Allow asv.conf.jsonc file by default

[lucascolley]

In SciPy, we have lots of comments in our asv.conf.json file. But comments are not allowed in JSON files (hence developers on GitHub or VSCode see lots of red highlighting).

This PR allows the config file to be named asv.conf.jsonc(a JSON with comments file), without having to pass the name to --config.

[lucascolley]

@HaoZeke @mattip does this sounds like something you'd be happy to support by default? Or do you think comments in the config file is a niche use case?

[HaoZeke]

Is the .jsonc file extension widely known / used? If not I don't know that this would make sense to explicitly search for from an ASV perspective, given that this is mostly run in CI anyway and since passing the file works via --config

[lucascolley]

500ish stars on Microsoft's parser (https://github.com/microsoft/node-jsonc-parser), so not widely known, but also not niche IMO.

My reason for opening this PR is that appending c to the file extension appears to be the idiomatic solution to all the red highlighting that appears on the GitHub interface. Unless I am missing a setting.

But no worries if this seems like more hassle than it is worth - I was just hoping that I could change to jsonc downstream without breaking peoples' local asv invocations.

---

The same applies to code blocks in conversation messages:

json:

// a comment

jsonc:

// a comment

[HaoZeke]

500ish stars on Microsoft's parser (microsoft/node-jsonc-parser), so not widely known, but also not niche IMO.

My reason for opening this PR is that appending c to the file extension appears to be the idiomatic solution to all the red highlighting that appears on the GitHub interface. Unless I am missing a setting.

But no worries if this seems like more hassle than it is worth - I was just hoping that I could change to jsonc downstream without breaking peoples' local asv invocations.

The same applies to code blocks in conversation messages:

json:

// a comment

jsonc:

// a comment

OK then I'd say it'd be in scope, since anything to make using ASV more ergonomic is a win, left some suggestions for a refactor though.

[lucascolley]

let me know if this is what you had in mind with Path!

[lucascolley]

thanks @HaoZeke !

Let me know if I can help make an asv release happen, btw - I would like to patch SciPy once this is released.

[HaoZeke]

thanks @HaoZeke !

Let me know if I can help make an asv release happen, btw - I would like to patch SciPy once this is released.

Sure, I'm planning to cut a new release today, and add some documentation on how we manage them and what the tasks generally are (fairly standard but good to have written down). Maybe a REL issue or something.