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 * The support to build with NO_PTHREADS has been resurrected. 13 14 * Compilation options has been updated a bit to support z/OS port 15 better. 16 17 18UI, Workflows & Features 19 20 * "git archive" learned to filter what gets archived with pathspec. 21 22 * "git config --edit --global" starts from a skeletal per-user 23 configuration file contents, instead of a total blank, when the 24 user does not already have any. This immediately reduces the 25 need for a later "Have you forgotten setting core.user?" and we 26 can add more to the template as we gain more experience. 27 28 * "git stash list -p" used to be almost always a no-op because each 29 stash entry is represented as a merge commit. It learned to show 30 the difference between the base commit version and the working tree 31 version, which is in line with what "git show" gives. 32 33 * Sometimes users want to report a bug they experience on their 34 repository, but they are not at liberty to share the contents of 35 the repository. "fast-export" was taught an "--anonymize" option 36 to replace blob contents, names of people and paths and log 37 messages with bland and simple strings to help them. 38 39 * "git difftool" learned an option to stop feeding paths to the 40 diff backend when it exits with a non-zero status. 41 42 * "log --date=iso" uses a slight variant of ISO 8601 format that is 43 made more human readable. A new "--date=iso-strict" option gives 44 datetime output that is more strictly conformant. 45 46 * The logic "git prune" uses is more resilient against various corner 47 cases. 48 49 * A broken reimplementation of Git could write an invalid index that 50 records both stage #0 and higher stage entries for the same path. 51 We now notice and reject such an index, as there is no sensible 52 fallback (we do not know if the broken tool wanted to resolve and 53 forgot to remove higher stage entries, or if it wanted to unresolve 54 and forgot to remove the stage#0 entry). 55 56 * The temporary files "git mergetool" uses are named to avoid too 57 many dots in them (e.g. a temporary file for "hello.c" used to be 58 named e.g. "hello.BASE.4321.c" but now uses underscore instead, 59 e.g. "hello_BASE_4321.c"). 60 61 * The temporary files "git mergetools" uses can be placed in a newly 62 creted temporary directory, instead of the current directory, by 63 setting the mergetool.writeToTemp configuration variable. 64 65 * "git mergetool" understands "--tool bc" now, as version 4 of 66 BeyondCompare can be driven the same way as its version 3 and it 67 feels awkward to say "--tool bc3". 68 69 * The "pre-receive" and "post-receive" hooks are no longer required 70 to consume their input fully (not following this requirement used 71 to result in intermittent errors in "git push"). 72 73 * The pretty-format specifier "%d", which expanded to " (tagname)" 74 for a tagged commit, gained a cousin "%D" that just gives the 75 "tagname" without frills. 76 77 * "git push" learned "--signed" push, that allows a push (i.e. 78 request to update the refs on the other side to point at a new 79 history, together with the transmission of necessary objects) to be 80 signed, so that it can be verified and audited, using the GPG 81 signature of the person who pushed, that the tips of branches at a 82 public repository really point the commits the pusher wanted to, 83 without having to "trust" the server. 84 85 * "git interpret-trailers" is a new filter to programatically edit 86 the tail end of the commit log messages. 87 88 * "git help everyday" shows the "Everyday Git in 20 commands or so" 89 document, whose contents have been updated to more modern Git 90 practice. 91 92 93Performance, Internal Implementation, etc. 94 95 * The API to manipulate the "refs" has been restructured to make it 96 more transactional, with the eventual goal to allow all-or-none 97 atomic updates and migrating the storage to something other than 98 the traditional filesystem based one (e.g. databases). 99 100 * The lockfile API and its users have been cleaned up. 101 102 * We no longer attempt to keep track of individual dependencies to 103 the header files in the build procedure, relying on automated 104 dependency generation support from modern compilers. 105 106 * In tests, we have been using NOT_{MINGW,CYGWIN} test prerequisites 107 long before negated prerequisites e.g. !MINGW were invented. 108 The former has been converted to the latter to avoid confusion. 109 110 * Looking up remotes configuration in a repository with very many 111 remotes defined has been optimized. 112 113 * There are cases where you lock and open to write a file, close it 114 to show the updated contents to external processes, and then have 115 to update the file again while still holding the lock, but the 116 lockfile API lacked support for such an access pattern. 117 118 * The API to allocate the structure to keep track of commit 119 decoration has been updated to make it less cumbersome to use. 120 121 * An in-core caching layer to let us avoid reading the same 122 configuration files number of times has been added. A few commands 123 have been converted to use this subsystem. 124 125 * Various code paths have been cleaned up and simplified by using 126 "strbuf", "starts_with()", and "skip_prefix()" APIs more. 127 128 * A few codepaths that died when large blobs that would not fit in 129 core are involved in their operation have been taught to punt 130 instead, by e.g. marking too large a blob as not to be diffed. 131 132 * A few more code paths in "commit" and "checkout" have been taught 133 to repopulate the cache-tree in the index, to help speed up later 134 "write-tree" (used in "commit") and "diff-index --cached" (used in 135 "status"). 136 137 * A common programming mistake to assign the same short option name 138 to two separate options is detected by parse_options() API to help 139 developers. 140 141 * The code path to write out the packed-refs file has been optimized, 142 which especially matters in a repository with a large number of 143 refs. 144 145 * The check to see if a ref $F can be created by making sure no 146 existing ref has $F/ as its prefix has been optimized, which 147 especially matters in a repository with a large number of existing 148 refs. 149 150 * "git fsck" was taught to check contents of tag objects a bit more. 151 152 * "git hash-object" was taught a "--literally" option to help 153 debugging. 154 155 * When running a required clean filter, we do not have to mmap the 156 original before feeding the filter. Instead, stream the file 157 contents directly to the filter and process its output. 158 159 * The scripts in the test suite can be run with "-x" option to show 160 a shell-trace of each command run in them. 161 162 * The "run-command" API learned to manage the argv and environment 163 array for child process, alleviating the need for the callers to 164 allocate and deallocate them. 165 166 * Some people use AsciiDoctor, instead of AsciiDoc, to format our 167 documentation set; the documentation has been adjusted, as 168 AsciiDoctor is pickier than AsciiDoc in its input mark-up. 169 170 171Also contains various documentation updates and code clean-ups. 172 173 174Fixes since v2.1 175---------------- 176 177Unless otherwise noted, all the fixes since v2.1 in the maintenance 178track are contained in this release (see the maintenance releases' 179notes for details). 180 181 * "git log --pretty/format=" with an empty format string did not 182 mean the more obvious "No output whatsoever" but "Use default 183 format", which was counterintuitive. 184 185 * "git -c section.var command" and "git -c section.var= command" 186 should pass the configuration differently (the former should be a 187 boolean true, the latter should be an empty string). 188 189 * Applying a patch not generated by Git in a subdirectory used to 190 check the whitespace breakage using the attributes for incorrect 191 paths. Also whitespace checks were performed even for paths 192 excluded via "git apply --exclude=<path>" mechanism. 193 194 * "git bundle create" with date-range specification were meant to 195 exclude tags outside the range, but it didn't. 196 197 * "git add x" where x that used to be a directory has become a 198 symbolic link to a directory misbehaved. 199 200 * The prompt script checked $GIT_DIR/ref/stash file to see if there 201 is a stash, which was a no-no. 202 203 * Pack-protocol documentation had a minor typo. 204 205 * "git checkout -m" did not switch to another branch while carrying 206 the local changes forward when a path was deleted from the index. 207 208 * "git daemon" (with NO_IPV6 build configuration) used to incorrectly 209 use the hostname even when gethostbyname() reported that the given 210 hostname is not found. 211 (merge 107efbe rs/daemon-fixes later to maint). 212 213 * With sufficiently long refnames, "git fast-import" could have 214 overflown an on-stack buffer. 215 216 * After "pack-refs --prune" packed refs at the top-level, it failed 217 to prune them. 218 219 * Progress output from "git gc --auto" was visible in "git fetch -q". 220 221 * We used to pass -1000 to poll(2), expecting it to also mean "no 222 timeout", which should be spelled as -1. 223 224 * "git rebase" documentation was unclear that it is required to 225 specify on what <upstream> the rebase is to be done when telling it 226 to first check out <branch>. 227 (merge 95c6826 so/rebase-doc later to maint). 228 229 * "git push" over HTTP transport had an artificial limit on number of 230 refs that can be pushed imposed by the command line length. 231 (merge 26be19b jk/send-pack-many-refspecs later to maint). 232 233 * When receiving an invalid pack stream that records the same object 234 twice, multiple threads got confused due to a race. 235 (merge ab791dd jk/index-pack-threading-races later to maint). 236 237 * An attempt to remove the entire tree in the "git fast-import" input 238 stream caused it to misbehave. 239 (merge 2668d69 mb/fast-import-delete-root later to maint). 240 241 * Reachability check (used in "git prune" and friends) did not add a 242 detached HEAD as a starting point to traverse objects still in use. 243 (merge c40fdd0 mk/reachable-protect-detached-head later to maint). 244 245 * "git config --add section.var val" used to lose existing 246 section.var whose value was an empty string. 247 (merge c1063be ta/config-add-to-empty-or-true-fix later to maint). 248 249 * "git fsck" failed to report that it found corrupt objects via its 250 exit status in some cases. 251 (merge 30d1038 jk/fsck-exit-code-fix later to maint). 252 253 * Use of "--verbose" option used to break "git branch --merged". 254 (merge 12994dd jk/maint-branch-verbose-merged later to maint). 255 256 * Some MUAs mangled a line in a message that begins with "From " to 257 ">From " when writing to a mailbox file and feeding such an input 258 to "git am" used to lose such a line. 259 (merge 85de86a jk/mbox-from-line later to maint). 260 261 * "rev-parse --verify --quiet $name" is meant to quietly exit with a 262 non-zero status when $name is not a valid object name, but still 263 gave error messages in some cases. 264 265 * A handful of C source files have been updated to include 266 "git-compat-util.h" as the first thing, to conform better to our 267 coding guidelines. 268 (merge 1c4b660 da/include-compat-util-first-in-c later to maint). 269 270 * t7004 test, which tried to run Git with small stack space, has been 271 updated to give a bit larger stack to avoid false breakage on some 272 platforms. 273 (merge b9a1907 sk/tag-contains-wo-recursion later to maint). 274 275 * A few documentation pages had example sections marked up not quite 276 correctly, which passed AsciiDoc but failed with AsciiDoctor. 277 (merge c30c43c bc/asciidoc-pretty-formats-fix later to maint). 278 (merge f8a48af bc/asciidoc later to maint). 279 280 * "gitweb" used deprecated CGI::startfrom, which was removed from 281 CGI.pm as of 4.04; use CGI::start_from instead. 282 (merge 4750f4b rm/gitweb-start-form later to maint). 283 284 * Newer versions of 'meld' breaks the auto-detection we use to see if 285 they are new enough to support the `--output` option. 286 (merge b12d045 da/mergetool-meld later to maint). 287 288 * "git pack-objects" forgot to disable the codepath to generate 289 object recheability bitmap when it needs to split the resulting 290 pack. 291 (merge 2113471 jk/pack-objects-no-bitmap-when-splitting later to maint).