It's possible to commit files that contains conflict data. Is there a way to mark these files as conflicted again, so that running git mergetool will generate the necessary files and run the merge tool?
If the index is already in a conflict state, simply check out the file with the --conflict=merge
flag:
git checkout --conflict=merge file
If the index is clean because the unresolved file has been [erroneously] added, just reset it before checking it out:
git reset file
git checkout --conflict=merge file
This will allow you to resume conflict resolution normally (e.g., git mergetool
).
NOTE: Promoting a comment to @jakub-narębski's answer into its own answer by request from @fourpastmidnight. :)
You can get contents of file with conflict markers using git checkout --conflict=merge -- file
, but if you have cleaned up index by using git add file
(or if GUI did that for you) it wouldn't work.
There is git update-index --unresolve
, but it is hacky, and does not work very reliably. I think the state it restores would be not enough for git-mergetool.
You would probably have to redo merge, or use git update-index --cacheinfo
to manually set stages version... git-stash can help you preserve correctly resolved conflicts.
git update-index --unresolve
was created in "ancient" times to allow to restore git diff --ours
etc. after (accidental) git add
confirming conflict resolution. It stuffs HEAD version (not version with automatically resolvable conflicts resolved) in stage #2, stuffs MERGE_HEAD version in stage #3, and doesn't put anything in stage #1, the ancestor version. –
Selfexistent git pull <repo>
you can do git merge <repo>
which is shortcut for git merge <repo>/<branch>
, and <repo>/<branch>
is remote-tracking branch. –
Selfexistent git diff
with the redone merge and this messed up merge commit. That diff hopefully contains roughly the changeset that was originally intended to be committed so you do not need to manually re-create all that work. –
Cream git reset file; git checkout --conflict=merge file
. –
Dint The most elegant solution would be to prevent this problem from the very beginning:
git config --global mergetool.[tool].cmd [command-line call]
git config --global mergetool.[tool].trustExitCode false
As far as I know, you won't be able to commit while a file still contain conflict markers in it.... which is not exactly true:
The OP mentions that you can (I copy here his pastbin), but that won't be enough for the mergetool to be trigger again:
Auto-merged README
CONFLICT (content): Merge conflict in README
Automatic merge failed; fix conflicts and then commit the result.
lynx:~/test_clone$ ls
README
lynx:~/test_clone$ git add README
lynx:~/test_clone$ git commit -a
Created commit 46ee062: It works!
lynx:~/test_clone$ ls
README
lynx:~/test_clone$ cat README
<<<<<<< HEAD:README
testingtesting
=======
hmm
>>>>>>> 881d60f5f738bc5716f5c9a9384e262b535717fd:README
lynx:~/test_clone$
As Charles Bailey comments, and illustrates in this SO answer, the mergetool is queried because there are 3 instances of the same file in the index:
For an unmerged file in a conflict git makes available the common base, local and remote versions of the file in the index. (This is where they are read from for use in a 3-way diff tool by
git mergetool
.) You can use git show to view them:
# common base:
git show :1:afile.txt
# 'ours'
git show :2:afile.txt
# 'theirs'
git show :3:afile.txt
git add
(with whatever content, including conflict markers) will automatically removes 2 of them, ensuring the mergetool
won't be called again.
git mergetool
detect it and re-trigger a merge? –
Interminable Please use git update-index --unresolve
Since git 1.7 it uses resolve-undo info from the index to restore all 3 stages (1:base, 2:ours, 3:theirs): https://github.com/git/git/commit/8aa38563b22c84b06ea1fff9638cc1f44fda726f
© 2022 - 2024 — McMap. All rights reserved.