[英]What should my automated git update process look like?
I come from a subversion background and recently my company made the switch to git. I used to have a cron entry, on my laptop, to update the multitude of checkouts, on a nightly basis.我来自颠覆背景,最近我的公司切换到 git。我曾经在我的笔记本电脑上有一个 cron 条目,每晚更新大量结账。 This way, I would be running against current versions of the different components of our system;这样,我将针对我们系统不同组件的当前版本运行; especially the parts that I wasn't actively developing, but had dependencies on.特别是那些我没有积极开发但有依赖性的部分。 I would like to achieve the same thing with git.我想用 git 达到同样的目的。
Here's my old update process with svn:这是我使用 svn 的旧更新过程:
#!/bin/bash -e
checkout="$1"
svn update --accept postpone ${checkout}
# Run a script to report conflicts that I would resolve in the morning.
I've read a lot of blogs posts on the topic and asked around, and I haven't found many consistent answers.我已经阅读了很多关于该主题的博客文章并四处询问,但没有找到很多一致的答案。 Also, none of the solutions I've seen so far are complete to the extent that I am looking for.此外,到目前为止,我所看到的解决方案都没有达到我正在寻找的程度。 I've taken all those opinions and created the script below.我采纳了所有这些意见并创建了下面的脚本。
How should I deal with submodules?我应该如何处理子模块?
Are there situations, or gotchas, I have not accounted for?是否存在我没有考虑到的情况或陷阱?
#!/bin/bash -e
checkout="$1"
now=$(date +%Y%m%dT%H%M%S)
cd ${checkout}
# If you are in the middle of a rebase, merge, bisect, or cherry pick, then don't update.
if [ -e .git/rebase-merge ]; then continue; fi
if [ -e .git/MERGE_HEAD ]; then continue; fi
if [ -e .git/BISECT_LOG ]; then continue; fi
if [ -e .git/CHERRY_PICK_HEAD ]; then continue; fi
# Determine what branch the project is on, if any.
ref=$(git branch | grep '^*' | sed 's/^* //')
if [[ $ref = "(no branch)" ]]; then
# The directory is in a headless state.
ref=$(git rev-parse HEAD)
fi
# If there are any uncommitted changes, stash them.
stashed=false
if [[ $(git status --ignore-submodules --porcelain | grep -v '^??') != "" ]]; then
stashed=true
git stash save "auto-${now}"
fi
# If there are any untracked files, add and stash them.
untracked=false
if [[ $(git status --ignore-submodules --porcelain) != "" ]]; then
untracked=true
git add .
git stash save "auto-untracked-${now}"
fi
# If status is non-empty, at this point, something is very wrong, fail.
if [[ $(git status --ignore-submodules --porcelain) != "" ]]; then continue; fi
# If not on master, checkout master.
if [[ $ref != "master" ]]; then
git checkout master
fi
# Rebase upstream changes.
git pull --rebase
# Restore branch, if necessary.
if [[ $ref != "master" ]]; then
git checkout ${ref}
fi
# Restore untracked files, unless there is a conflict.
if $untracked; then
stash_name=$(git stash list | grep ": auto-untracked-${now}\$" | sed "s/^\([^:]*\):.*$/\\1/")
git stash pop ${stash_name}
git reset HEAD .
fi
# Restore uncommitted changes, unless there is a conflict.
if $stashed; then
stash_name=$(git stash list | grep ": auto-${now}\$" | sed "s/^\([^:]*\):.*$/\\1/")
git stash pop ${stash_name}
fi
# Update submodules.
git submodule init
git submodule update --recursive
Thank you.谢谢你。
I got this error when I had duplicate submodule request in my.gitmodules file.当我在 my.gitmodules 文件中有重复的子模块请求时,我得到了这个错误。
You need to check if there are changes in the submodules as well. 您还需要检查子模块中是否也有更改。 See git submodule foreach
. 参见git submodule foreach
。
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.