support parse_and_remainder

[sundy-li]

Hi, we are migrating chrono into jiff, is there any function that works same as parse_and_remainder ?

https://docs.rs/chrono/latest/chrono/format/fn.parse_and_remainder.html

[BurntSushi]

I don't believe so, but I designed the internal parsers to be able to support this style of API.

However, I wanted a concrete use case for where and how it's used before exposing anything. Can you share an MRE of your Chrono code and the high level use case?

[sundy-li]

1. postgresql/greenplum support parse flexible timestamp as the issue describes

2. we are using a setting named parse_datetime_ignore_remainder defaults to 1 and use chrono to parse it.

https://github.com/databendlabs/databend/blob/7969d365b5003a4cf1224d88050787a179653ac4/src/query/functions/src/scalars/timestamp/src/datetime.rs#L367-L472

[BurntSushi]

Looking at that code, it looks like you aren't actually using the remainder right? You're just looking to parse a prefix?

Also, if I do this, I'll probably return an offset into the input given instead of the &str. That's because Jiff's parsing routines work on both &[u8] and &str, and returning the right type of slice would require more generic machinery that I'm not keen on adding. But with an offset, I think you should be fine.

[sundy-li]

Looking at that code, it looks like you aren't actually using the remainder right? You're just looking to parse a prefix?

Yes, parse the prefix and ignore the remainder that unable to be parsed.

Return an offset would be ok.

[BurntSushi]

OK, this should now be available via jiff::fmt::strtime::BrokenDownTime::parse_prefix in jiff 0.1.26. Please let me know if it works for you. I'm hoping to do a jiff 0.2 release soon, so if I've flubbed the API, I can do a quick breaking change if I know about it. Thanks!

[BurntSushi]

https://docs.rs/jiff/latest/jiff/fmt/strtime/struct.BrokenDownTime.html#method.parse_prefix