git使用中的change_id

git提交时的Change-Id问题 为什么要使用change-id Change_id的作用是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。 1.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 2.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chmodu+x .git/hooks/commit-msg保证hook可执行。 commit-msg hook脚本 #!/bin/sh # FromGerrit Code Review 2.6 # # Part of Gerrit Code Review () # # 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 # # # # 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" | gitstripspace` if test -z "$clean_message" then return fi # Does Change-Id: already exist? if so, exit (no change). ifgrep -i '^Change-Id:' "$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 } /]$/ && (footerComment == 1) { footerComment = 2 } # We have a non-blank line after blank lines. Handle this. (blankLines> 0) { print lines for (i = 0; i "$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 $parent" fi echo "author `gitvar GIT_AUTHOR_IDENT`" echo "committer `gitvar GIT_COMMITTER_IDENT`" echo printf '%s' "$clean_message" } _gen_ChangeId() { _gen_ChangeIdInput | git hash-object -t commit --stdin }