1Git v2.3 Release Notes 2====================== 3 4Updates since v2.2 5------------------ 6 7Ports 8 9 * Recent gcc toolchain on Cygwin started throwing compilation warning, 10 which has been squelched. 11 12 * A few updates to build on platforms that lack tv_nsec, 13 clock_gettime, CLOCK_MONOTONIC and HMAC_CTX_cleanup (e.g. older 14 RHEL) have been added. 15 16 17UI, Workflows & Features 18 19 * It was cumbersome to use "GIT_SSH" mechanism when the user wanted 20 to pass an extra set of arguments to the underlying ssh. A new 21 environment variable GIT_SSH_COMMAND can be used for this. 22 23 * A request to store an empty note via "git notes" meant to remove 24 note from the object but with --allow-empty we will store a 25 (surprise!) note that is empty. 26 27 * "git interpret-trailers" learned to properly handle the 28 "Conflicts:" block at the end. 29 30 * "git am" learned "--message-id" option to copy the message ID of 31 the incoming e-mail to the log message of resulting commit. 32 33 * "git clone --reference=<over there>" learned the "--dissociate" 34 option to go with it; it borrows objects from the reference object 35 store while cloning only to reduce network traffic and then 36 dissociates the resulting clone from the reference by performing 37 local copies of borrowed objects. 38 39 * "git send-email" learned "--transfer-encoding" option to force a 40 non-fault Content-Transfer-Encoding header (e.g. base64). 41 42 * "git send-email" normally identifies itself via X-Mailer: header in 43 the message it sends out. A new command line flag --no-xmailer 44 allows the user to squelch the header. 45 46 * "git push" into a repository with a working tree normally refuses 47 to modify the branch that is checked out. The command learned to 48 optionally do an equivalent of "git reset --hard" only when there 49 is no change to the working tree and the index instead, which would 50 be useful to "deploy" by pushing into a repository. 51 52 * "git new-workdir" (in contrib/) can be used to populate an empty 53 and existing directory now. 54 55 * Credential helpers are asked in turn until one of them give 56 positive response, which is cumbersome to turn off when you need to 57 run Git in an automated setting. The credential helper interface 58 learned to allow a helper to say "stop, don't ask other helpers." 59 Also GIT_TERMINAL_PROMPT environment can be set to false to disable 60 our built-in prompt mechanism for passwords. 61 62 * "git branch -d" (delete) and "git branch -m" (move) learned to 63 honor "-f" (force) flag; unlike many other subcommands, the way to 64 force these have been with separate "-D/-M" options, which was 65 inconsistent. 66 67 * "diff-highlight" filter (in contrib/) allows its color output to be 68 customized via configuration variables. 69 70 * "git imap-send" learned to take "-v" (verbose) and "-q" (quiet) 71 command line options. 72 73 * "git remote add $name $URL" is now allowed when "url.$URL.insteadOf" 74 is already defined. 75 76 * "git imap-send" now can be built to use cURL library to talk to 77 IMAP servers (if the library is recent enough, of course). 78 This allows you to use authenticate method other than CRAM-MD5, 79 among other things. 80 81 * "git imap-send" now allows GIT_CURL_VERBOSE environment variable to 82 control the verbosity when talking via the cURL library. 83 84 * The prompt script (in contrib/) learned to optionally hide prompt 85 when in an ignored directory by setting GIT_PS1_HIDE_IF_PWD_IGNORED 86 shell variable. 87 88 89Performance, Internal Implementation, Development Support etc. 90 91 * Earlier we made "rev-list --object-edge" more aggressively list the 92 objects at the edge commits, in order to reduce number of objects  93 fetched into a shallow repository, but the change affected cases 94 other than "fetching into a shallow repository" and made it 95 unusably slow (e.g. fetching into a normal repository should not 96 have to suffer the overhead from extra processing). Limit it to a 97 more specific case by introducing --objects-edge-aggressive, a new 98 option to rev-list. 99 100 * Squelched useless compiler warnings on Mac OS X regarding the 101 crypto API. 102 103 * The procedure to generate unicode table has been simplified. 104 105 * Some filesystems assign filemodes in a strange way, fooling then 106 automatic "filemode trustability" check done during a new 107 repository creation. The initialization codepath has been hardened 108 against this issue. 109 110 * The codepath in "git remote update --prune" to drop many refs has 111 been optimized. 112 113 * The API into get_merge_bases*() family of functions was easy to 114 misuse, which has been corrected to make it harder to do so. 115 116 * Long overdue departure from the assumption that S_IFMT is shared by 117 everybody made in 2005, which was necessary to port to z/OS. 118 119 * "git push" and "git fetch" did not communicate an overlong refname 120 correctly. Now it uses 64kB sideband to accommodate longer ones. 121 122 * Recent GPG changes the keyring format and drops support for RFC1991 123 formatted signatures, breaking our existing tests. 124 125 * "git-prompt" (in contrib/) used a variable from the global scope, 126 possibly contaminating end-user's namespace. 127 128 129Also contains various documentation updates and code clean-ups. 130 131 132Fixes since v2.2 133---------------- 134 135Unless otherwise noted, all the fixes since v2.2 in the maintenance 136track are contained in this release (see the maintenance releases' 137notes for details). 138 139 * The logic in "git bisect bad HEAD" etc. to avoid forcing the test 140 of the common ancestor of bad and good commits was broken. 141 (merge 07913d5 cc/bisect-rev-parsing later to maint). 142 143 * "git checkout-index --temp=$target $path" did not work correctly 144 for paths outside the current subdirectory in the project. 145 (merge 74c4de5 es/checkout-index-temp later to maint). 146 147 * The report from "git checkout" on a branch that builds on another 148 local branch by setting its branch.*.merge to branch name (not a 149 full refname) incorrectly said that the upstream is gone. 150 (merge 05e7368 jc/checkout-local-track-report later to maint). 151 152 * With The git-prompt support (in contrib/), using the exit status of 153 the last command in the prompt, e.g. PS1='$(__git_ps1) $? ', did 154 not work well, because the helper function stomped on the exit 155 status. 156 (merge 6babe76 tf/prompt-preserve-exit-status later to maint). 157 158 * Recent update to "git commit" broke amending an existing commit 159 with bogus author/committer lines without a valid e-mail address. 160 (merge c83a509 jk/commit-date-approxidate later to maint). 161 162 * The lockfile API used to get confused which file to clean up when 163 the process moved the $cwd after creating a lockfile. 164 (merge fa137f6 nd/lockfile-absolute later to maint). 165 166 * Traditionally we tried to avoid interpreting date strings given by 167 the user as future dates, e.g. GIT_COMMITTER_DATE=2014-12-10 when 168 used early November 2014 was taken as "October 12, 2014" because it 169 is likely that a date in the future, December 10, is a mistake. 170 This heuristics has been loosened to allow people to express future 171 dates (most notably, --until=<date> may want to be far in the 172 future) and we no longer tiebreak by future-ness of the date when 173 174 (1) ISO-like format is used, and 175 (2) the string can make sense interpreted as both y-m-d and y-d-m. 176 177 Git may still have to use the heuristics to tiebreak between dd/mm/yy 178 and mm/dd/yy, though. 179 (merge d372395 jk/approxidate-avoid-y-d-m-over-future-dates later to maint). 180 181 * Git did not correctly read an overlong refname from a packed refs 182 file. 183 (merge ea41783 jk/read-packed-refs-without-path-max later to maint). 184 185 * "git apply" was described in the documentation to take --ignore-date 186 option, which it does not. 187 (merge 0cef4e7 rw/apply-does-not-take-ignore-date later to maint). 188 189 * "git add -i" did not notice when the interactive command input 190 stream went away and kept asking the same question. 191 (merge a8bec7a jk/add-i-read-error later to maint). 192 193 * "git send-email" did not handle RFC 2047 encoded headers quite 194 right. 195 (merge ab47e2a rd/send-email-2047-fix later to maint). 196 197 * New tag object format validation added in 2.2 showed garbage after 198 a tagname it reported in its error message. 199 (merge a1e920a js/fsck-tag-validation later to maint). 200 201 * The code that reads the reflog from the newer to the older entries 202 did not handle an entry that crosses a boundary of block it uses to 203 read them correctly. 204 (merge 69216bf jk/for-each-reflog-ent-reverse later to maint). 205 206 * "git diff -B -M" after making a new copy B out of an existing file 207 A and then editing A extensively ought to report that B was created 208 by copying A and A was modified, which is what "git diff -C" 209 reports, but it instead said A was renamed to B and A was edited 210 heavily in place. This was not just incoherent but also failed to 211 apply with "git apply". The report has been corrected to match what 212 "git diff -C" produces for this case. 213 (merge 6936b58 jc/diff-b-m later to maint). 214 215 * In files we pre-populate for the user to edit with commented hints, 216 a line of hint that is indented with a tab used to show as '#' (or 217 any comment char), ' ' (space), and then the hint text that began 218 with the tab, which some editors flag as an indentation error (tab 219 following space). We now omit the space after the comment char in 220 such a case. 221 (merge d55aeb7 jc/strbuf-add-lines-avoid-sp-ht-sequence later to maint). 222 223 * "git ls-tree" does not support path selection based on negative 224 pathspecs, but did not error out when negative pathspecs are given. 225 (merge f1f6224 nd/ls-tree-pathspec later to maint). 226 227 * The function sometimes returned a non-freeable memory and some 228 other times returned a piece of memory that must be freed, leading 229 to inevitable leaks. 230 (merge 59362e5 jc/exec-cmd-system-path-leak-fix later to maint). 231 232 * The code to abbreviate an object name to its short unique prefix 233 has been optimized when no abbreviation was requested. 234 (merge 61e704e mh/find-uniq-abbrev later to maint). 235 236 * "git add --ignore-errors ..." did not ignore an error to 237 give a file that did not exist. 238 (merge 1d31e5a mg/add-ignore-errors later to maint). 239 240 * "git checkout $treeish $path", when $path in the index and the 241 working tree already matched what is in $treeish at the $path, 242 still overwrote the $path unnecessarily. 243 (merge c5326bd jk/checkout-from-tree later to maint). 244 245 * "git config --get-color" did not parse its command line arguments 246 carefully. 247 (merge cb35722 jk/colors-fix later to maint). 248 249 * open() emulated on Windows platforms did not give EISDIR upon 250 an attempt to open a directory for writing. 251 (merge ba6fad0 js/windows-open-eisdir-error later to maint). 252 253 * A few code paths used abs() when they should have used labs() on 254 long integers. 255 (merge 83915ba rs/maint-config-use-labs later to maint). 256 (merge 31a8aa1 rs/receive-pack-use-labs later to maint). 257 258 * "gitweb" used to depend on a behaviour recent CGI.pm deprecated. 259 (merge 13dbf46 jk/gitweb-with-newer-cgi-multi-param later to maint). 260 261 * "git init" (hence "git clone") initialized the per-repository 262 configuration file .git/config with x-bit by mistake. 263 (merge 1f32ecf mh/config-flip-xbit-back-after-checking later to maint). 264 265 * Recent update in Git 2.2 started creating objects/info/packs and 266 info/refs files with permission bits tighter than user's umask. 267 (merge d91175b jk/prune-packed-server-info later to maint). 268 269 * Git 2.0 was supposed to make the "simple" mode for the default of 270 "git push", but it didn't. 271 (merge 00a6fa0 jk/push-simple later to maint). 272 273 * "Everyday" document had a broken link. 274 (merge 366c8d4 po/everyday-doc later to maint). 275 276 * A few test fixes. 277 (merge 880ef58 jk/no-perl-tests later to maint). 278 279 * The build procedure did not bother fixing perl and python scripts 280 when NO_PERL and NO_PYTHON build-time configuration changed. 281 (merge ca2051d jk/rebuild-perl-scripts-with-no-perl-seting-change later to maint). 282 283 * The usage string of "git log" command was marked incorrectly for 284 l10n. 285 (merge e66dc0c km/log-usage-string-i18n later to maint). 286 287 * "git for-each-ref" mishandled --format="%(upstream:track)" when a 288 branch is marked to have forked from a non-existing branch. 289 (merge b6160d9 rc/for-each-ref-tracking later to maint).