"brian m. carlson" <sandals@xxxxxxxxxxxxxxxxxxxx> writes:
> On 2025-07-18 at 22:26:31, Eric Frederickson wrote:
>> Hello everyone,
>
> Hi,
>
>> I hope that this message finds you well! I'm a software developer and passionate
>> git user, and I'd like to try my hand at contributing to the project. I'm
>> sending this message in order to ask Junio and the team if there's anything that
>> would be particularly useful / appropriate for me to start looking into on that
>> front. (Reading through the last few "What's cooking" messages, I didn't see
>> anything that jumped out at me as needing a new contributor, which is why I'm
>> asking in a separate message instead of replying to one of those.)
>
> It's great to hear that you want to contribute.
Thank you for your welcoming response :). Excited for the journey!
>> (Some notes on my skills: working on docs or tests is always a favorite for me,
>> so things in those areas would be a great time. Also comfortable with low-level
>> code, and any kind of scripting. Note too that academic background centers
>> around programming language design and parsing related stuff, so I've got some
>> fluency in those areas that I could hopefully apply well to the project if ever
>> needed.)
>
> We have a Git FAQ because I answered questions on Stack Overflow and got
> tired of answering the same question again and again, so if you want to
> add any common problems there, that's of course welcome.
Noted.
> (Now, to be clear, people did not stop asking those questions, but I had
> prepared text and a link to provide to them, and then people _knew_ that we
> had a FAQ and could refer to it.) If there's anything that you think is poorly
> documented and colleagues or people online have trouble with, we would
> absolutely love for that documentation to be improved.
Sounds good. Would love to submit some patches to this effect at some point.
> One thing I did to get started some years ago is to grep through the
> codebase for TODO statements and to pick something that seemed simple
> enough to do, and then send a patch.
>
> If there are small functions that you think would benefit from unit
> tests, we now have the framework for that that we lacked until recently.
>
> If you see things in the tests marked test_expect_failure, then that
> means that's a TODO test: it should pass, but it presently does not, and
> that can sometimes be a source of good things to work on. Some of those
> things are kind of big (especially a lot of the submodule TODOs), but
> sometimes they're small and approachable.
This was very helpful to learn!
I've been looking through them, and have found a pair that's drawn my eye: the
pair of known breakages in:
t3903-stash, tests 52 and 53
("stash directory to file" and "stash file to directory").
