1 Git v2.14.2 Release Notes
2 =========================
7 * Because recent Git for Windows do come with a real msgfmt, the
8 build procedure for git-gui has been updated to use it instead of a
9 hand-rolled substitute.
11 * "%C(color name)" in the pretty print format always produced ANSI
12 color escape codes, which was an early design mistake. They now
13 honor the configuration (e.g. "color.ui = never") and also tty-ness
16 * The http.{sslkey,sslCert} configuration variables are to be
17 interpreted as a pathname that honors "~[username]/" prefix, but
18 weren't, which has been fixed.
20 * Numerous bugs in walking of reflogs via "log -g" and friends have
23 * "git commit" when seeing an totally empty message said "you did not
24 edit the message", which is clearly wrong. The message has been
27 * When a directory is not readable, "gitweb" fails to build the
28 project list. Work this around by skipping such a directory.
30 * A recently added test for the "credential-cache" helper revealed
31 that EOF detection done around the time the connection to the cache
32 daemon is torn down were flaky. This was fixed by reacting to
33 ECONNRESET and behaving as if we got an EOF.
35 * Some versions of GnuPG fail to kill gpg-agent it auto-spawned
36 and such a left-over agent can interfere with a test. Work it
37 around by attempting to kill one before starting a new test.
39 * "git log --tag=no-such-tag" showed log starting from HEAD, which
40 has been fixed---it now shows nothing.
42 * The "tag.pager" configuration variable was useless for those who
43 actually create tag objects, as it interfered with the use of an
44 editor. A new mechanism has been introduced for commands to enable
45 pager depending on what operation is being carried out to fix this,
46 and then "git tag -l" is made to run pager by default.
48 * "git push --recurse-submodules $there HEAD:$target" was not
49 propagated down to the submodules, but now it is.
51 * Commands like "git rebase" accepted the --rerere-autoupdate option
52 from the command line, but did not always use it. This has been
55 * "git clone --recurse-submodules --quiet" did not pass the quiet
56 option down to submodules.
58 * "git am -s" has been taught that some input may end with a trailer
59 block that is not Signed-off-by: and it should refrain from adding
60 an extra blank line before adding a new sign-off in such a case.
62 * "git svn" used with "--localtime" option did not compute the tz
63 offset for the timestamp in question and instead always used the
64 current time, which has been corrected.
66 * Memory leaks in a few error codepaths have been plugged.
68 * bash 4.4 or newer gave a warning on NUL byte in command
69 substitution done in "git stash"; this has been squelched.
71 * "git grep -L" and "git grep --quiet -L" reported different exit
72 codes; this has been corrected.
74 * When handshake with a subprocess filter notices that the process
75 asked for an unknown capability, Git did not report what program
76 the offending subprocess was running. This has been corrected.
78 * "git apply" that is used as a better "patch -p1" failed to apply a
79 taken from a file with CRLF line endings to a file with CRLF line
80 endings. The root cause was because it misused convert_to_git()
81 that tried to do "safe-crlf" processing by looking at the index
82 entry at the same path, which is a nonsense---in that mode, "apply"
83 is not working on the data in (or derived from) the index at all.
86 * Killing "git merge --edit" before the editor returns control left
87 the repository in a state with MERGE_MSG but without MERGE_HEAD,
88 which incorrectly tells the subsequent "git commit" that there was
89 a squash merge in progress. This has been fixed.
91 * "git archive" did not work well with pathspecs and the
92 export-ignore attribute.
94 Also contains various documentation updates and code clean-ups.