Documentation / RelNotes / 2.2.0.txton commit Merge branch 'rs/ref-transaction' (3c85452)
   1Git v2.2 Release Notes
   2======================
   3
   4Updates since v2.1
   5------------------
   6
   7Ports
   8
   9 * Building on older MacOS X systems automatically sets
  10   the necessary NO_APPLE_COMMON_CRYPTO build-time option.
  11
  12
  13UI, Workflows & Features
  14
  15 * "git archive" learned to filter what gets archived with pathspec.
  16
  17 * "git config --edit --global" starts from a skeletal per-user
  18   configuration file contents, instead of a total blank, when the
  19   user does not already have any.  This immediately reduces the
  20   need for a later "Have you forgotten setting core.user?" and we
  21   can add more to the template as we gain more experience.
  22
  23 * "git stash list -p" used to be almost always a no-op because each
  24   stash entry is represented as a merge commit.  It learned to show
  25   the difference between the base commit version and the working tree
  26   version, which is in line with what "git show" gives.
  27
  28 * Sometimes users want to report a bug they experience on their
  29   repository, but they are not at liberty to share the contents of
  30   the repository.  "fast-export" was taught an "--anonymize" option
  31   to replace blob contents, names of people and paths and log
  32   messages with bland and simple strings to help them.
  33
  34 * "log --date=iso" uses a slight variant of ISO 8601 format that is
  35   made more human readable.  A new "--date=iso-strict" option gives
  36   datetime output that is more strictly conformant.
  37
  38 * A broken reimplementation of Git could write an invalid index that
  39   records both stage #0 and higher stage entries for the same path.
  40   We now notice and reject such an index, as there is no sensible
  41   fallback (we do not know if the broken tool wanted to resolve and
  42   forgot to remove higher stage entries, or if it wanted to unresolve
  43   and forgot to remove the stage#0 entry).
  44
  45 * The "pre-receive" and "post-receive" hooks are no longer required
  46   to consume their input fully (not following this requirement used
  47   to result in intermittent errors in "git push").
  48
  49 * The pretty-format specifier "%d", which expanded to " (tagname)"
  50   for a tagged commit, gained a cousin "%D" that just gives the
  51   "tagname" without frills.
  52
  53 * "git push" learned "--signed" push, that allows a push (i.e.
  54   request to update the refs on the other side to point at a new
  55   history, together with the transmission of necessary objects) to be
  56   signed, so that it can be verified and audited, using the GPG
  57   signature of the person who pushed, that the tips of branches at a
  58   public repository really point the commits the pusher wanted to,
  59   without having to "trust" the server.
  60
  61 * "git interpret-trailers" is a new filter to programatically edit
  62    the tail end of the commit log messages.
  63
  64 * "git help everyday" shows the "Everyday Git in 20 commands or so"
  65   document, whose contents have been updated to more modern Git
  66   practice.
  67
  68
  69Performance, Internal Implementation, etc.
  70
  71 * The API to manipulate the "refs" is currently undergoing a revamp
  72   to make it more transactional, with the eventual goal to allow
  73   all-or-none atomic updates and migrating the storage to something
  74   other than the traditional filesystem based one (e.g. databases).
  75
  76 * The lockfile API and its users have been cleaned up.
  77
  78 * We no longer attempt to keep track of individual dependencies to
  79   the header files in the build procedure, relying on automated
  80   dependency generation support from modern compilers.
  81
  82 * In tests, we have been using NOT_{MINGW,CYGWIN} test prerequisites
  83   long before negated prerequisites e.g. !MINGW were invented.
  84   The former has been converted to the latter to avoid confusion.
  85
  86 * Looking up remotes configuration in a repository with very many
  87   remotes defined has been optimized.
  88
  89 * There are cases where you lock and open to write a file, close it
  90   to show the updated contents to external processes, and then have
  91   to update the file again while still holding the lock, but the
  92   lockfile API lacked support for such an access pattern.
  93
  94 * The API to allocate the structure to keep track of commit
  95   decoration has been updated to make it less cumbersome to use.
  96
  97 * An in-core caching layer to let us avoid reading the same
  98   configuration files number of times has been added.  A few commands
  99   have been converted to use this subsystem.
 100
 101 * Various code paths have been cleaned up and simplified by using
 102   "strbuf", "starts_with()", and "skip_prefix()" APIs more.
 103
 104 * A few codepaths that died when large blobs that would not fit in
 105   core are involved in their operation have been taught to punt
 106   instead, by e.g. marking too large a blob as not to be diffed.
 107
 108 * A few more code paths in "commit" and "checkout" have been taught
 109   to repopulate the cache-tree in the index, to help speed up later
 110   "write-tree" (used in "commit") and "diff-index --cached" (used in
 111   "status").
 112
 113 * A common programming mistake to assign the same short option name
 114   to two separate options is detected by parse_options() API to help
 115   developers.
 116
 117 * The code path to write out the packed-refs file has been optimized,
 118   which especially matters in a repository with a large number of
 119   refs.
 120
 121 * The check to see if a ref $F can be created by making sure no
 122   existing ref has $F/ as its prefix has been optimized, which
 123   especially matters in a repository with a large number of existing
 124   refs.
 125
 126 * "git fsck" was taught to check contents of tag objects a bit more.
 127
 128 * "git hash-object" was taught a "--literally" option to help
 129   debugging.
 130
 131 * When running a required clean filter, we do not have to mmap the
 132   original before feeding the filter.  Instead, stream the file
 133   contents directly to the filter and process its output.
 134
 135 * The scripts in the test suite can be run with "-x" option to show
 136   a shell-trace of each command run in them.
 137
 138
 139Also contains various documentation updates and code clean-ups.
 140
 141
 142Fixes since v2.1
 143----------------
 144
 145Unless otherwise noted, all the fixes since v2.1 in the maintenance
 146track are contained in this release (see the maintenance releases'
 147notes for details).
 148
 149 * "git log --pretty/format=" with an empty format string did not
 150   mean the more obvious "No output whatsoever" but "Use default
 151   format", which was counterintuitive.
 152
 153 * "git -c section.var command" and "git -c section.var= command"
 154   should pass the configuration differently (the former should be a
 155   boolean true, the latter should be an empty string).
 156
 157 * Applying a patch not generated by Git in a subdirectory used to
 158   check the whitespace breakage using the attributes for incorrect
 159   paths. Also whitespace checks were performed even for paths
 160   excluded via "git apply --exclude=<path>" mechanism.
 161
 162 * "git bundle create" with date-range specification were meant to
 163   exclude tags outside the range, but it didn't.
 164
 165 * "git add x" where x that used to be a directory has become a
 166   symbolic link to a directory misbehaved.
 167
 168 * The prompt script checked $GIT_DIR/ref/stash file to see if there
 169   is a stash, which was a no-no.
 170
 171 * Pack-protocol documentation had a minor typo.
 172
 173 * "git checkout -m" did not switch to another branch while carrying
 174   the local changes forward when a path was deleted from the index.
 175
 176 * "git daemon" (with NO_IPV6 build configuration) used to incorrectly
 177   use the hostname even when gethostbyname() reported that the given
 178   hostname is not found.
 179   (merge 107efbe rs/daemon-fixes later to maint).
 180
 181 * With sufficiently long refnames, "git fast-import" could have
 182   overflown an on-stack buffer.
 183
 184 * After "pack-refs --prune" packed refs at the top-level, it failed
 185   to prune them.
 186
 187 * Progress output from "git gc --auto" was visible in "git fetch -q".
 188
 189 * We used to pass -1000 to poll(2), expecting it to also mean "no
 190   timeout", which should be spelled as -1.
 191
 192 * "git rebase" documentation was unclear that it is required to
 193   specify on what <upstream> the rebase is to be done when telling it
 194   to first check out <branch>.
 195   (merge 95c6826 so/rebase-doc later to maint).
 196
 197 * "git push" over HTTP transport had an artificial limit on number of
 198   refs that can be pushed imposed by the command line length.
 199   (merge 26be19b jk/send-pack-many-refspecs later to maint).
 200
 201 * When receiving an invalid pack stream that records the same object
 202   twice, multiple threads got confused due to a race.
 203   (merge ab791dd jk/index-pack-threading-races later to maint).
 204
 205 * An attempt to remove the entire tree in the "git fast-import" input
 206   stream caused it to misbehave.
 207   (merge 2668d69 mb/fast-import-delete-root later to maint).
 208
 209 * Reachability check (used in "git prune" and friends) did not add a
 210   detached HEAD as a starting point to traverse objects still in use.
 211   (merge c40fdd0 mk/reachable-protect-detached-head later to maint).
 212
 213 * "git config --add section.var val" used to lose existing
 214   section.var whose value was an empty string.
 215   (merge c1063be ta/config-add-to-empty-or-true-fix later to maint).
 216
 217 * "git fsck" failed to report that it found corrupt objects via its
 218   exit status in some cases.
 219   (merge 30d1038 jk/fsck-exit-code-fix later to maint).
 220
 221 * Use of "--verbose" option used to break "git branch --merged".
 222   (merge 12994dd jk/maint-branch-verbose-merged later to maint).
 223
 224 * Some MUAs mangled a line in a message that begins with "From " to
 225   ">From " when writing to a mailbox file and feeding such an input
 226   to "git am" used to lose such a line.
 227   (merge 85de86a jk/mbox-from-line later to maint).
 228
 229 * "rev-parse --verify --quiet $name" is meant to quietly exit with a
 230   non-zero status when $name is not a valid object name, but still
 231   gave error messages in some cases.
 232
 233 * A handful of C source files have been updated to include
 234   "git-compat-util.h" as the first thing, to conform better to our
 235   coding guidelines.
 236   (merge 1c4b660 da/include-compat-util-first-in-c later to maint).
 237
 238 * t7004 test, which tried to run Git with small stack space, has been
 239   updated to give a bit larger stack to avoid false breakage on some
 240   platforms.
 241   (merge b9a1907 sk/tag-contains-wo-recursion later to maint).
 242
 243 * A few documentation pages had example sections marked up not quite
 244   correctly, which passed AsciiDoc but failed with AsciiDoctor.
 245   (merge c30c43c bc/asciidoc-pretty-formats-fix later to maint).
 246   (merge f8a48af bc/asciidoc later to maint).
 247
 248 * "gitweb" used deprecated CGI::startfrom, which was removed from
 249   CGI.pm as of 4.04; use CGI::start_from instead.
 250   (merge 4750f4b rm/gitweb-start-form later to maint).