Shane F. Carr

Also, a possible advantage of this naming scheme is that the data struct name is fully a substring of the data marker name. If the data struct is `SymbolsV1`, you...

WG discussion: - @Manishearth I like `V1Marker` being a thing, but the point about it becoming `V2Marker` is valid. So splitting it in the middle is fine. We have a...

I talked this over with @echeran today. He pushed back on the position from https://github.com/unicode-org/icu4x/issues/4991#issuecomment-2350760637 that we can't just use the Rust code path as the marker name and as...

@echeran also astutely asked, why are we even discussing this? What problem are we trying to solve? The problem we are trying to solve is making a clear framework for...

> I think we should minimize the need to carry or compare long strings in run-time binaries (i.e. should prefer small strings both due to binary size and for comparison...

@robertbastian Do you agree with https://github.com/unicode-org/icu4x/issues/4991#issuecomment-2504732608? I updated it to include the change I suggested in https://github.com/unicode-org/icu4x/issues/4991#issuecomment-2505073780

To reiterate part of https://github.com/unicode-org/icu4x/issues/4991#issuecomment-2504732608 that we didn't discuss: I proposed that: 1. Data markers do not have the crate name in their name 2. The crate name, without `icu_`,...

> nope 😦 Haven't looked if there's a workaround. Agreed this is unfortunate. Seems like a tooling problem, though, which _could_ be fixed. > Both of this is possible with...

The proposed solution still doesn't taste good to me; there are the following two requirements I have that aren't stated in the previous post, both of which are mentioned earlier...

@zbraniecki Thoughts on this?