-1

my company is currently working on Azure Cloud, managing our code repositories using Azure Devops and Git.

For some months we decided to use the Squash-commit when we wanted to merge our feature branches into our develop branch through the means of Pull Requests.

Since then, we are experiencing big issues in merging files as in the File-Updates sections we can see thousands files updates listed. Most of them have already been commited to our Development branch, so they are fake and they only raise many concerns about the possible implications of committing them again.

Why does git recognize these files as changed (newly)? Is there anything we re missing? or doing wrong?

Regards

Alessio Orlando
  • 35
  • 1
  • 5
  • Does this answer your question? [Git merge conflicts during PRs](https://stackoverflow.com/questions/65041738/git-merge-conflicts-during-prs) – matt May 25 '21 at 03:47
  • "Is there anything we re missing? or doing wrong?" Evidently yes. A squash merge is not a merge! So the merge base never moves, the PRs get longer and longer, and conflicts can easily arise and they get worse and worse. You picked a wrong strategy. – matt May 25 '21 at 03:49

1 Answers1

1

On the files comparison of a PR, it does not only compare the file changes between the source branch and target branch. It will compare the commit history between them.

It is a correct behavior if you use the squash merge multiple times to merge changes from a branch to another one. Because instead of each commit on the topic branch being added to the history of the default branch, a squash merge takes all the file changes and adds them to a single new commit on the default branch.

This means that when you use squash merge to merge a PR, only the file changes on the source branch (feature in your case) will be merged and generate a new commit on the target branch (develop in your case). The commit history for these file changes on the source branch will not be merged to the target branch.

Generally, when completing the squash merge, you should delete the source branch. And in the next time if you still want to use the squash merge, you can create a new source branch (you can still use the same branch name feature) based on the target branch. Then make changes on the new source branch, merge changes to the target branch using squash merge, delete the source branch.

If you do not delete the source branch when using the squash merge, and in the next time you continue making and merging changes from this source branch to the target branch, you may get some conflicts and extra changes just like the problem you are facing.

At this time, if there are conflicts, resolve the conflicts then you can complete the merge as usual. If no conflicts, you can directly complete the merge as usual. When completing the merge, do not forget to delete the source branch.

To view more details, you can see "Squash merge pull requests".

enter image description here

Bright Ran-MSFT
  • 5,190
  • 1
  • 5
  • 12