Documentation/RelNotes/1.9.txt

   1 Git v1.9 Release Notes
   2 ======================
   3
   4 Backward compatibility notes
   5 ----------------------------
   6
   7 "git submodule foreach $cmd $args" used to treat "$cmd $args" the same
   8 way "ssh" did, concatenating them into a single string and letting the
   9 shell unquote. Careless users who forget to sufficiently quote $args
  10 gets their argument split at $IFS whitespaces by the shell, and got
  11 unexpected results due to this. Starting from this release, the
  12 command line is passed directly to the shell, if it has an argument.
  13
  14 Read-only support for experimental loose-object format, in which users
  15 could optionally choose to write in their loose objects for a short
  16 while between v1.4.3 to v1.5.3 era, has been dropped.
  17
  18 The meanings of "--tags" option to "git fetch" has changed; the
  19 command fetches tags _in addition to_ what are fetched by the same
  20 command line without the option.
  21
  22 The way "git push $there $what" interprets $what part given on the
  23 command line, when it does not have a colon that explicitly tells us
  24 what ref at the $there repository is to be updated, has been enhanced.
  25
  26 A handful of ancient commands that have long been deprecated are
  27 finally gone (repo-config, tar-tree, lost-found, and peek-remote).
  28
  29
  30 Backward compatibility notes (for Git 2.0)
  31 ------------------------------------------
  32
  33 When "git push [$there]" does not say what to push, we have used the
  34 traditional "matching" semantics so far (all your branches were sent
  35 to the remote as long as there already are branches of the same name
  36 over there).  In Git 2.0, the default will change to the "simple"
  37 semantics, which pushes:
  38
  39  - only the current branch to the branch with the same name, and only
  40    when the current branch is set to integrate with that remote
  41    branch, if you are pushing to the same remote as you fetch from; or
  42
  43  - only the current branch to the branch with the same name, if you
  44    are pushing to a remote that is not where you usually fetch from.
  45
  46 Use the user preference configuration variable "push.default" to
  47 change this.  If you are an old-timer who is used to the "matching"
  48 semantics, you can set the variable to "matching" to keep the
  49 traditional behaviour.  If you want to live in the future early, you
  50 can set it to "simple" today without waiting for Git 2.0.
  51
  52 When "git add -u" (and "git add -A") is run inside a subdirectory and
  53 does not specify which paths to add on the command line, it
  54 will operate on the entire tree in Git 2.0 for consistency
  55 with "git commit -a" and other commands.  There will be no
  56 mechanism to make plain "git add -u" behave like "git add -u .".
  57 Current users of "git add -u" (without a pathspec) should start
  58 training their fingers to explicitly say "git add -u ."
  59 before Git 2.0 comes.  A warning is issued when these commands are
  60 run without a pathspec and when you have local changes outside the
  61 current directory, because the behaviour in Git 2.0 will be different
  62 from today's version in such a situation.
  63
  64 In Git 2.0, "git add <path>" will behave as "git add -A <path>", so
  65 that "git add dir/" will notice paths you removed from the directory
  66 and record the removal.  Versions before Git 2.0, including this
  67 release, will keep ignoring removals, but the users who rely on this
  68 behaviour are encouraged to start using "git add --ignore-removal <path>"
  69 now before 2.0 is released.
  70
  71 The default prefix for "git svn" will change in Git 2.0.  For a long
  72 time, "git svn" created its remote-tracking branches directly under
  73 refs/remotes, but it will place them under refs/remotes/origin/ unless
  74 it is told otherwise with its --prefix option.
  75
  76
  77 Updates since v1.8.5
  78 --------------------
  79
  80 Foreign interfaces, subsystems and ports.
  81
  82  * The HTTP transport, when talking GSS-Negotiate, uses "100
  83    Continue" response to avoid having to rewind and resend a large
  84    payload, which may not be always doable.
  85
  86  * Various bugfixes to remote-bzr and remote-hg (in contrib/).
  87
  88  * The build procedure is aware of MirBSD now.
  89
  90
  91 UI, Workflows & Features
  92
  93  * Just like we give a reasonable default for "less" via the LESS
  94    environment variable, we now specify a reasonable default for "lv"
  95    via the "LV" environment variable when spawning the pager.
  96
  97  * Two-level configuration variable names in "branch.*" and "remote.*"
  98    hierarchies, whose variables are predominantly three-level, were
  99    not completed by hitting a <TAB> in bash and zsh completions.
 100
 101  * Fetching 'frotz' branch with "git fetch", while 'frotz/nitfol'
 102    remote-tracking branch from an earlier fetch was still there, would
 103    error out, primarily because the command was not told that it is
 104    allowed to lose any information on our side.  "git fetch --prune"
 105    now can be used to remove 'frotz/nitfol' to make room to fetch and
 106    store 'frotz' remote-tracking branch.
 107
 108  * "diff.orderfile=<file>" configuration variable can be used to
 109    pretend as if the "-O<file>" option were given from the command
 110    line of "git diff", etc.
 111
 112  * The negative pathspec syntax allows "git log -- . ':!dir'" to tell
 113    us "I am interested in everything but 'dir' directory".
 114
 115  * "git difftool" shows how many different paths there are in total,
 116    and how many of them have been shown so far, to indicate progress.
 117
 118  * "git push origin master" used to push our 'master' branch to update
 119    the 'master' branch at the 'origin' repository.  This has been
 120    enhanced to use the same ref mapping "git push origin" would use to
 121    determine what ref at the 'origin' to be updated with our 'master'.
 122    For example, with this configuration
 123
 124    [remote "origin"]
 125       push = refs/heads/*:refs/review/*
 126
 127    that would cause "git push origin" to push out our local branches
 128    to corresponding refs under refs/review/ hierarchy at 'origin',
 129    "git push origin master" would update 'refs/review/master' over
 130    there.  Alternatively, if push.default is set to 'upstream' and our
 131    'master' is set to integrate with 'topic' from the 'origin' branch,
 132    running "git push origin" while on our 'master' would update their
 133    'topic' branch, and running "git push origin master" while on any
 134    of our branches does the same.
 135
 136  * "gitweb" learned to treat ref hierarchies other than refs/heads as
 137    if they are additional branch namespaces (e.g. refs/changes/ in
 138    Gerrit).
 139
 140  * "git for-each-ref --format=..." learned a few formatting directives;
 141    e.g. "%(color:red)%(HEAD)%(color:reset) %(refname:short) %(subject)".
 142
 143  * The command string given to "git submodule foreach" is passed
 144    directly to the shell, without being eval'ed.  This is a backward
 145    incompatible change that may break existing users.
 146
 147  * "git log" and friends learned the "--exclude=<glob>" option, to
 148    allow people to say "list history of all branches except those that
 149    match this pattern" with "git log --exclude='*/*' --branches".
 150
 151  * "git rev-parse --parseopt" learned a new "--stuck-long" option to
 152    help scripts parse options with an optional parameter.
 153
 154  * The "--tags" option to "git fetch" no longer tells the command to
 155    fetch _only_ the tags. It instead fetches tags _in addition to_
 156    what are fetched by the same command line without the option.
 157
 158
 159 Performance, Internal Implementation, etc.
 160
 161  * When parsing a 40-hex string into the object name, the string is
 162    checked to see if it can be interpreted as a ref so that a warning
 163    can be given for ambiguity. The code kicked in even when the
 164    core.warnambiguousrefs is set to false to squelch this warning, in
 165    which case the cycles spent to look at the ref namespace were an
 166    expensive no-op, as the result was discarded without being used.
 167
 168  * The naming convention of the packfiles has been updated; it used to
 169    be based on the enumeration of names of the objects that are
 170    contained in the pack, but now it also depends on how the packed
 171    result is represented---packing the same set of objects using
 172    different settings (or delta order) would produce a pack with
 173    different name.
 174
 175  * "git diff --no-index" mode used to unnecessarily attempt to read
 176    the index when there is one.
 177
 178  * The deprecated parse-options macro OPT_BOOLEAN has been removed;
 179    use OPT_BOOL or OPT_COUNTUP in new code.
 180
 181  * A few duplicate implementations of prefix/suffix string comparison
 182    functions have been unified to starts_with() and ends_with().
 183
 184  * The new PERLLIB_EXTRA makefile variable can be used to specify
 185    additional directories Perl modules (e.g. the ones necessary to run
 186    git-svn) are installed on the platform when building.
 187
 188  * "git merge-base" learned the "--fork-point" mode, that implements
 189    the same logic used in "git pull --rebase" to find a suitable fork
 190    point out of the reflog entries for the remote-tracking branch the
 191    work has been based on.  "git rebase" has the same logic that can be
 192    triggered with the "--fork-point" option.
 193
 194  * A third-party "receive-pack" (the responder to "git push") can
 195    advertise the "no-thin" capability to tell "git push" not to use
 196    the thin-pack optimization. Our receive-pack has always been
 197    capable of accepting and fattening a thin-pack, and will continue
 198    not to ask "git push" to use a non-thin pack.
 199
 200
 201 Also contains various documentation updates and code clean-ups.
 202
 203
 204 Fixes since v1.8.5
 205 ------------------
 206
 207 Unless otherwise noted, all the fixes since v1.8.5 in the maintenance
 208 track are contained in this release (see the maintenance releases' notes
 209 for details).
 210
 211  * The implementation of 'git stash $cmd "stash@{...}"' did not quote
 212    the stash argument properly and left it split at IFS whitespace.
 213    (merge 2a07e43 ow/stash-with-ifs later to maint).
 214
 215  * The "--[no-]informative-errors" options to "git daemon" were parsed
 216    a bit too loosely, allowing any other string after these option
 217    names.
 218    (merge 82246b7 nd/daemon-informative-errors-typofix later to maint).
 219
 220  * There is no reason to have a hardcoded upper limit of the number of
 221    parents for an octopus merge, created via the graft mechanism, but
 222    there was.
 223    (merge e228c17 js/lift-parent-count-limit later to maint).
 224
 225  * The basic test used to leave unnecessary trash directories in the
 226    t/ directory.
 227    (merge 738a8be jk/test-framework-updates later to maint).
 228
 229  * "git merge-base --octopus" used to leave cleaning up suboptimal
 230    result to the caller, but now it does the clean-up itself.
 231    (merge 8f29299 bm/merge-base-octopus-dedup later to maint).
 232
 233  * A "gc" process running as a different user should be able to stop a
 234    new "gc" process from starting, but it didn't.
 235    (merge ed7eda8 km/gc-eperm later to maint).
 236
 237  * An earlier "clean-up" introduced an unnecessary memory leak.
 238    (merge e1c1a32 jk/credential-plug-leak later to maint).
 239
 240  * "git add -A" (no other arguments) in a totally empty working tree
 241    used to emit an error.
 242    (merge 64ed07c nd/add-empty-fix later to maint).
 243
 244  * "git log --decorate" did not handle a tag pointed by another tag
 245    nicely.
 246    (merge 5e1361c bc/log-decoration later to maint).
 247
 248  * When we figure out how many file descriptors to allocate for
 249    keeping packfiles open, a system with non-working getrlimit() could
 250    cause us to die(), but because we make this call only to get a
 251    rough estimate of how many is available and we do not even attempt
 252    to use up all file descriptors available ourselves, it is nicer to
 253    fall back to a reasonable low value rather than dying.
 254    (merge 491a8de jh/rlimit-nofile-fallback later to maint).
 255
 256  * read_sha1_file(), that is the workhorse to read the contents given
 257    an object name, honoured object replacements, but there was no
 258    corresponding mechanism to sha1_object_info() that was used to
 259    obtain the metainfo (e.g. type & size) about the object.  This led
 260    callers to weird inconsistencies.
 261    (merge 663a856 cc/replace-object-info later to maint).
 262
 263  * "git cat-file --batch=", an admittedly useless command, did not
 264    behave very well.
 265    (merge 6554dfa jk/cat-file-regression-fix later to maint).
 266
 267  * "git rev-parse <revs> -- <paths>" did not implement the usual
 268    disambiguation rules the commands in the "git log" family used in
 269    the same way.
 270    (merge 62f162f jk/rev-parse-double-dashes later to maint).
 271
 272  * "git mv A B/", when B does not exist as a directory, should error
 273    out, but it didn't.
 274    (merge c57f628 mm/mv-file-to-no-such-dir-with-slash later to maint).
 275
 276  * A workaround to an old bug in glibc prior to glibc 2.17 has been
 277    retired; this would remove a side effect of the workaround that
 278    corrupts system error messages in non-C locales.
 279
 280  * SSL-related options were not passed correctly to underlying socket
 281    layer in "git send-email".
 282    (merge 5508f3e tr/send-email-ssl later to maint).
 283
 284  * "git commit -v" appends the patch to the log message before
 285    editing, and then removes the patch when the editor returned
 286    control. However, the patch was not stripped correctly when the
 287    first modified path was a submodule.
 288    (merge 1a72cfd jl/commit-v-strip-marker later to maint).
 289
 290  * "git fetch --depth=0" was a no-op, and was silently ignored.
 291    Diagnose it as an error.
 292    (merge 5594bca nd/transport-positive-depth-only later to maint).
 293
 294  * Remote repository URL expressed in scp-style host:path notation are
 295    parsed more carefully (e.g. "foo/bar:baz" is local, "[::1]:/~user" asks
 296    to connect to user's home directory on host at address ::1.
 297    (merge a2036d7 tb/clone-ssh-with-colon-for-port later to maint).
 298
 299  * "git diff -- ':(icase)makefile'" was unnecessarily rejected at the
 300    command line parser.
 301    (merge 887c6c1 nd/magic-pathspec later to maint).
 302
 303  * "git cat-file --batch-check=ok" did not check the existence of
 304    the named object.
 305    (merge 4ef8d1d sb/sha1-loose-object-info-check-existence later to maint).
 306
 307  * "git am --abort" sometimes complained about not being able to write
 308    a tree with an 0{40} object in it.
 309    (merge 77b43ca jk/two-way-merge-corner-case-fix later to maint).
 310
 311  * Two processes creating loose objects at the same time could have
 312    failed unnecessarily when the name of their new objects started
 313    with the same byte value, due to a race condition.
 314    (merge b2476a6 jh/loose-object-dirs-creation-race later to maint).