Adds CRAN compliance helpers and vignette

[JosiahParry]

This PR superceds #313. The purpose of this PR is create a pathforward for developers who wish to publish extendr packages on CRAN. It is based on prqlr and my experience publishing rsgeo to CRAN.

The PR adds:

• use_cran_defaults() which adds configure and configure.win files, modifies Makevars and Makevars.win to use CRAN requirements
• vendor_pkgs() which vendors crates appropriately
  • uses cargo generate-lockfile for brand new packages without a Cargo.lock file
  • automatically updates vendor-config.toml
• vignettes/articles/cran-compliance.Rmd which describes the process for making R packages CRAN compliant.

It does not add functionality to document crate authors. rsgeo does not do this and has not been problematic in getting the package published—though it may in the future, I am unsure.

Nonetheless, I used these functions to publish rsgeo and it has been successful. This PR provides minimal funcitonality to provide a path forward towards CRAN compliant packages in an opt-in manner.

Follow up tasks

• [ ] merge SystemRequirements change #310
  • To be CRAN compliant the change proposed in #310 is required. Initially, i added it as part of this PR but I believe it should be for all new extendr packages not just CRAN ones
• [ ] Identify appropriate way for handling crate author attribution. I have not encountered this issue.

[JosiahParry]

@Ilia-Kosenkov I cannot figure out what the error is with these snapshot tests for testing vendoring output. @CGMossa was able to pull the branch, run the tests, and they all passed. Do you have any idea what might be causing the failure? I'm tempted to remove to test all together to appease the CI gods.

[Ilia-Kosenkov]

Ok let me check

[Ilia-Kosenkov]

I got this snapshot failure -- you incorrectly parsed stdout

[JosiahParry]

Which is yet different than the CI errors 🙈 ! Thank you for testing

[Ilia-Kosenkov]

Maybe you can extract vendored packages (if I udnerstand your logic) using this regex? Vendoring\s([a-z0-9_][a-z0-9_-]*?)(?=\s)\s(.+?)(?=\s)

[Ilia-Kosenkov]

You changed snapshot to snapshot x, but the actual snapshot still has print(x), hence it complains :)

[JosiahParry]

@Ilia-Kosenkov true! But i don't think that's the core of the problem. If you look at the CI output nothing is showing up :O

I'll save it for another time.

[Ilia-Kosenkov]

Yep which means some sort of parsing had failed. I will check this out in an hour or so

[Ilia-Kosenkov]

The reason I insist on this is that I believe this flaky test shows instability in parsing, which may negatively affect users. So we better figure it out

[Ilia-Kosenkov]

@JosiahParry , can I push to your branch here?

[JosiahParry]

@Ilia-Kosenkov confirmed

[JosiahParry]

@Ilia-Kosenkov your changes have fixed CI! I've added a note to NEWS.md as well.

[JosiahParry]

I'd like to revisit this PR. I used this fork to publish {arcpbf} to CRAN. Using these functions has made it very easy to publish extendr packages to CRAN. There was no back and forth about the use of Rust in this package, no back and forth about the build or compilation process etc.

I think making it easy to publish extendr packages to CRAN is the best way to make this project widely useful and interesting.

Please let me know what is needed to merge this PR.

[Ilia-Kosenkov]

Hey @JosiahParry , what' the status of this PR?

[JosiahParry]

Please let me know what is needed to merge this PR

What do you need from me to merge this PR? I've been using it myself for cran publications.

[Ilia-Kosenkov]

Tests are failing due to version mismatch, can you pull it locally and run tests - it should tell you which snapshot is wrong.

[Ilia-Kosenkov]

And add a reference to your Pr in News.md, I think it is missing

[JosiahParry]

Cool sounds good! I'll appease PR gods tomorrow-ish and update news to get this through!

[Ilia-Kosenkov]

Looks solid! If you have no more changes to make @JosiahParry , should we merge it?

[JosiahParry]

No more changes from me at this time :)