git commit 中的change-id

git commit 中的change-id

为什么要使用change-id

保证已经提交审核的修订通过审核入库后，被别的分支 cherry-pick 后再推送至服务器时不会产生新的重复的评审任务。Gerrit 设计了一套方法，即要求每个提交包含唯一的 Change-Id，这个 Change-Id 因为出现在日志中，当执行 cherry-pick 时也会保持，Gerrit 一旦发现新的提交包含了已经处理过的 Change-Id ，就不再为该修订创建新的评审任务和 task-id，而直接将提交入库。

解决“ERROR：missing Change-Id in commit message”

在执行git push origin Head:refs/for/xxxx时有时会报出这样的错误，也就是在commit Message仅仅包含如“feature：xxxx”等title这样的信息而缺少Change-Id。

手动解决

执行git commit --amend，在title下面空出一行后，将错误提示中的change-id复制到Message中。再push。注意，一定要空出一行，否则Git会把其作为title的一部分处理

利用commit-msg hook自动生成change-id

将如下的的hook脚本复制到Git项目中.git/hooks下，并命名为“commit-msg”。一般情况 .git/hooks 下会包含一个叫“commit-msg.sample”的文件，可以把它删除。添加完成后，执行chmod u+x .git/hooks/commit-msg保证hook可执行。

commit-msg hook脚本

#!/bin/sh
# From Gerrit Code Review 2.6
#
# Part of Gerrit Code Review (http://code.google.com/p/gerrit/)
#
# Copyright (C) 2009 The Android Open Source Project
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
#

unset GREP_OPTIONS

CHANGE_ID_AFTER="Bug|Issue"
MSG="$1"

# Check for, and add if missing, a unique Change-Id
#
add_ChangeId() {
clean_message=`sed -e '
/^diff --git a\/.*/{
s///
q
}
/^Signed-off-by:/d
/^#/d
' "$MSG" | git stripspace`
if test -z "$clean_message"
then
return
fi

# Does Change-Id: already exist? if so, exit (no change).
if grep -i '^Change-Id:' "$MSG" >/dev/null
then
return
fi

id=`_gen_ChangeId`
T="$MSG.tmp.$$"
AWK=awk
if [ -x /usr/xpg4/bin/awk ]; then
# Solaris AWK is just too broken
AWK=/usr/xpg4/bin/awk
fi

# How this works:
# - parse the commit message as (textLine+ blankLine*)*
# - assume textLine+ to be a footer until proven otherwise
# - exception: the first block is not footer (as it is the title)
# - read textLine+ into a variable
# - then count blankLines
# - once the next textLine appears, print textLine+ blankLine* as these
# aren't footer
# - in END, the last textLine+ block is available for footer parsing
$AWK '
BEGIN {
# while we start with the assumption that textLine+
# is a footer, the first block is not.
isFooter = 0
footerComment = 0
blankLines = 0
}

# Skip lines starting with "#" without any spaces before it.
/^#/ { next }

# Skip the line starting with the diff command and everything after it,
# up to the end of the file, assuming it is only patch data.
# If more than one line before the diff was empty, strip all but one.
/^diff --git a/ {
blankLines = 0
while (getline) { }
next
}

# Count blank lines outside footer comments
/^$/ && (footerComment == 0) {
blankLines++
next
}

# Catch footer comment
/^\[[a-zA-Z0-9-]+:/ && (isFooter == 1) {
footerComment = 1
}

/]$/ && (footerComment == 1) {
footerComment = 2
}

# We have a non-blank line after blank lines. Handle this.
(blankLines > 0) {
print lines
for (i = 0; i < blankLines; i++) {
print ""
}

lines = ""
blankLines = 0
isFooter = 1
footerComment = 0
}

# Detect that the current block is not the footer
(footerComment == 0) && (!/^\[?[a-zA-Z0-9-]+:/ || /^[a-zA-Z0-9-]+:\/\//) {
isFooter = 0
}

{
# We need this information about the current last comment line
if (footerComment == 2) {
footerComment = 0
}
if (lines != "") {
lines = lines "\n";
}
lines = lines $0
}

# Footer handling:
# If the last block is considered a footer, splice in the Change-Id at the
# right place.
# Look for the right place to inject Change-Id by considering
# CHANGE_ID_AFTER. Keys listed in it (case insensitive) come first,
# then Change-Id, then everything else (eg. Signed-off-by:).
#
# Otherwise just print the last block, a new line and the Change-Id as a
# block of its own.
END {
unprinted = 1
if (isFooter == 0) {
print lines "\n"
lines = ""
}
changeIdAfter = "^(" tolower("'"$CHANGE_ID_AFTER"'") "):"
numlines = split(lines, footer, "\n")
for (line = 1; line <= numlines; line++) {
if (unprinted && match(tolower(footer[line]), changeIdAfter) != 1) {
unprinted = 0
print "Change-Id: I'"$id"'"
}
print footer[line]
}
if (unprinted) {
print "Change-Id: I'"$id"'"
}
}' "$MSG" > "$T" && mv "$T" "$MSG" || rm -f "$T"
}
_gen_ChangeIdInput() {
echo "tree `git write-tree`"
if parent=`git rev-parse "HEAD^0" 2>/dev/null`
then
echo "parent $parent"
fi
echo "author `git var GIT_AUTHOR_IDENT`"
echo "committer `git var GIT_COMMITTER_IDENT`"
echo
printf '%s' "$clean_message"
}
_gen_ChangeId() {
_gen_ChangeIdInput |
git hash-object -t commit --stdin
}

add_ChangeId