Patrick Steinhardt [off-list ref] writes:
Changes compared to v2:
- `append_fetch_head` now takes a `struct object_id *old_oid`
instead of a `char *old_oid`.
- Handling of the merge status marker was moved into
`append_fetch_head`.
- I've unified code paths to format the line we're about to write to
FETCH_HEAD in atomic and non-atomic cases, which is the new 2/5
patch.
- We now always initialize `fetch_head->fp` and use it to derive
whether anything should be written at all instead of using the
global `write_fetch_head` variable.
I think this should address all of Junio's feedback. Thanks for your
input!
Looking good. I left a few comments on remaining or newly
introduced issues, but IIRC all of them were rather minor.
Thanks, will replace.