Git merge subtree problems

[Annirak]

I have two branches, one was used for development and one is used for integration. Extensive development has been done in the development branch, and I want to merge a portion of it back in. The specific portion that I want to merge is all contained in one subtree of the development branch.

My directory structure is like this:

Code: Select all

[Branch A]              [Branch B]
    |                       |
    +--Dir1                 +--Dir1
    +--Dir2                 +--Dir2
        |                   |   |
        +--DirA             |   +--DirA
            |               |       |
            +--File1        |       +--File1
            +--File2        |       +--File2
                            |       +--File3
                            |       +--File4
                            +--Dir3


I want to merge Branch B/Dir2/DirA into Branch A/Dir2/DirA. I want File1 and File2 to be merged and File3 and File4 should be created in Branch A. I don't want to pick up Dir3, or any changes in Dir1.

I've tried the steps outlined by kernel.org for merging subtrees, but they fail when I do the git read-tree with:

Code: Select all

error: Entry 'Dir1/DirA/File1' overlaps with 'Dir1/DirA/File1'.  Cannot bind.

I could cherrypick the files to merge, but that seems unnecessarily convoluted for what should be a common and straight-forward problem.

Thanks,
Annirak

[Aaeriele]

unnecessarily convoluted for what should be a common and straight-forward problem.

It's not actually as straightforward as you think, due to how Git works. Git doesn't track per-file history - it tracks per-repository history, which happens to be mappable onto files. Thus, Git's merging functionality is not designed to merge only "some" files from a given commit, because doing so would fundamentally alter the commit (and thus the cryptographic hashes used by Git as keys to link commits together).

The most straightforward way to do this if you're only merging in one direction (and that would allow you to merge the same set of directories again in the future without undue conflicts) would be to actually merge everything, and then modify the merge commit on your integration branch to not include the parts you didn't want to merge. (The reason for this is that it's easier to *add* changes than to ignore them, so it's easier to effectively ignore-via-negation.)

Something like this:

Code: Select all

git checkout BranchA
git merge --no-commit BranchB
git rm Dir3
git commit -m "Merge everything except Dir3"

[Annirak]

I came up with another solution:

Code: Select all

git log Branch_B -- dir2/dirA | grep ^commit | cut -d ' ' -f 2 | xargs git merge --no-commit

But, then it turned out that the commits I was trying to replicate had been made in enormous blobs. I have to do the merge manually because the branch I wanted to merge from is in bad shape.