Use the semver trick? about cortex-m-rt HOT 3 CLOSED

Are there many crates that depend on cortex-m-rt besides PACs? I had hoped it wouldn't be too bad an upgrade (compared to cortex-m for example). We've had all sorts of annoying problems coordinating the semver-trick releases (and it's a ton of work to go through every public item and work out if it can be safely re-exported from the new version or you have to maintain the old version), so I'd rather avoid it if we can...

I haven't thought about whether it would be viable for c-m-rt. I think basically the only thing other crates rely on is the interrupt macro for PACs, which hasn't changed and so should be easy to re-export. So it's probably possible...

from cortex-m-rt.

Yeah, maybe this is a non-issue after all. I'll try around a little more ( I mean I haven't really mentioned an example in the OP) and see if that's really a problem and if not I will close this issue.

Thanks for the insight @adamgreig

from cortex-m-rt.

Are there many crates that depend on cortex-m-rt besides PACs? I had hoped it wouldn't be too bad an upgrade (compared to cortex-m for example).

So I checked the dependency tree of stm32f3xx-hal again and the dependencies which use are panic-probe and defmt-test, firstly which is a special usecase and secondly are optional dependencies. So you are right, that almost nobody besides PACs are using it.

Also solving the possible conflict via PACs and HALs can be done with version = ">=0.6.13, <0.8". This was just not working locally for me, because I forgot to call cargo update before running build again.

So I guess, even though the semver-trick might be possible it is not worth doing it right now, because there is no immediate need for it.

from cortex-m-rt.