[GSoC] pack-objects: plural-aware thread message
- From
- Manoj <manojkumar148700@gmail.com>
- Date
- Nov 9, 2026, 15:07 UTC
- Message-ID
- <CAOVdN544jyuLG7XNE8jTBeA90iSBS_NanW-Cfn3XgOqTFgQJEg@mail.gmail.com>
Hi,
I'm Manoj, a third-year CSE student interested in contributing to Git for GSoC 2027. I've had a few open-source contributions merged, including uutils/coreutils#15558, kaelzhang/node-ignore#175 and bitranox/igittigitt#48. This is my first mail to the Git mailing list.
While looking for a possible microproject, I noticed that "Delta compression using up to %d threads" in builtin/pack-objects.c uses _(), whereas the path-based delta compression message in the same file uses Q_().
I tested this with --progress --threads=2 and --progress --threads=1. The message appears with two threads but not with one. I also found that ll_find_deltas() returns early when delta_search_threads <= 1, before reaching the message.
The path-based message has the same early return, so its singular form cannot be reached either, but it still uses Q_(). This made me wonder whether the first message should follow the same approach.
I understand that Q_() provides plural-aware translation based on the count. I'm not entirely sure how much practical difference this makes for existing translations, though.
This wouldn't change English output or runtime behaviour. The original message was made translatable using _() in 2018, while the path-based message was introduced with Q_() in 2025. I also understand that changing the translation string would cause existing translations to need review.
I checked current master at fb255012ac. On 9 November, I fetched origin/master, origin/seen and origin/next and grepped builtin/pack-objects.c for "Delta compression". All three still contain the original message, and I found no other in-flight change touching it. I also searched Lore for existing discussions and patches concerning it.
Would this be a reasonable microproject to work on, or is it already being considered? I'd appreciate any suggestions before I start working on the patch.
Sorry if this has already been discussed or isn't worth the maintainers' time.
Thanks, Manoj