tests: rewrite rebase-submodule .gitmodule file#4275
Conversation
|
Crossing fingers (a.k.a thanks Apple, for .DS_Store files) 🔥. |
383d593 to
1fc265e
Compare
|
The change looks good to me, thanks. There are more tests with a ".gitmodules" file, though, any particular reason why you didn't adjust them, as well? |
|
The rationale is that I'm still in a bit of a conundrum here, as even with that change, on a freshly cloned repository, I get Note, this is not using |
1fc265e to
7e71668
Compare
|
Unfortunately, this breaks |
7e71668 to
5cd7d42
Compare
|
Gah, that gave me a hard time. Tests should be A-OK, but I'm unsure of the cleanliness of my fix. I had to reset after doing the commit, least the index would be weird. |
pks-t
left a comment
There was a problem hiding this comment.
You forgot to free some variables in the initialization. You could also merge your two commits together and add some reasoning why this commit is required inside the commit message. Looks fine otherwise and passes ony my local machine, will merge when my comments are addressed. Thanks!
|
|
||
| /* And a final reset, for good measure */ | ||
| git_object_lookup(&obj, repo, &commit_id, GIT_OBJ_COMMIT); | ||
| cl_git_pass(git_reset(repo, obj, GIT_RESET_HARD, &opts)); |
There was a problem hiding this comment.
This misses cleanup of some variables, namely parent, obj and master_ref.
5cd7d42 to
35cb7b8
Compare
|
@pks-t Fixed the missing frees, and the space-indent. Would it be possible to have that back-ported to 0.26 as well (this is against master), as it's intended to help downstream users/wrappers ? |
|
Good to go, Travis choked on proxies again. |
Fixes #4274