Documentation / RelNotes / 1.8.3.txton commit remote-curl: die directly with http error messages (de89f0b)
   1Git v1.8.3 Release Notes
   2========================
   3
   4Backward compatibility notes (for Git 2.0)
   5------------------------------------------
   6
   7When "git push [$there]" does not say what to push, we have used the
   8traditional "matching" semantics so far (all your branches were sent
   9to the remote as long as there already are branches of the same name
  10over there).  In Git 2.0, the default will change to the "simple"
  11semantics that pushes the current branch to the branch with the same
  12name, only when the current branch is set to integrate with that
  13remote branch.  There is a user preference configuration variable
  14"push.default" to change this.  If you are an old-timer who is used
  15to the "matching" semantics, you can set it to "matching" to keep the
  16traditional behaviour.  If you want to live in the future early,
  17you can set it to "simple" today without waiting for Git 2.0.
  18
  19When "git add -u" and "git add -A", that does not specify what paths
  20to add on the command line is run from inside a subdirectory, these
  21commands will operate on the entire tree in Git 2.0 for consistency
  22with "git commit -a" and other commands. Because there will be no
  23mechanism to make "git add -u" behave as if "git add -u .", it is
  24important for those who are used to "git add -u" (without pathspec)
  25updating the index only for paths in the current subdirectory to start
  26training their fingers to explicitly say "git add -u ." when they mean
  27it before Git 2.0 comes.
  28
  29
  30Updates since v1.8.2
  31--------------------
  32
  33UI, Workflows & Features
  34
  35 * When the interactive access to git-shell is not enabled, it issues
  36   a message meant to help the system admininstrator to enable it.
  37   An explicit way to help the end users who connect to the service by
  38   issuing custom messages to refuse such an access has been added.
  39
  40 * In addition to the case where the user edits the log message with
  41   the "e)dit" option of "am -i", replace the "Applying: this patch"
  42   message with the final log message contents after applymsg hook
  43   munges it.
  44
  45 * "git status" suggests users to look into using --untracked=no option
  46   when it takes too long.
  47
  48 * "git status" shows a bit more information to "git status" during a
  49   rebase/bisect session.
  50
  51 * "git fetch" learned to fetch a commit at the tip of an unadvertised
  52   ref by specifying a raw object name from the command line when the
  53   server side supports this feature.
  54
  55 * "git count-objects -v" learned to report leftover temporary
  56   packfiles and other garbage in the object store.
  57
  58 * A new read-only credential helper (in contrib/) to interact with
  59   the .netrc/.authinfo files has been added.
  60
  61 * "git send-email" can be used with the credential helper system.
  62
  63 * There was no Porcelain way to say "I no longer am interested in
  64   this submodule", once you express your interest in a submodule with
  65   "submodule init".  "submodule deinit" is the way to do so.
  66
  67 * "git pull --rebase" learned to pass "-v/-q" options to underlying
  68   "git rebase".
  69
  70 * The new "--follow-tags" option tells "git push" to push relevant
  71   annotated tags when pushing branches out.
  72
  73 * "git merge" and "git pull" can optionally be told to inspect and
  74   reject when merging a commit that does not carry a trusted GPG
  75   signature.
  76
  77 * "git mergetool" now feeds files to the "p4merge" backend in the
  78   order that matches the p4 convention, where "theirs" is usually
  79   shown on the left side, which is the opposite from other backend
  80   expects.
  81
  82 * "show/log" now honors gpg.program configuration just like other
  83   parts of the code that use GnuPG.
  84
  85 * "git log" that shows the difference between the parent and the
  86   child has been optimized somewhat.
  87
  88 * "git difftool" allows the user to write into the temporary files
  89   being shown; if the user makes changes to the working tree at the
  90   same time, one of the changes has to be lost in such a case, but it
  91   tells the user what happened and refrains from overwriting the copy
  92   in the working tree.
  93
  94 * There was no good way to ask "I have a random string that came from
  95   outside world. I want to turn it into a 40-hex object name while
  96   making sure such an object exists".  A new peeling suffix ^{object}
  97   can be used for that purpose, together with "rev-parse --verify".
  98
  99
 100Performance, Internal Implementation, etc.
 101
 102 * Updates for building under msvc.
 103
 104 * A few codepaths knew how much data they need to put in the
 105   hashtables they use upfront, but still started from a small table
 106   repeatedly growing and rehashing.
 107
 108 * The API to walk reflog entries from the latest to older, which was
 109   necessary for operations such as "git checkout -", was cumbersome
 110   to use correctly and also inefficient.
 111
 112 * Codepaths that inspect log-message-to-be and decide when to add a
 113   new Signed-off-by line in various commands have been consolidated.
 114
 115 * The pkt-line API, implementation and its callers have been cleaned
 116   up to make them more robust.
 117
 118 * Cygwin port has a faster-but-lying lstat(2) emulation whose
 119   incorrectness does not matter in practice except for a few
 120   codepaths, and setting permission bits to directories is a codepath
 121   that needs to use a more correct one.
 122
 123 * "git checkout" had repeated pathspec matches on the same paths,
 124   which have been consolidated.  Also a bug in "git checkout dir/"
 125   that is started from an unmerged index has been fixed.
 126
 127
 128Also contains minor documentation updates and code clean-ups.
 129
 130
 131Fixes since v1.8.2
 132------------------
 133
 134Unless otherwise noted, all the fixes since v1.8.2 in the maintenance
 135track are contained in this release (see release notes to them for
 136details).
 137
 138 * "git merge $(git rev-parse v1.8.2)" behaved quite differently from
 139   "git merge v1.8.2", as if v1.8.2 were written as v1.8.2^0 and did
 140   not pay much attention to the annotated tag payload.  Make the code
 141   notice the type of the tag object, in addition to the dwim_ref()
 142   based classification the current code uses (i.e. the name appears
 143   in refs/tags/) to decide when to special case merging of tags.
 144   (merge a38d3d7 jc/merge-tag-object later to maint).
 145
 146 * Fix 1.8.1.x regression that stopped matching "dir" (without
 147   trailing slash) to a directory "dir".
 148   (merge efa5f82 jc/directory-attrs-regression-fix later to maint-1.8.1).
 149
 150 * "git apply --whitespace=fix" was not prepared to see a line getting
 151   longer after fixing whitespaces (e.g. tab-in-indent aka Python).
 152   (merge 329b26e jc/apply-ws-fix-tab-in-indent later to maint-1.8.1).
 153
 154 * The prompt string generator (in contrib/completion/) did not notice
 155   when we are in a middle of a "git revert" session.
 156   (merge 3ee4452 rr/prompt-revert-head later to maint).
 157
 158 * "submodule summary --summary-limit" option did not support
 159   "--option=value" form.
 160   (merge 862ae6c rs/submodule-summary-limit later to maint).
 161
 162 * "index-pack --fix-thin" used uninitialize value to compute delta
 163   depths of objects it appends to the resulting pack.
 164   (merge 57165db jk/index-pack-correct-depth-fix later to maint).
 165
 166 * "index-pack --verify-stat" used a few counters outside protection
 167   of mutex, possibly showing incorrect numbers.
 168   (merge 8f82aad nd/index-pack-threaded-fixes later to maint).
 169
 170 * The code to keep track of what directory names are known to Git on
 171   platforms with case insensitive filesystems can get confused upon a
 172   hash collision between these pathnames and looped forever.
 173
 174 * Annotated tags outside refs/tags/ hierarchy were not advertised
 175   correctly to the ls-remote and fetch with recent version of Git.
 176
 177 * Recent optimization broke shallow clones.
 178   (merge f59de5d jk/peel-ref later to maint).
 179
 180 * "git cmd -- ':(top'" was not diagnosed as an invalid syntax, and
 181   instead the parser kept reading beyond the end of the string.
 182
 183 * "git tag -f <tag>" always said "Updated tag '<tag>'" even when
 184   creating a new tag (i.e. not overwriting nor updating).
 185
 186 * "git p4" did not behave well when the path to the root of the P4
 187   client was not its real path.
 188   (merge bbd8486 pw/p4-symlinked-root later to maint).
 189
 190 * "git archive" reports a failure when asked to create an archive out
 191   of an empty tree.  It would be more intuitive to give an empty
 192   archive back in such a case.
 193
 194 * When "format-patch" quoted a non-ascii strings on the header files,
 195   it incorrectly applied rfc2047 and chopped a single character in
 196   the middle of it.
 197
 198 * An aliased command spawned from a bare repository that does not say
 199   it is bare with "core.bare = yes" is treated as non-bare by mistake.
 200
 201 * In "git reflog expire", REACHABLE bit was not cleared from the
 202   correct objects.
 203
 204 * The logic used by "git diff -M --stat" to shorten the names of
 205   files before and after a rename did not work correctly when the
 206   common prefix and suffix between the two filenames overlapped.
 207
 208 * The "--match=<pattern>" option of "git describe", when used with
 209   "--all" to allow refs that are not annotated tags to be used as a
 210   base of description, did not restrict the output from the command
 211   to those that match the given pattern.
 212
 213 * Clarify in the documentation "what" gets pushed to "where" when the
 214   command line to "git push" does not say these explicitly.
 215
 216 * The "--color=<when>" argument to the commands in the diff family
 217   was described poorly.
 218
 219 * The arguments given to pre-rebase hook were not documented.
 220
 221 * The v4 index format was not documented.
 222
 223 * The "--match=<pattern>" argument "git describe" takes uses glob
 224   pattern but it wasn't obvious from the documentation.
 225
 226 * Some sources failed to compile on systems that lack NI_MAXHOST in
 227   their system header (e.g. z/OS).
 228
 229 * Add an example use of "--env-filter" in "filter-branch"
 230   documentation.
 231
 232 * "git bundle verify" did not say "records a complete history" for a
 233   bundle that does not have any prerequisites.
 234
 235 * In the v1.8.0 era, we changed symbols that do not have to be global
 236   to file scope static, but a few functions in graph.c were used by
 237   CGit from sideways bypassing the entry points of the API the
 238   in-tree users use.
 239
 240 * "git update-index -h" did not do the usual "-h(elp)" thing.
 241
 242 * "git index-pack" had a buffer-overflow while preparing an
 243   informational message when the translated version of it was too
 244   long.
 245
 246 * 'git commit -m "$msg"' used to add an extra newline even when
 247   $msg already ended with one.
 248
 249 * The SSL peer verification done by "git imap-send" did not ask for
 250   Server Name Indication (RFC 4366), failing to connect SSL/TLS
 251   sites that serve multiple hostnames on a single IP.
 252
 253 * perl/Git.pm::cat_blob slurped everything in core only to write it
 254   out to a file descriptor, which was not a very smart thing to do.
 255
 256 * "git branch" did not bother to check nonsense command line
 257   parameters and issue errors in many cases.
 258
 259 * Verification of signed tags were not done correctly when not in C
 260   or en/US locale.
 261
 262 * Some platforms and users spell UTF-8 differently; retry with the
 263   most official "UTF-8" when the system does not understand the
 264   user-supplied encoding name that are the common alternative
 265   spellings of UTF-8.
 266
 267 * When export-subst is used, "zip" output recorded incorrect
 268   size of the file.
 269
 270 * "git am $maildir/" applied messages in an unexpected order; sort
 271   filenames read from the maildir/ in a way that is more likely to
 272   sort messages in the order the writing MUA meant to, by sorting
 273   numeric segment in numeric order and non-numeric segment in
 274   alphabetical order.
 275
 276 * "git submodule update", when recursed into sub-submodules, did not
 277   acccumulate the prefix paths.