improve docs of generated library

[robamu]

This adds feature gate displays, for example like the ones in the syn docs.
The generated library only has a few of those, but I find them to be very useful. Fixes #855

[burrbull]

It is nightly feature, is not it?

[robamu]

The additional feature gate displays require building docs with nightly, but it is still possible to generate docs using stable without issues. docs.rs always builds the documentation with nightly.

[burrbull]

We possibly need a "doc workflow" (on stable & nightly). It can be independent job or added to one of existent. Could you work on this?

[robamu]

You mean building the docs for the PACs generated as part of the CI? I could look into this.. I would probably have to adapt the ci/script.sh for this? Maybe it would be a good idea to move the CI from the shell script to the regress tool as a pre-requisite?

[burrbull]

Maybe it would be a good idea to move the CI from the shell script to the regress tool as a pre-requisite?

Sure. #689
But better in separate PR.

[robamu]

#905 is WIP and I already was thinking of how to best add this workflow. I saw that the bash script is used for the clippy check as well. If this is done by the regress tool, I am wondering what the best way would be to do the CI. The regress tool already takes care of creating new projects for each chip. One way would be to also run clippy and doc for each job directly after the check. The alternative would be to run clippy and docs only for a small selection of chips in s separate job.. what would be the better/preferred way here? @burrbull @Emilgardis

[Emilgardis]

For checking if generated doc comments is correct it'd be enough to just run it with the regular per-target tests imo.

[burrbull]

run clippy and docs only for a small selection of chips in s separate job

This is enough