1Git ???? Release Notes 2====================== 3 4Updates since v2.3 5------------------ 6 7Ports 8 9 * Our default I/O size (8 MiB) for large files was too large for some 10 platforms with smaller SSIZE_MAX, leading to read(2)/write(2) 11 failures. 12 13 * We did not check the curl library version before using 14 CURLOPT_PROXYAUTH feature that may not exist. 15 16 17UI, Workflows & Features 18 19 * The command usage info strings given by "git cmd -h" and in 20 documentation have been tweaked for consistency. 21 22 * The "sync" subcommand of "git p4" now allows users to exclude 23 subdirectories like its "clone" subcommand does. 24 25 * "git log --invert-grep --grep=WIP" will show only commits that do 26 not have the string "WIP" in their messages. 27 28 * "git push" has been taught a "--atomic" option that makes push to 29 update more than one ref an "all-or-none" affair. 30 31 * Extending the "push to deploy" added in 2.3, the behaviour of "git 32 push" when updating the branch that is checked out can now be 33 tweaked by push-to-checkout hook. 34 35 * Using environment variable LANGUAGE and friends on the client side, 36 HTTP-based transports now send Accept-Language when making requests. 37 38 * "git send-email" used to accept a mistaken "y" (or "yes") as an 39 answer to "What encoding do you want to use [UTF-8]? " without 40 questioning. Now it asks for confirmation when the answer looks 41 too short to be a valid encoding name. 42 43 * When "git apply --whitespace=fix" fixed whitespace errors in the 44 common context lines, the command reports that it did so. 45 46 47Performance, Internal Implementation, Development Support etc. 48 49 * Implementation of N_() macro has been updated slightly to help us 50 detect mistakes. 51 52 * Implementation of "reflog expire" has been restructured to fit the 53 reflogs better with the recently updated ref API. 54 55 * The transport-helper did not give transport options such as 56 verbosity, progress, cloning, etc. to import and export based 57 helpers, like it did for fetch and push based helpers, robbing them 58 the chance to honor the wish of the end-users better. 59 60 * The tests that wanted to see that file becomes unreadable after 61 running "chmod a-r file", and the tests that wanted to make sure it 62 is not run as root, we used "can we write into the / directory?" as 63 a cheap substitute, but on some platforms that is not a good 64 heuristics. The tests and their prerequisites have been updated to 65 check what they really require. 66 (merge f400e51 jk/sanity later to maint). 67 68 * The strbuf API was explained between the API documentation and in 69 the header file. Move missing bits to strbuf.h so that programmers 70 can check only one place for all necessary information. 71 72 * The error handling functions and conventions are now documented in 73 the API manual. 74 75 * Optimize attribute look-up, mostly useful in "git grep" on a 76 project that does not use many attributes, by avoiding it when we 77 (should) know that the attributes are not defined in the first 78 place. 79 80 * Typofix in comments. 81 (merge ef2956a ak/git-pm-typofix later to maint). 82 83 * Code clean-up. 84 (merge 0b868f0 sb/hex-object-name-is-at-most-41-bytes-long later to maint). 85 (merge 5d30851 dp/remove-duplicated-header-inclusion later to maint). 86 87 88 89Also contains various documentation updates and code clean-ups. 90 91 92Fixes since v2.3 93---------------- 94 95Unless otherwise noted, all the fixes since v2.3 in the maintenance 96track are contained in this release (see the maintenance releases' 97notes for details). 98 99 * "git blame HEAD -- missing" failed to correctly say "HEAD" when it 100 tried to say "No such path 'missing' in HEAD". 101 (merge a46442f jk/blame-commit-label later to maint). 102 103 * "git rerere" (invoked internally from many mergy operations) did 104 not correctly signal errors when told to update the working tree 105 files and failed to do so for whatever reason. 106 (merge 89ea903 jn/rerere-fail-on-auto-update-failure later to maint). 107 108 * Setting diff.submodule to 'log' made "git format-patch" produce 109 broken patches. 110 (merge 339de50 dk/format-patch-ignore-diff-submodule later to maint). 111 112 * After attempting and failing a password-less authentication 113 (e.g. kerberos), libcURL refuses to fall back to password based 114 Basic authentication without a bit of help/encouragement. 115 (merge 4dbe664 bc/http-fallback-to-password-after-krb-fails later to maint). 116 117 * The "git push" documentation made the "--repo=<there>" option 118 easily misunderstood. 119 (merge 57b92a7 mg/push-repo-option-doc later to maint). 120 121 * Code to read branch name from various files in .git/ directory 122 would have misbehaved if the code to write them left an empty file. 123 (merge 66ec904 jk/status-read-branch-name-fix later to maint). 124 125 * A misspelled conditional that is always true has been fixed. 126 (merge 94ee8e2 jk/remote-curl-an-array-in-struct-cannot-be-null later to maint). 127 128 * The documentation incorrectly said that C(opy) and R(ename) are the 129 only ones that can be followed by the score number in the output in 130 the --raw format. 131 (merge ac1c2d9 jc/diff-format-doc later to maint). 132 133 * A broken pack .idx file in the receiving repository prevented the 134 dumb http transport from fetching a good copy of it from the other 135 side. 136 (merge 8b9c2dd jk/dumb-http-idx-fetch-fix later to maint). 137 138 * The error message from "git commit", when a non-existing author 139 name was given as value to the "--author=" parameter, has been 140 reworded to avoid misunderstanding. 141 (merge 1044b1f mg/commit-author-no-match-malformed-message later to maint). 142 143 * "git log --help" used to show rev-list options that are irrelevant 144 to the "log" command. 145 (merge 3cab02d jc/doc-log-rev-list-options later to maint). 146 147 * "git apply --whitespace=fix" used to under-allocate the memory when 148 the fix resulted in a longer text than the original patch. 149 (merge 407a792 jc/apply-ws-fix-expands later to maint). 150 151 * The interactive "show a list and let the user choose from it" 152 interface "add -i" used showed and prompted to the user even when 153 the candidate list was empty, against which the only "choice" the 154 user could have made was to choose nothing. 155 (merge a9c4641 ak/add-i-empty-candidates later to maint). 156 157 * The insn sheet "git rebase -i" creates did not fully honor 158 core.abbrev settings. 159 (merge edb72d5 ks/rebase-i-abbrev later to maint). 160 161 * "git fetch" over a remote-helper that cannot respond to "list" 162 command could not fetch from a symbolic reference e.g. HEAD. 163 (merge 33cae54 mh/deref-symref-over-helper-transport later to maint). 164 165 * "git push --signed" gave an incorrectly worded error message when 166 the other side did not support the capability. 167 (merge 45917f0 jc/push-cert later to maint). 168 169 * We didn't format an integer that wouldn't fit in "int" but in 170 "uintmax_t" correctly. 171 (merge d306f3d jk/decimal-width-for-uintmax later to maint). 172 173 * Reading configuration from a blob object, when it ends with a lone 174 CR, use to confuse the configuration parser. 175 (merge 1d0655c jk/config-no-ungetc-eof later to maint). 176 177 * The pack bitmap support did not build with older versions of GCC. 178 (merge bd4e882 jk/pack-bitmap later to maint). 179 180 * The documentation wasn't clear that "remote.<nick>.pushURL" and 181 "remote.<nick>.URL" are there to name the same repository accessed 182 via different transports, not two separate repositories. 183 (merge 697f652 jc/remote-set-url-doc later to maint). 184 185 * Older GnuPG implementations may not correctly import the keyring 186 material we prepare for the tests to use. 187 (merge 1f985d6 ch/new-gpg-drops-rfc-1991 later to maint). 188 189 * The credential helper for Windows (in contrib/) used to mishandle 190 a user name with an at-sign in it. 191 (merge 13d261e av/wincred-with-at-in-username-fix later to maint). 192 193 * Longstanding configuration variable naming rules has been added to 194 the documentation. 195 (merge 35840a3 jc/conf-var-doc later to maint). 196 197 * An earlier workaround to squelch unhelpful deprecation warnings 198 from the complier on Mac OSX unnecessarily set minimum required 199 version of the OS, which the user might want to raise (or lower) 200 for other reasons. 201 (merge 88c03eb es/squelch-openssl-warnings-on-macosx later to maint). 202 203 * Certain older vintages of cURL give irregular output from 204 "curl-config --vernum", which confused our build system. 205 (merge 3af6792 tc/curl-vernum-output-broken-in-7.11 later to maint). 206 207 * In v2.2.0, we broke "git prune" that runs in a repository that 208 borrows from an alternate object store. 209 (merge b0a4264 jk/prune-mtime later to maint). 210 211 * "git submodule add" failed to squash "path/to/././submodule" to 212 "path/to/submodule". 213 (merge 8196e72 ps/submodule-sanitize-path-upon-add later to maint). 214 215 * "git merge-file" did not work correctly in a subdirectory. 216 (merge 204a8ff ab/merge-file-prefix later to maint). 217 218 * "git blame" died, trying to free an uninitialized piece of memory. 219 (merge e600592 es/blame-commit-info-fix later to maint). 220 221 * "git fast-import" used to crash when it could not close and 222 conclude the resulting packfile cleanly. 223 (merge 5e915f3 jk/fast-import-die-nicely-fix later to maint). 224 225 * "update-index --refresh" used to leak when an entry cannot be 226 refreshed for whatever reason. 227 (merge bc1c2ca sb/plug-leak-in-make-cache-entry later to maint). 228 229 * The "interpolated-path" option of "git daemon" inserted any string 230 client declared on the "host=" capability request without checking. 231 Sanitize and limit %H and %CH to a saner and a valid DNS name. 232 (merge b485373 jk/daemon-interpolate later to maint). 233 234 * "git daemon" looked up the hostname even when "%CH" and "%IP" 235 interpolations are not requested, which was unnecessary. 236 (merge dc8edc8 rs/daemon-interpolate later to maint). 237 238 * Even though we officially haven't dropped Perl 5.8 support, the 239 Getopt::Long package that came with it does not support "--no-" 240 prefix to negate a boolean option; manually add support to help 241 people with older Getopt::Long package. 242 (merge f471494 km/send-email-getopt-long-workarounds later to maint). 243 244 * "git apply" was not very careful about reading from, removing, 245 updating and creating paths outside the working tree (under 246 --index/--cached) or the current directory (when used as a 247 replacement for GNU patch). 248 (merge e0d201b jc/apply-beyond-symlink later to maint).