Date & Time

[zeenix]

This should have been supported from the day 1. I think we need to take the user-facing API from toml_datetime (so we don't need to depend on another external crate) and perhaps also take the parsing code from toml_edit (with some adaptation of course).

I think this will also fix a bunch of failing cases in the from the test harness (#11).

[zeenix]

cc @CosmicHorrorDev. BTW, if you haven't already, I'd suggest subscribing to all activity on this repo since you said you intend to help develop and maintain it. :)

[CosmicHorrorDev]

I would personally just depend on toml_datetime at that point. Virtually the entire crate is it's public API

[zeenix]

I would personally just depend on toml_datetime at that point. Virtually the entire crate is it's public API

Good point.

[zeenix]

I would personally just depend on toml_datetime at that point. Virtually the entire crate is it's public API

Good point.

It's not a no_std lib though so we'll need to make it one, to use it.

[CosmicHorrorDev]

Ah I didn't realize 😅

It looks like it only needs alloc, so I can handle working with upstream on that

[zeenix]

It looks like it only needs alloc, so I can handle working with upstream on that

Yeah, I asked the maintainer in a PM if he'd be interested in a PR for that. Unlikely that he's not but you never know. :)

[zeenix]

It looks like it only needs alloc, so I can handle working with upstream on that

Yeah, I asked the maintainer in a PM if he'd be interested in a PR for that. Unlikely that he's not but you never know. :)

He says, he's good with that. 👍 So please go ahead.

[zeenix]

He says, he's good with that. 👍 So please go ahead.

One issue though: Error trait only became stable in core in Rust 1.81, which is way too recent and I doubt Ed would be happy to bump from 1.65 to 1.81 just for this. In tomling, I added an std feature only for this and Error trait is implemented if std is enabled. However, that would be a breaking change and seems Ed (understandably) isn't too enthusiastic about that. Unless there is another solution to this issue that doesn't involve breaking changes or too big an MSRV bump, I'm afraid we'll be only left with my first suggestion. :(

Btw, Ed said to add #![warn(clippy::std_instead_of_core)] to the crate if any std API will still be used. I already did that for tomling.

[CosmicHorrorDev]

Ah yeah that's a good point 😿

Whelp then I think we carry on with copying out the parts we need

[zeenix]

Whelp then I think we carry on with copying out the parts we need

Yeah. 🤷 Still if we try and keep modifications to minimum, we can ditch our copy once toml_edit can bump its MSRV to 1.81.