Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 98.54%. Comparing base (3d19a8b) to head (ef5d948).
Report is 250 commits behind head on master.

@@           Coverage Diff           @@
##           master   #17301   +/-   ##
=======================================
  Coverage   98.54%   98.54%           
=======================================
  Files         169      169           
  Lines       21898    21898           
=======================================
  Hits        21579    21579           
  Misses        319      319           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

Not sure if it's important enough to be featured in that document, but maybe it'd be useful to document the special test file name prefixes/suffixes recognised by tests/run-tests.py that act as rudimentary feature checks to skip certain tests?

For example, test files whose name starts with native_ or viper_ will be automatically skipped if @micropython.native or @micropython.viper support was not detected in the target interpreter, and so on. There's a dozen or so of those tags so maybe it's not worth documenting them all, but again, the LLM may sort all that stuff out by itself :)

I think that is a good addition

special test file name prefixes/suffixes

Yeah I do think it'd be helpful to have these itemised somewhere as it's certainly not abvious for new developers!

I'm not too sure about how best to divide what does in the rst docs and what goes in the readme here... I kinda think the readme is perhaps better for "internals" things like that which are maybe of more interest to low level developers, with the rst docs more focussing on "end users" interested in running tests?

I personally use the README files as a quickstart guide myself, always favouring explicit documentation if it's available. It used to be like that back then, when dinosaurs still roamed freely and whatnot, so maybe somebody from a younger demographics should weigh in on this.

Considering the users/developers split, I wonder if this possible TOC for three (or two) documents - names are generic on purpose - can be a starting point for a discussion on this:

• "Quickstart"/"End users"
  • Standard tests
    • How to run tests (simplified version, no need to bring emitters and mpy in, f.ex.)
    • How a successful run should look like
    • Where to look for failure info to send back to a developer
  • Performance benchmarks
    • How to run benchmarks (simplified version)
    • How to read output and compare with previous runs
• "Developers"
  • How to run tests (what's left to document)
  • Standard tests (almost everything that's in writingtests.rst for standard tests, maybe skip more niche sub-topics)
  • Multi-instance tests
  • SSL tests
  • Performance benchmarks (what's left to document)
• "Internals" (fold this into "Developers" if needed)
  • How run-tests.py enumerates tests and how it decides which ones to skip
  • Natmod tests
  • Internal benchmarks
  • Everything else that's not in "Developers" ?

I can help out on the natmod tests bit if needed, as long as somebody else fixes the grammar :)

for context : I just got an inflatable Dino for my birthday.
There is no need to settle on just a single topilogy
With Sphinx & MyST it is simple to pull in the Readme.MD and other markdown docs into the online documentation.
That allows the information to be found and accessed va different ways and personal preferences , without increasing the effort to maintain it.
( possibly that is also a way to pull in some of the information in the Wiki , which is rather hard to search )

Thanks for the suggestions, I've restructured the content in the latest push. I haven't got the rst docs importing the readme, not sure how much adjusting would be needed in the sphinx stuff to support that?
The readme includes links to the rst docs which would need to be stripped out or handled differently if it was included directly?
I don't think there's too much duplication there really...

The content looks pretty streamlined now, that's great. I'm not much of a sphinx user so I can't comment on how rst inclusion/handling can be improved - but content-wise it's really good!