wasm-pkg-client: validate semver compliance on publish

[ryan-surname-p]

closes #128

Summary

By default publish:

• rejects duplicate registrations
• ensures component is compatible with it's nearest server series neighbors using wit_component::check_semver

The changeset is fairly large I know though a material amount of the new code are tests and moving functions around to keep things readable.

[vados-cosmonic]

Hey @ryan-surname-p it looks like some conflicts creeped in -- apologies for the inconvenience but would you mind resolving those? Then we can get this merged.

[ryan-surname-p]

No problem at all.

@mkatychev actually reached out to me about this and some work they have in flight as well.

#215

If I understood their request correctly we'd like for that to land first and then work some version of this onto that.

This request seemed reasonable enough to me so the current plan is to monitor the progress of 215 and then adjust accordingly.

[mkatychev]

I'd like to hear @vados-cosmonic's thoughts on which should land first.

I suppose one of us will be handling gnarly merge conflicts regardless so it may be fairer for @ryan-surname-p's changes to land first.

[ryan-surname-p]

FWIW I'm happy enough to handle either approach I just wanted to make sure everyone was working with the same set of assumptions.

[mkatychev]

@ryan-surname-p the multi publish package PR has landed, could you address the conflicts and #213 (comment)?

[ryan-surname-p]

Thanks for the heads up, I'll get to this first thing tomorrow.

I followed up on #213 (comment) but I realize now you weren't actually able to see it because I never submitted my review 🤦

[mkatychev]

wit_component::check_semver is really cool, like cargo-semver-checks for wit!

If we can address||change the nits about Shape, skip_semver_check, impl PartialOrd for DecodedComponent, and assert_matches! then this LGTM!

[mkatychev]

LGTM! Excited to use this.