从叉子到父存储库的堆叠拉动请求
堆叠拉动请求当从叉子到父仓库的PR时，使用模式？
我经常使用此模式，当我们需要将大型功能分为小型可审查的PRS时。在一个回购中，它只需要在子分支之间进行PRS，这很简单。但是，当我想在存储库之间创建PR时，我找不到如何实现此目标。
例如，我想创建一个修复时间轴组件的表顺序的PR，但是此更改取决于修复侧向箭头和菜单按钮的关注性的PR。在单个存储库中，我将具有这样的PRS依赖关系结构：
master＆lt; - sideways-fixsideways-fix＆lt; - 菜单框架菜单>菜单fix＆lt; - tabbing-fix
，但是交叉repo pr允许我仅选择master分支作为目标。这导致拥有3个PR：
Master＆lt; - 侧向框架master＆lt; - 菜单框架master＆lt; - tabbing-fix
最后一个PR重叠了上面的2个PR的变化，使其更大，更难审查。
How can the Stacked Pull Requests pattern be used when doing PRs from the fork to the parent repo?
I use this pattern often when we need to split a large feature into small reviewable PRs. In one repo it just requires doing PRs between sub-branches which is quite simple. But I wasn't able to find how to implement this when I want to create a PR between repos.
For example, I want to create a PR that fixes the tabbing order of the Timeline component, but this change depends on the PRs that fix the sideways arrows and menu buttons' focusability. In single repo I would have such PRs dependencies structure:
master <- sideways-fixsideways-fix <- menu-fixmenu-fix <- tabbing-fix
But the cross-repo PR allows me to only select the master branch as the target. This leads to having the 3 PRs:
master <- sideways-fixmaster <- menu-fixmaster <- tabbing-fix
And the last PR overlaps the changes from the 2 PRs above which makes it larger and harder to review.
如果你对这篇内容有疑问，欢迎到本站社区发帖提问 参与讨论，获取更多帮助，或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱，如果其他用户或者作者回复了您的评论，将不能在第一时间通知您！
发布评论
评论（1）
虽然不直接支持这一点，但自2021年10月以来，https://github.blog/changelog/2021-10-27-pull-request-request-merge-queue-limited-beta/” “ nofollow noreferrer”>拉动请求合并队列有限β正在进行中（现在，2023年2月，GA：一般可用性）。
更多的是验证与“准备合并”并行确定为“准备合并”的不同组合，以便拉动请求可以有效合并，而没有当今合并之间存在的典型延迟。
因此，不完全是您的用例，而是关闭的，因为您不必在合并PR之前重新重新使用PR。
在客户端，您有
cape> cape lylyze/pr-train可以提供帮助：While this is not directly supported, there is since Oct. 2021 a Pull Request Merge Queue Limited Beta in progress (now, Feb. 2023, in GA: General Availability).
It is more to validating different combinations of pull requests identified as “ready to merge” in parallel so that pull requests can merge efficiently and without the typical delays that exist between merges today.
So, not exactly your use case, but close, since you don't have to rebase a PR before it can be merged.
On the client side, you have
realyze/pr-trainwhich can help: