1Git v1.8.1 Release Notes 2======================== 3 4Backward compatibility notes 5---------------------------- 6 7In the next major release (not *this* one), we will change the 8behavior of the "git push" command. 9 10When "git push [$there]" does not say what to push, we have used the 11traditional "matching" semantics so far (all your branches were sent 12to the remote as long as there already are branches of the same name 13over there). We will use the "simple" semantics that pushes the 14current branch to the branch with the same name, only when the current 15branch is set to integrate with that remote branch. There is a user 16preference configuration variable "push.default" to change this, and 17"git push" will warn about the upcoming change until you set this 18variable in this release. 19 20"git branch --set-upstream" is deprecated and may be removed in a 21relatively distant future. "git branch [-u|--set-upstream-to]" has 22been introduced with a saner order of arguments to replace it. 23 24 25Updates since v1.8.0 26-------------------- 27 28UI, Workflows & Features 29 30 * Command-line completion scripts for tcsh and zsh have been added. 31 32 * "git-prompt" scriptlet (in contrib/completion) can be told to paint 33 pieces of the hints in the prompt string in colors. 34 35 * Some documentation pages that used to ship only in the plain text 36 format are now formatted in HTML as well. 37 38 * We used to have a workaround for a bug in ancient "less" that 39 causes it to exit without any output when the terminal is resized. 40 The bug has been fixed in "less" version 406 (June 2007), and the 41 workaround has been removed in this release. 42 43 * When "git checkout" checks out a branch, it tells the user how far 44 behind (or ahead) the new branch is relative to the remote tracking 45 branch it builds upon. The message now also advises how to sync 46 them up by pushing or pulling. This can be disabled with the 47 advice.statusHints configuration variable. 48 49 * "git config --get" used to diagnose presence of multiple 50 definitions of the same variable in the same configuration file as 51 an error, but it now applies the "last one wins" rule used by the 52 internal configuration logic. Strictly speaking, this may be an 53 API regression but it is expected that nobody will notice it in 54 practice. 55 56 * A new configuration variable "diff.context" can be used to 57 give the default number of context lines in the patch output, to 58 override the hardcoded default of 3 lines. 59 60 * "git format-patch" learned the "--notes=<ref>" option to give 61 notes for the commit after the three-dash lines in its output. 62 63 * "git log -p -S<string>" now looks for the <string> after applying 64 the textconv filter (if defined); earlier it inspected the contents 65 of the blobs without filtering. 66 67 * "git log --grep=<pcre>" learned to honor the "grep.patterntype" 68 configuration set to "perl". 69 70 * "git replace -d <object>" now interprets <object> as an extended 71 SHA-1 (e.g. HEAD~4 is allowed), instead of only accepting full hex 72 object name. 73 74 * "git rm $submodule" used to punt on removing a submodule working 75 tree to avoid losing the repository embedded in it. Because 76 recent git uses a mechanism to separate the submodule repository 77 from the submodule working tree, "git rm" learned to detect this 78 case and removes the submodule working tree when it is safe to do so. 79 80 * "git send-email" used to prompt for the sender address, even when 81 the committer identity is well specified (e.g. via user.name and 82 user.email configuration variables). The command no longer gives 83 this prompt when not necessary. 84 85 * "git send-email" did not allow non-address garbage strings to 86 appear after addresses on Cc: lines in the patch files (and when 87 told to pick them up to find more recipients), e.g. 88 89 Cc: Stable Kernel <stable@k.org> # for v3.2 and up 90 91 The command now strips " # for v3.2 and up" part before adding the 92 remainder of this line to the list of recipients. 93 94 * "git submodule add" learned to add a new submodule at the same 95 path as the path where an unrelated submodule was bound to in an 96 existing revision via the "--name" option. 97 98 * "git submodule sync" learned the "--recursive" option. 99 100 * "diff.submodule" configuration variable can be used to give custom 101 default value to the "git diff --submodule" option. 102 103 * "git symbolic-ref" learned the "-d $symref" option to delete the 104 named symbolic ref, which is more intuitive way to spell it than 105 "update-ref -d --no-deref $symref". 106 107 108Foreign Interface 109 110 * "git cvsimport" can be told to record timezones (other than GMT) 111 per-author via its author info file. 112 113 * The remote helper interface to interact with subversion 114 repositories (one of the GSoC 2012 projects) has been merged. 115 116 * A new remote-helper interface for Mercurial has been added to 117 contrib/remote-helpers. 118 119 * The documentation for git(1) was pointing at a page at an external 120 site for the list of authors that no longer existed. The link has 121 been updated to point at an alternative site. 122 123 124Performance, Internal Implementation, etc. 125 126 * Compilation on Cygwin with newer header files are supported now. 127 128 * A couple of low-level implementation updates on MinGW. 129 130 * The logic to generate the initial advertisement from "upload-pack" 131 (i.e. what is invoked by "git fetch" on the other side of the 132 connection) to list what refs are available in the repository has 133 been optimized. 134 135 * The logic to find set of attributes that match a given path has 136 been optimized. 137 138 * Use preloadindex in "git diff-index" and "git update-index", which 139 has a nice speedup on systems with slow stat calls (and even on 140 Linux). 141 142 143Also contains minor documentation updates and code clean-ups. 144 145 146Fixes since v1.8.0 147------------------ 148 149Unless otherwise noted, all the fixes since v1.8.0 in the maintenance 150track are contained in this release (see release notes to them for 151details). 152 153 * The configuration parser had an unnecessary hardcoded limit on 154 variable names that was not checked consistently. 155 156 * The "say" function in the test scaffolding incorrectly allowed 157 "echo" to interpret "\a" as if it were a C-string asking for a 158 BEL output. 159 160 * "git mergetool" feeds /dev/null as a common ancestor when dealing 161 with an add/add conflict, but p4merge backend cannot handle 162 it. Work it around by passing a temporary empty file. 163 164 * "git log -F -E --grep='<ere>'" failed to use the given <ere> 165 pattern as extended regular expression, and instead looked for the 166 string literally. 167 168 * "git grep -e pattern <tree>" asked the attribute system to read 169 "<tree>:.gitattributes" file in the working tree, which was 170 nonsense. 171 172 * A symbolic ref refs/heads/SYM was not correctly removed with "git 173 branch -d SYM"; the command removed the ref pointed by SYM 174 instead. 175 176 * Update "remote tracking branch" in the documentation to 177 "remote-tracking branch". 178 179 * "git pull --rebase" run while the HEAD is detached tried to find 180 the upstream branch of the detached HEAD (which by definition 181 does not exist) and emitted unnecessary error messages. 182 183 * The refs/replace hierarchy was not mentioned in the 184 repository-layout docs. 185 186 * Various rfc2047 quoting issues around a non-ASCII name on the 187 From: line in the output from format-patch have been corrected. 188 189 * Sometimes curl_multi_timeout() function suggested a wrong timeout 190 value when there is no file descriptor to wait on and the http 191 transport ended up sleeping for minutes in select(2) system call. 192 A workaround has been added for this. 193 194 * For a fetch refspec (or the result of applying wildcard on one), 195 we always want the RHS to map to something inside "refs/" 196 hierarchy, but the logic to check it was not exactly right. 197 (merge 5c08c1f jc/maint-fetch-tighten-refname-check later to maint). 198 199 * "git diff -G<pattern>" did not honor textconv filter when looking 200 for changes. 201 202 * Some HTTP servers ask for auth only during the actual packing phase 203 (not in ls-remote phase); this is not really a recommended 204 configuration, but the clients used to fail to authenticate with 205 such servers. 206 (merge 2e736fd jk/maint-http-half-auth-fetch later to maint). 207 208 * "git p4" used to try expanding malformed "$keyword$" that spans 209 across multiple lines. 210 211 * Syntax highlighting in "gitweb" was not quite working. 212 213 * RSS feed from "gitweb" had a xss hole in its title output. 214 215 * "git config --path $key" segfaulted on "[section] key" (a boolean 216 "true" spelled without "=", not "[section] key = true"). 217 218 * "git checkout -b foo" while on an unborn branch did not say 219 "Switched to a new branch 'foo'" like other cases. 220 221 * Various codepaths have workaround for a common misconfiguration to 222 spell "UTF-8" as "utf8", but it was not used uniformly. Most 223 notably, mailinfo (which is used by "git am") lacked this support. 224 225 * We failed to mention a file without any content change but whose 226 permission bit was modified, or (worse yet) a new file without any 227 content in the "git diff --stat" output. 228 229 * When "--stat-count" hides a diffstat for binary contents, the total 230 number of added and removed lines at the bottom was computed 231 incorrectly. 232 233 * When "--stat-count" hides a diffstat for unmerged paths, the total 234 number of affected files at the bottom of the "diff --stat" output 235 was computed incorrectly. 236 237 * "diff --shortstat" miscounted the total number of affected files 238 when there were unmerged paths. 239 240 * "update-ref -d --deref SYM" to delete a ref through a symbolic ref 241 that points to it did not remove it correctly.