1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
|
Git v1.8.3 Release Notes
========================
Backward compatibility notes (for Git 2.0)
------------------------------------------
When "git push [$there]" does not say what to push, we have used the
traditional "matching" semantics so far (all your branches were sent
to the remote as long as there already are branches of the same name
over there). In Git 2.0, the default will change to the "simple"
semantics that pushes the current branch to the branch with the same
name, only when the current branch is set to integrate with that
remote branch. There is a user preference configuration variable
"push.default" to change this. If you are an old-timer who is used
to the "matching" semantics, you can set it to "matching" to keep the
traditional behaviour. If you want to live in the future early,
you can set it to "simple" today without waiting for Git 2.0.
When "git add -u" and "git add -A", that does not specify what paths
to add on the command line is run from inside a subdirectory, these
commands will operate on the entire tree in Git 2.0 for consistency
with "git commit -a" and other commands. Because there will be no
mechanism to make "git add -u" behave as if "git add -u .", it is
important for those who are used to "git add -u" (without pathspec)
updating the index only for paths in the current subdirectory to start
training their fingers to explicitly say "git add -u ." when they mean
it before Git 2.0 comes.
Updates since v1.8.2
--------------------
UI, Workflows & Features
* "git help" learned "-g" option to show the list of guides just like
list of commands are given with "-a".
* A triangular "pull from one place, push to another place" workflow
is supported better by new remote.pushdefault (overrides the
"origin" thing) and branch.*.pushremote (overrides the
branch.*.remote) configuration variables.
* "git status" learned to report that you are in the middle of a
revert session, just like it does for a cherry-pick and a bisect
session.
* The handling by "git branch --set-upstream-to" against various forms
of erroneous inputs was suboptimal and has been improved.
* When the interactive access to git-shell is not enabled, it issues
a message meant to help the system administrator to enable it.
An explicit way to help the end users who connect to the service by
issuing custom messages to refuse such an access has been added.
* In addition to the case where the user edits the log message with
the "e)dit" option of "am -i", replace the "Applying: this patch"
message with the final log message contents after applymsg hook
munges it.
* "git status" suggests users to look into using --untracked=no option
when it takes too long.
* "git status" shows a bit more information to "git status" during a
rebase/bisect session.
* "git fetch" learned to fetch a commit at the tip of an unadvertised
ref by specifying a raw object name from the command line when the
server side supports this feature.
* "git count-objects -v" learned to report leftover temporary
packfiles and other garbage in the object store.
* A new read-only credential helper (in contrib/) to interact with
the .netrc/.authinfo files has been added.
* "git send-email" can be used with the credential helper system.
* There was no Porcelain way to say "I no longer am interested in
this submodule", once you express your interest in a submodule with
"submodule init". "submodule deinit" is the way to do so.
* "git pull --rebase" learned to pass "-v/-q" options to underlying
"git rebase".
* The new "--follow-tags" option tells "git push" to push relevant
annotated tags when pushing branches out.
* "git merge" and "git pull" can optionally be told to inspect and
reject when merging a commit that does not carry a trusted GPG
signature.
* "git mergetool" now feeds files to the "p4merge" backend in the
order that matches the p4 convention, where "theirs" is usually
shown on the left side, which is the opposite from other backend
expects.
* "show/log" now honors gpg.program configuration just like other
parts of the code that use GnuPG.
* "git log" that shows the difference between the parent and the
child has been optimized somewhat.
* "git difftool" allows the user to write into the temporary files
being shown; if the user makes changes to the working tree at the
same time, one of the changes has to be lost in such a case, but it
tells the user what happened and refrains from overwriting the copy
in the working tree.
* There was no good way to ask "I have a random string that came from
outside world. I want to turn it into a 40-hex object name while
making sure such an object exists". A new peeling suffix ^{object}
can be used for that purpose, together with "rev-parse --verify".
Performance, Internal Implementation, etc.
* Updates for building under msvc.
* The logic to coalesce the same lines removed from the parents in
the output from "diff -c/--cc" has been updated, but with an O(n^2)
complexity, so this might turn out to be undesirable.
* The code to enforce permission bits on files in $GIT_DIR/ for
shared repositories have been simplified.
* A few codepaths knew how much data they need to put in the
hashtables they use upfront, but still started from a small table
repeatedly growing and rehashing.
* The API to walk reflog entries from the latest to older, which was
necessary for operations such as "git checkout -", was cumbersome
to use correctly and also inefficient.
* Codepaths that inspect log-message-to-be and decide when to add a
new Signed-off-by line in various commands have been consolidated.
* The pkt-line API, implementation and its callers have been cleaned
up to make them more robust.
* Cygwin port has a faster-but-lying lstat(2) emulation whose
incorrectness does not matter in practice except for a few
codepaths, and setting permission bits to directories is a codepath
that needs to use a more correct one.
* "git checkout" had repeated pathspec matches on the same paths,
which have been consolidated. Also a bug in "git checkout dir/"
that is started from an unmerged index has been fixed.
* A few bugfixes to "git rerere" working on corner case merge
conflicts have been applied.
Also contains minor documentation updates and code clean-ups.
Fixes since v1.8.2
------------------
Unless otherwise noted, all the fixes since v1.8.2 in the maintenance
track are contained in this release (see release notes to them for
details).
* "git log -S/-G" started paying attention to textconv filter, but
there was no way to disable this. Make it honor --no-textconv
option.
(merge 61690bf sr/log-SG-no-textconv later to maint).
* When used with "-d temporary-directory" option, "git filter-branch"
failed to come back to the original working tree to perform the
final clean-up procedure.
(merge 9727601 jk/filter-branch-come-back-to-original later to maint).
* "git merge $(git rev-parse v1.8.2)" behaved quite differently from
"git merge v1.8.2", as if v1.8.2 were written as v1.8.2^0 and did
not pay much attention to the annotated tag payload. Make the code
notice the type of the tag object, in addition to the dwim_ref()
based classification the current code uses (i.e. the name appears
in refs/tags/) to decide when to special case merging of tags.
(merge a38d3d7 jc/merge-tag-object later to maint).
* Fix 1.8.1.x regression that stopped matching "dir" (without
trailing slash) to a directory "dir".
(merge efa5f82 jc/directory-attrs-regression-fix later to maint-1.8.1).
* "git apply --whitespace=fix" was not prepared to see a line getting
longer after fixing whitespaces (e.g. tab-in-indent aka Python).
(merge 329b26e jc/apply-ws-fix-tab-in-indent later to maint-1.8.1).
* The prompt string generator (in contrib/completion/) did not notice
when we are in a middle of a "git revert" session.
(merge 3ee4452 rr/prompt-revert-head later to maint).
* "submodule summary --summary-limit" option did not support
"--option=value" form.
(merge 862ae6c rs/submodule-summary-limit later to maint).
* "index-pack --fix-thin" used an uninitialized value to compute
delta depths of objects it appends to the resulting pack.
(merge 57165db jk/index-pack-correct-depth-fix later to maint).
* "index-pack --verify-stat" used a few counters outside protection
of mutex, possibly showing incorrect numbers.
(merge 8f82aad nd/index-pack-threaded-fixes later to maint).
* The code to keep track of what directory names are known to Git on
platforms with case insensitive filesystems can get confused upon a
hash collision between these pathnames and looped forever.
* Annotated tags outside refs/tags/ hierarchy were not advertised
correctly to the ls-remote and fetch with recent version of Git.
* Recent optimization broke shallow clones.
(merge f59de5d jk/peel-ref later to maint).
* "git cmd -- ':(top'" was not diagnosed as an invalid syntax, and
instead the parser kept reading beyond the end of the string.
* "git tag -f <tag>" always said "Updated tag '<tag>'" even when
creating a new tag (i.e. not overwriting nor updating).
* "git p4" did not behave well when the path to the root of the P4
client was not its real path.
(merge bbd8486 pw/p4-symlinked-root later to maint).
* "git archive" reports a failure when asked to create an archive out
of an empty tree. It would be more intuitive to give an empty
archive back in such a case.
* When "format-patch" quoted a non-ascii strings on the header files,
it incorrectly applied rfc2047 and chopped a single character in
the middle of it.
* An aliased command spawned from a bare repository that does not say
it is bare with "core.bare = yes" is treated as non-bare by mistake.
* In "git reflog expire", REACHABLE bit was not cleared from the
correct objects.
* The logic used by "git diff -M --stat" to shorten the names of
files before and after a rename did not work correctly when the
common prefix and suffix between the two filenames overlapped.
* The "--match=<pattern>" option of "git describe", when used with
"--all" to allow refs that are not annotated tags to be used as a
base of description, did not restrict the output from the command
to those that match the given pattern.
* Clarify in the documentation "what" gets pushed to "where" when the
command line to "git push" does not say these explicitly.
* The "--color=<when>" argument to the commands in the diff family
was described poorly.
* The arguments given to pre-rebase hook were not documented.
* The v4 index format was not documented.
* The "--match=<pattern>" argument "git describe" takes uses glob
pattern but it wasn't obvious from the documentation.
* Some sources failed to compile on systems that lack NI_MAXHOST in
their system header (e.g. z/OS).
* Add an example use of "--env-filter" in "filter-branch"
documentation.
* "git bundle verify" did not say "records a complete history" for a
bundle that does not have any prerequisites.
* In the v1.8.0 era, we changed symbols that do not have to be global
to file scope static, but a few functions in graph.c were used by
CGit from sideways bypassing the entry points of the API the
in-tree users use.
* "git update-index -h" did not do the usual "-h(elp)" thing.
* "git index-pack" had a buffer-overflow while preparing an
informational message when the translated version of it was too
long.
* 'git commit -m "$msg"' used to add an extra newline even when
$msg already ended with one.
* The SSL peer verification done by "git imap-send" did not ask for
Server Name Indication (RFC 4366), failing to connect SSL/TLS
sites that serve multiple hostnames on a single IP.
* perl/Git.pm::cat_blob slurped everything in core only to write it
out to a file descriptor, which was not a very smart thing to do.
* "git branch" did not bother to check nonsense command line
parameters and issue errors in many cases.
* Verification of signed tags were not done correctly when not in C
or en/US locale.
* Some platforms and users spell UTF-8 differently; retry with the
most official "UTF-8" when the system does not understand the
user-supplied encoding name that are the common alternative
spellings of UTF-8.
* When export-subst is used, "zip" output recorded incorrect
size of the file.
* "git am $maildir/" applied messages in an unexpected order; sort
filenames read from the maildir/ in a way that is more likely to
sort messages in the order the writing MUA meant to, by sorting
numeric segment in numeric order and non-numeric segment in
alphabetical order.
* "git submodule update", when recursed into sub-submodules, did not
accumulate the prefix paths.
|
