| blob | ac5c3fafe59be32e5f276ad4276a5f9d6305fe85 |
1 Git v1.8.5 Release Notes
2 ========================
4 Backward compatibility notes (for Git 2.0)
5 ------------------------------------------
7 When "git push [$there]" does not say what to push, we have used the
8 traditional "matching" semantics so far (all your branches were sent
9 to the remote as long as there already are branches of the same name
10 over there). In Git 2.0, the default will change to the "simple"
11 semantics that pushes:
13 - only the current branch to the branch with the same name, and only
14 when the current branch is set to integrate with that remote
15 branch, if you are pushing to the same remote as you fetch from; or
17 - only the current branch to the branch with the same name, if you
18 are pushing to a remote that is not where you usually fetch from.
20 Use the user preference configuration variable "push.default" to
21 change this. If you are an old-timer who is used to the "matching"
22 semantics, you can set the variable to "matching" to keep the
23 traditional behaviour. If you want to live in the future early, you
24 can set it to "simple" today without waiting for Git 2.0.
26 When "git add -u" (and "git add -A") is run inside a subdirectory and
27 does not specify which paths to add on the command line, it
28 will operate on the entire tree in Git 2.0 for consistency
29 with "git commit -a" and other commands. There will be no
30 mechanism to make plain "git add -u" behave like "git add -u .".
31 Current users of "git add -u" (without a pathspec) should start
32 training their fingers to explicitly say "git add -u ."
33 before Git 2.0 comes. A warning is issued when these commands are
34 run without a pathspec and when you have local changes outside the
35 current directory, because the behaviour in Git 2.0 will be different
36 from today's version in such a situation.
38 In Git 2.0, "git add <path>" will behave as "git add -A <path>", so
39 that "git add dir/" will notice paths you removed from the directory
40 and record the removal. Versions before Git 2.0, including this
41 release, will keep ignoring removals, but the users who rely on this
42 behaviour are encouraged to start using "git add --ignore-removal <path>"
43 now before 2.0 is released.
46 Updates since v1.8.4
47 --------------------
49 Foreign interfaces, subsystems and ports.
51 * "git-svn" used with SVN 1.8.0 when talking over https:// connection
52 dumped core due to a bug in the serf library that SVN uses. Work
53 it around on our side, even though the SVN side is being fixed.
55 * On MacOS X, we detected if the filesystem needs the "pre-composed
56 unicode strings" workaround, but did not automatically enable it.
57 Now we do.
59 * remote-hg remote helper misbehaved when interacting with a local Hg
60 repository relative to the home directory, e.g. "clone hg::~/there".
62 * imap-send ported to OS X uses Apple's security framework instead of
63 OpenSSL one.
65 * Subversion 1.8.0 that was recently released breaks older subversion
66 clients coming over http/https in various ways.
68 * "git fast-import" treats an empty path given to "ls" as the root of
69 the tree.
72 UI, Workflows & Features
74 * Instead of typing four capital letters "HEAD", you can say "@" now,
75 e.g. "git log @".
77 * "git check-ignore" follows the same rule as "git add" and "git
78 status" in that the ignore/exclude mechanism does not take effect
79 on paths that are already tracked. With "--no-index" option, it
80 can be used to diagnose which paths that should have been ignored
81 have been mistakenly added to the index.
83 * Some irrelevant "advice" messages that are shared with "git status"
84 output have been removed from the commit log template.
86 * "update-refs" learnt a "--stdin" option to read multiple update
87 requests and perform them in an all-or-none fashion.
89 * Just like "make -C <directory>", "git -C <directory> ..." tells Git
90 to go there before doing anything else.
92 * Just like "git checkout -" knows to check out and "git merge -"
93 knows to merge the branch you were previously on, "git cherry-pick"
94 now understands "git cherry-pick -" to pick from the previous
95 branch.
97 * "git status" now omits the prefix to make its output a comment in a
98 commit log editor, which is not necessary for human consumption.
100 * Make "foo^{tag}" to peel a tag to itself, i.e. no-op., and fail if
101 "foo" is not a tag. "git rev-parse --verify v1.0^{tag}" would be
102 a more convenient way to say "test $(git cat-file -t v1.0) = tag".
104 * "git branch -v -v" (and "git status") did not distinguish among a
105 branch that does not build on any other branch, a branch that is in
106 sync with the branch it builds on, and a branch that is configured
107 to build on some other branch that no longer exists.
109 * A packfile that stores the same object more than once is broken and
110 will be rejected by "git index-pack" that is run when receiving
111 data over the wire.
113 * Earlier we started rejecting an attempt to add 0{40} object name to
114 the index and to tree objects, but it sometimes is necessary to
115 allow so to be able to use tools like filter-branch to correct such
116 broken tree objects. "filter-branch" can again be used to to do
117 so.
119 * "git config" did not provide a way to set or access numbers larger
120 than a native "int" on the platform; it now provides 64-bit signed
121 integers on all platforms.
123 * "git pull --rebase" always chose to do the bog-standard flattening
124 rebase. You can tell it to run "rebase --preserve-merges" by
125 setting "pull.rebase" configuration to "preserve".
127 * "git push --no-thin" actually disables the "thin pack transfer"
128 optimization.
130 * Magic pathspecs like ":(icase)makefile" that matches both
131 Makefile and makefile can be used in more places.
133 * The "http.*" variables can now be specified per URL that the
134 configuration applies. For example,
136 [http]
137 sslVerify = true
138 [http "https://weak.example.com/"]
139 sslVerify = false
141 would flip http.sslVerify off only when talking to that specified
142 site.
144 * "git mv A B" when moving a submodule A has been taught to
145 relocate its working tree and to adjust the paths in the
146 .gitmodules file.
148 * "git blame" can now take more than one -L option to discover the
149 origin of multiple blocks of the lines.
151 * The http transport clients can optionally ask to save cookies
152 with http.savecookies configuration variable.
154 * "git push" learned a more fine grained control over a blunt
155 "--force" when requesting a non-fast-forward update with the
156 "--force-with-lease=<refname>:<expected object name>" option.
158 * "git diff --diff-filter=<classes of changes>" can now take
159 lowercase letters (e.g. "--diff-filter=d") to mean "show
160 everything but these classes". "git diff-files -q" is now a
161 deprecated synonym for "git diff-files --diff-filter=d".
163 * "git fetch" (hence "git pull" as well) learned to check
164 "fetch.prune" and "remote.*.prune" configuration variables and
165 to behave as if the "--prune" command line option was given.
167 * "git check-ignore -z" applied the NUL termination to both its input
168 (with --stdin) and its output, but "git check-attr -z" ignored the
169 option on the output side. Make both honor -z on the input and
170 output side the same way.
172 * "git whatchanged" may still be used by old timers, but mention of
173 it in documents meant for new users will only waste readers' time
174 wonderig what the difference is between it and "git log". Make it
175 less prominent in the general part of the documentation and explain
176 that it is merely a "git log" with different default behaviour in
177 its own document.
180 Performance, Internal Implementation, etc.
182 * If a build-time fallback is set to "cat" instead of "less", we
183 should apply the same "no subprocess or pipe" optimization as we
184 apply to user-supplied GIT_PAGER=cat.
186 * Many commands use --dashed-option as a operation mode selector
187 (e.g. "git tag --delete") that the user can use at most one
188 (e.g. "git tag --delete --verify" is a nonsense) and you cannot
189 negate (e.g. "git tag --no-delete" is a nonsense). parse-options
190 API learned a new OPT_CMDMODE macro to make it easier to implement
191 such a set of options.
193 * OPT_BOOLEAN() in parse-options API was misdesigned to be "counting
194 up" but many subcommands expect it to behave as "on/off". Update
195 them to use OPT_BOOL() which is a proper boolean.
197 * "git gc" exits early without doing a double-work when it detects
198 that another instance of itself is already running.
200 * Under memory pressure and/or file descriptor pressure, we used to
201 close pack windows that are not used and also closed filehandle to
202 an open but unused packfiles. These are now controlled separately
203 to better cope with the load.
205 Also contains various documentation updates and code clean-ups.
208 Fixes since v1.8.4
209 ------------------
211 Unless otherwise noted, all the fixes since v1.8.4 in the maintenance
212 track are contained in this release (see release notes to them for
213 details).
215 * When running "fetch -q", a long silence while the sender side
216 computes the set of objects to send can be mistaken by proxies as
217 dropped connection. The server side has been taught to send a
218 small empty messages to keep the connection alive.
219 (merge 115dedd jk/upload-pack-keepalive later to maint).
221 * "git rebase" had a portability regression in v1.8.4 to trigger a
222 bug in some BSD shell implementations.
223 (merge 99855dd mm/rebase-continue-freebsd-WB later to maint).
225 * "git branch --track" had a minor regression in v1.8.3.2 and later
226 that made it impossible to base your local work on anything but a
227 local branch of the upstream repository you are tracking from.
228 (merge b0f49ff jh/checkout-auto-tracking later to maint).
230 * When the webserver responds with "405 Method Not Allowed", "git
231 http-backend" should tell the client what methods are allowed with
232 the "Allow" header.
233 (merge 9247be0 bc/http-backend-allow-405 later to maint).
235 * When there is no sufficient overlap between old and new history
236 during a "git fetch" into a shallow repository, objects that the
237 sending side knows the receiving end has were unnecessarily sent.
238 (merge f21d2a7 nd/fetch-into-shallow later to maint).
240 * "git cvsserver" computed the permission mode bits incorrectly for
241 executable files.
242 (merge 1b48d56 jc/cvsserver-perm-bit-fix later to maint).
244 * When send-email comes up with an error message to die with upon
245 failure to start an SSL session, it tried to read the error string
246 from a wrong place.
247 (merge 6cb0c88 bc/send-email-ssl-die-message-fix later to maint).
249 * The implementation of "add -i" has a crippling code to work around
250 ActiveState Perl limitation but it by mistake also triggered on Git
251 for Windows where MSYS perl is used.
252 (merge df17e77 js/add-i-mingw later to maint).
254 * We made sure that we notice the user-supplied GIT_DIR is actually a
255 gitfile, but did not do the same when the default ".git" is a
256 gitfile.
257 (merge 487a2b7 nd/git-dir-pointing-at-gitfile later to maint).
259 * When an object is not found after checking the packfiles and then
260 loose object directory, read_sha1_file() re-checks the packfiles to
261 prevent racing with a concurrent repacker; teach the same logic to
262 has_sha1_file().
263 (merge 45e8a74 jk/has-sha1-file-retry-packed later to maint).
265 * "git commit --author=$name", when $name is not in the canonical
266 "A. U. Thor <au.thor@example.xz>" format, looks for a matching name
267 from existing history, but did not consult mailmap to grab the
268 preferred author name.
269 (merge ea16794 ap/commit-author-mailmap later to maint).
271 * "git ls-files -k" needs to crawl only the part of the working tree
272 that may overlap the paths in the index to find killed files, but
273 shared code with the logic to find all the untracked files, which
274 made it unnecessarily inefficient.
275 (merge 680be04 jc/ls-files-killed-optim later to maint).
277 * The commit object names in the insn sheet that was prepared at the
278 beginning of "rebase -i" session can become ambiguous as the
279 rebasing progresses and the repository gains more commits. Make
280 sure the internal record is kept with full 40-hex object names.
281 (merge 75c6976 es/rebase-i-no-abbrev later to maint).
283 * "git rebase --preserve-merges" internally used the merge machinery
284 and as a side effect, left merge summary message in the log, but
285 when rebasing, there should not be a need for merge summary.
286 (merge a9f739c rt/rebase-p-no-merge-summary later to maint).
288 * A call to xread() was used without a loop around to cope with short
289 read in the codepath to stream new contents to a pack.
290 (merge e92527c js/xread-in-full later to maint).
292 * "git rebase -i" forgot that the comment character can be
293 configurable while reading its insn sheet.
294 (merge 7bca7af es/rebase-i-respect-core-commentchar later to maint).
296 * The mailmap support code read past the allocated buffer when the
297 mailmap file ended with an incomplete line.
298 (merge f972a16 jk/mailmap-incomplete-line later to maint).
300 * We used to send a large request to read(2)/write(2) as a single
301 system call, which was bad from the latency point of view when
302 the operation needs to be killed, and also triggered an error on
303 broken 64-bit systems that refuse to take more than 2GB read or
304 write in one go.
305 (merge a487916 sp/clip-read-write-to-8mb later to maint).
307 * "git fetch" that auto-followed tags incorrectly reused the
308 connection with Git-aware transport helper (like the sample "ext::"
309 helper shipped with Git).
310 (merge 0f73f8b jc/transport-do-not-use-connect-twice-in-fetch later to maint).
312 * "git log --full-diff -- <pathspec>" showed a huge diff for paths
313 outside the given <pathspec> for each commit, instead of showing
314 the change relative to the parent of the commit. "git reflog -p"
315 had a similar problem.
316 (merge 838f9a1 tr/log-full-diff-keep-true-parents later to maint).
318 * Setting submodule.*.path configuration variable to true (without
319 giving "= value") caused Git to segfault.
320 (merge 4b05440 jl/some-submodule-config-are-not-boolean later to maint).
322 * "git rebase -i" (there could be others, as the root cause is pretty
323 generic) fed a random, data dependeant string to 'echo' and
324 expects it to come out literally, corrupting its error message.
325 (merge 89b0230 mm/no-shell-escape-in-die-message later to maint).
327 * Some people still use rather old versions of bash, which cannot
328 grok some constructs like 'printf -v varname' the prompt and
329 completion code started to use recently.
330 (merge a44aa69 bc/completion-for-bash-3.0 later to maint).
332 * Code to read configuration from a blob object did not compile on
333 platforms with fgetc() etc. implemented as macros.
334 (merge 49d6cfa hv/config-from-blob later to maint-1.8.3).
336 * The recent "short-cut clone connectivity check" topic broke a
337 shallow repository when a fetch operation tries to auto-follow tags.
338 (merge 6da8bdc nd/fetch-pack-shallow-fix later to maint-1.8.3).