
  26   update a remote ref whose name is different from the pushing side
  27   does not work yet, and (2) the helper may not know how to do
  28   --dry-run; these problematic cases are disabled for now.
  29
  30 * git-remote-hg/bzr (in contrib/) updates.
  31
  32 * git-remote-mw (in contrib/) hints users to check the certificate,
  33   when https:// connection failed.
  34
  35
  36UI, Workflows & Features
  37
  38 * Many tutorials teach users to set "color.ui" to "auto" as the first
  39   thing after you set "user.name/email" to introduce yourselves to
  40   Git.  Now the variable defaults to "auto".
  41
  42 * "git cmd <name>", when <name> happens to be a 40-hex string,
  43   directly uses the 40-hex string as an object name, even if a ref
  44   "refs/<some hierarchy>/<name>" exists.  This disambiguation order
  45   is unlikely to change, but we should warn about the ambiguity just
  46   like we warn when more than one refs/ hierachies share the same
  47   name.
  48
  49 * "git rebase" learned "--[no-]autostash" option to save local
  50   changes instead of refusing to run (to which people's normal
  51   response was to stash them and re-run).
  52
  53 * Instead of typing four capital letters "HEAD", you can say "@" now,
  54   e.g. "git log @".
  55
  56 * "check-ignore" (new feature since 1.8.2) has been updated to work
  57   more like "check-attr" over bidi-pipes.
  58
  59 * "git describe" learned "--first-parent" option to limit its closest
  60   tagged commit search to the first-parent chain.
  61
  62 * "git merge foo" that might have meant "git merge origin/foo" is
  63   diagnosed with a more informative error message.
  64
  65 * "git log -L<line>,<range>:<filename>" has been added.  This may
  66   still have leaks and rough edges, though.
  67
  68 * We used the approxidate() parser for "--expire=<timestamp>" options
  69   of various commands, but it is better to treat --expire=all and
  70   --expire=now a bit more specially than using the current timestamp.
  71   "git gc" and "git reflog" have been updated with a new parsing
  72   function for expiry dates.
  73
  74 * Updates to completion (both bash and zsh) helpers.
  75
  76 * The behaviour of the "--chain-reply-to" option of "git send-email"
  77   have changed at 1.7.0, and we added a warning/advice message to
  78   help users adjust to the new behaviour back then, but we kept it
  79   around for too long.  The message has finally been removed.
  80
  81 * "git fetch origin master" unlike "git fetch origin" or "git fetch"
  82   did not update "refs/remotes/origin/master"; this was an early
  83   design decision to keep the update of remote tracking branches
  84   predictable, but in practice it turns out that people find it more
  85   convenient to opportunistically update them whenever we have a
  86   chance, and we have been updating them when we run "git push" which
  87   already breaks the original "predictability" anyway.
  88
  89 * The configuration variable core.checkstat was advertised in the
  90   documentation but the code expected core.statinfo instead.
  91   For now, we accept both core.checkstat and core.statinfo, but the
  92   latter will be removed in the longer term.
  93
  94
  95Performance, Internal Implementation, etc.
  96
  97 * Uses of the platform fnmatch(3) function (many places in the code,
  98   matching pathspec, .gitignore and .gitattributes to name a few)
  99   have been replaced with wildmatch, allowing "foo/**/bar" that would
 100   match foo/bar, foo/a/bar, foo/a/b/bar, etc.
 101
 102 * Memory ownership and lifetime rules for what for-each-ref feeds to
 103   its callbacks have been clarified (in short, "you do not own it, so
 104   make a copy if you want to keep it").
 105
 106 * The revision traversal logic to improve culling of irrelevant
 107   parents while traversing a mergy history has been updated.
 108
 109 * Some leaks in unpack-trees (used in merge, cherry-pick and other
 110   codepaths) have been plugged.
 111
 112 * The codepath to read from marks files in fast-import/export did not
 113   have to accept anything but 40-hex representation of the object
 114   name.  Further, fast-export did not need full in-core object
 115   representation to have parsed wen reading from them.  These
 116   codepaths have been optimized by taking advantage of these access
 117   patterns.
 118
 119 * Object lookup logic, when the object hashtable starts to become
 120   crowded, has been optimized.
 121
 122 * When TEST_OUTPUT_DIRECTORY setting is used, it was handled somewhat
 123   inconsistently between the test framework and t/Makefile, and logic
 124   to summarize the results looked at a wrong place.
 125
 126 * "git clone" uses a lighter-weight implementation when making sure
 127   that the history behind refs are complete.
 128
 129 * Many warnings from sparse source checker in compat/ area has been
 130   squelched.
 131
 132 * The code to reading and updating packed-refs file has been updated,
 133   correcting corner case bugs.
 134
 135
 136Also contains various documentation updates and code clean-ups.
 137
 138
 139Fixes since v1.8.3
 140------------------
 141
 142Unless otherwise noted, all the fixes since v1.8.3 in the maintenance
 143track are contained in this release (see release notes to them for
 144details).
 145
 146 * Logic used by git-send-email to suppress cc mishandled names like
 147   "A U. Thor" <author@example.xz>, where the human readable part
 148   needs to be quoted (the user input may not have the double quotes
 149   around the name, and comparison was done between quoted and
 150   unquoted strings).
 151   (merge 1495266 mt/send-email-cc-match-fix later to maint).
 152
 153 * Call to discard_cache/discard_index (used when we use different
 154   contents of the index in-core, in many operations like commit,
 155   apply, and merge) used to leak memory that held the array of index
 156   entries, which has been plugged.
 157   (merge a0fc4db rs/discard-index-discard-array later to maint).
 158
 159 * "gitweb" forgot to clear a global variable $search_regexp upon each
 160   request, mistakenly carrying over the previous search to a new one
 161   when used as a persistent CGI.
 162   (merge ca7a5dc cm/gitweb-project-list-persistent-cgi-fix later to maint).
 163
 164 * The wildmatch engine did not honor WM_CASEFOLD option correctly.
 165   (merge b79c0c3 ar/wildmatch-foldcase later to maint).
 166
 167 * "git log -c --follow $path" segfaulted upon hitting the commit that
 168   renamed the $path being followed.
 169   (merge 46ec510 cb/log-follow-with-combined later to maint).
 170
 171 * When a reflog notation is used for implicit "current branch", we
 172   did not say which branch and worse said "branch ''".
 173   (merge 305ebea rr/die-on-missing-upstream later to maint).
 174
 175 * "difftool --dir-diff" did not copy back changes made by the
 176   end-user in the diff tool backend to the working tree in some
 177   cases.
 178   (merge 32eaf1d ks/difftool-dir-diff-copy-fix later to maint).
 179
 180 * "git push $there HEAD:branch" did not resolve HEAD early enough, so
 181   it was easy to flip it around while push is still going on and push
 182   out a branch that the user did not originally intended when the
 183   command was started.
 184   (merge 0f075b2 rr/push-head later to maint).
 185
 186 * The bash prompt code (in contrib/) displayed the name of the branch
 187   being rebased when "rebase -i/-m/-p" modes are in use, but not the
 188   plain vanilla "rebase".
 189   (merge 1306321 fc/show-branch-in-rebase-am later to maint).
 190
 191 * Handling of negative exclude pattern for directories "!dir" was
 192   broken in the update to v1.8.3.
 193   (merge c3c327d kb/status-ignored-optim-2 later to maint).
 194
 195 * zsh prompt script that borrowed from bash prompt script did not
 196   work due to slight differences in array variable notation between
 197   these two shells.
 198   (merge d0583da tg/maint-zsh-svn-remote-prompt later to maint).
 199
 200 * An entry for "file://" scheme in the enumeration of URL types Git
 201   can take in the HTML documentation was made into a clickable link
 202   by mistake.
 203   (merge 4c32e36 nd/urls-doc-no-file-hyperlink-fix later to maint).
 204
 205 * "git push --[no-]verify" was not documented.
 206   (merge 90d32d1 tr/push-no-verify-doc later to maint).
 207
 208 * Stop installing the git-remote-testpy script that is only used for
 209   testing.
 210   (merge 416fda6 fc/makefile later to maint).
 211
 212 * "git commit --allow-empty-message -m ''" should not start an
 213   editor.
 214   (merge 2520677 rs/commit-m-no-edit later to maint).
 215
 216 * "git merge @{-1}~22" was rewritten to "git merge frotz@{1}~22"
 217   incorrectly when your previous branch was "frotz" (it should be
 218   rewritten to "git merge frotz~22" instead).
 219   (merge 84cf246 jc/strbuf-branchname-fix later to maint).
 220
 221 * "git diff -c -p" was not showing a deleted line from a hunk when
 222   another hunk immediately begins where the earlier one ends.
 223   (merge aac3857 mk/combine-diff-context-horizon-fix later to maint).
 224
 225 * "git log --ancestry-path A...B" did not work as expected, as it did
 226   not pay attention to the fact that the merge base between A and B
 227   was the bottom of the range being specified.
 228   (merge a765499 kb/ancestry-path-threedots later to maint).
 229
 230 * Mac OS X does not like to write(2) more than INT_MAX number of
 231   bytes; work it around by chopping write(2) into smaller pieces.
 232   (merge 6c642a8 fc/macos-x-clipped-write later to maint).
 233
 234 * Newer MacOS X encourages the programs to compile and link with
 235   their CommonCrypto, not with OpenSSL.
 236   (merge be4c828 da/darwin later to maint).
 237
 238 * "git clone foo/bar:baz" cannot be a request to clone from a remote
 239   over git-over-ssh specified in the scp style.  This case is now
 240   detected and clones from a local repository at "foo/bar:baz".
 241   (merge 6000334 nd/clone-local-with-colon later to maint).
 242
 243 * When $HOME is misconfigured to point at an unreadable directory, we
 244   used to complain and die. Loosen the check.
 245   (merge 4698c8f jn/config-ignore-inaccessible later to maint).
 246
 247 * "git subtree" (in contrib/) had one codepath with loose error
 248   checks to lose data at the remote side.
 249   (merge 3212d56 jk/subtree-do-not-push-if-split-fails later to maint).
 250
 251 * "git fetch" into a shallow repository from a repository that does
 252   not know about the shallow boundary commits (e.g. a different fork
 253   from the repository the current shallow repository was cloned from)
 254   did not work correctly.
 255   (merge 71d5f93 mh/fetch-into-shallow later to maint).
 256
 257 * "git checkout foo" DWIMs the intended "upstream" and turns it into
 258   "git checkout -t -b foo remotes/origin/foo". This codepath has been
 259   updated to correctly take existing remote definitions into account.
 260   (merge 229177a jh/checkout-auto-tracking later to maint).