1 Git v2.10.2 Release Notes
2 =========================
7 * The code that parses the format parameter of for-each-ref command
8 has seen a micro-optimization.
10 * The "graph" API used in "git log --graph" miscounted the number of
11 output columns consumed so far when drawing a padding line, which
12 has been fixed; this did not affect any existing code as nobody
13 tried to write anything after the padding on such a line, though.
15 * Almost everybody uses DEFAULT_ABBREV to refer to the default
16 setting for the abbreviation, but "git blame" peeked into
17 underlying variable bypassing the macro for no good reason.
19 * Doc update to clarify what "log -3 --reverse" does.
21 * An author name, that spelled a backslash-quoted double quote in the
22 human readable part "My \"double quoted\" name", was not unquoted
23 correctly while applying a patch from a piece of e-mail.
25 * The original command line syntax for "git merge", which was "git
26 merge <msg> HEAD <parent>...", has been deprecated for quite some
27 time, and "git gui" was the last in-tree user of the syntax. This
28 is finally fixed, so that we can move forward with the deprecation.
30 * Codepaths that read from an on-disk loose object were too loose in
31 validating what they are reading is a proper object file and
32 sometimes read past the data they read from the disk, which has
33 been corrected. H/t to Gustavo Grieco for reporting.
35 * "git worktree", even though it used the default_abbrev setting that
36 ought to be affected by core.abbrev configuration variable, ignored
37 the variable setting. The command has been taught to read the
38 default set of configuration variables to correct this.
40 * A low-level function verify_packfile() was meant to show errors
41 that were detected without dying itself, but under some conditions
42 it didn't and died instead, which has been fixed.
44 * When "git fetch" tries to find where the history of the repository
45 it runs in has diverged from what the other side has, it has a
46 mechanism to avoid digging too deep into irrelevant side branches.
47 This however did not work well over the "smart-http" transport due
48 to a design bug, which has been fixed.
50 * When we started cURL to talk to imap server when a new enough
51 version of cURL library is available, we forgot to explicitly add
52 imap(s):// before the destination. To some folks, that didn't work
53 and the library tried to make HTTP(s) requests instead.
55 * The ./configure script generated from configure.ac was taught how
56 to detect support of SSL by libcurl better.
58 * http.emptyauth configuration is a way to allow an empty username to
59 pass when attempting to authenticate using mechanisms like
60 Kerberos. We took an unspecified (NULL) username and sent ":"
61 (i.e. no username, no password) to CURLOPT_USERPWD, but did not do
62 the same when the username is explicitly set to an empty string.
64 * "git clone" of a local repository can be done at the filesystem
65 level, but the codepath did not check errors while copying and
66 adjusting the file that lists alternate object stores.
68 * Documentation for "git commit" was updated to clarify that "commit
69 -p <paths>" adds to the current contents of the index to come up
72 * A stray symbolic link in $GIT_DIR/refs/ directory could make name
73 resolution loop forever, which has been corrected.
75 * The "submodule.<name>.path" stored in .gitmodules is never copied
76 to .git/config and such a key in .git/config has no meaning, but
77 the documentation described it and submodule.<name>.url next to
78 each other as if both belong to .git/config. This has been fixed.
80 * Recent git allows submodule.<name>.branch to use a special token
81 "." instead of the branch name; the documentation has been updated
84 * In a worktree connected to a repository elsewhere, created via "git
85 worktree", "git checkout" attempts to protect users from confusion
86 by refusing to check out a branch that is already checked out in
87 another worktree. However, this also prevented checking out a
88 branch, which is designated as the primary branch of a bare
89 reopsitory, in a worktree that is connected to the bare
90 repository. The check has been corrected to allow it.
92 * "git rebase" immediately after "git clone" failed to find the fork
93 point from the upstream.
95 * When fetching from a remote that has many tags that are irrelevant
96 to branches we are following, we used to waste way too many cycles
97 when checking if the object pointed at by a tag (that we are not
98 going to fetch!) exists in our repository too carefully.
100 * The Travis CI configuration we ship ran the tests with --verbose
101 option but this risks non-TAP output that happens to be "ok" to be
102 misinterpreted as TAP signalling a test that passed. This resulted
103 in unnecessary failure. This has been corrected by introducing a
104 new mode to run our tests in the test harness to send the verbose
105 output separately to the log file.
107 * Some AsciiDoc formatter mishandles a displayed illustration with
108 tabs in it. Adjust a few of them in merge-base documentation to
111 Also contains minor documentation updates and code clean-ups.