shym

I also think “ignored” is confusing. I suppose you meant something like “will not be used to check consistency between linearized and non-linearized runs”, didn’t you? Part of the confusion...

Regarding `Bigarray` failures on macos, I opened #160 and #161, the combination of which I’m testing on my integration branch. They give reproducibility in my first rounds of testing. #161...

As `main` and `multiple-ts` contain naturally conflicting changes, I find it makes more sense (at least to me) to merge instead of rebasing. That’s why I created my integration branch,...

That’s probably a matter of preference, but by merging instead of rebasing, _while the branch is still WIP_, you record a trace of how conflicts were resolved. When the branch...

I do agree it’d be nicer to go without `bash`. I had tried first to see if it would be possible to do in dune, but I couldn’t figure it...

Now, to plead the case for that PR: - we are the end users for the test suite and can make choices that would otherwise be clearly dangerously fragile, so...

For the record, I thought using the `echo` dune stanza could at least provide the feature of having the path of the test being run before it starts but my...

The documentation explains [here](https://docs.github.com/en/actions/using-workflows/workflow-commands-for-github-actions#setting-an-error-message) the syntax to report an error and create the anchor; tldr: `echo "::error title=Failure in STM Bytes test parallel::STM Bytes test parallel failed on its 11th...

In fact `bash` should create the temporary file in `/tmp` or some other system temporary directory. I wonder whether the sandboxing allows that. Another workaround could be to use a...

> If you replace `-cpu cortex-a53` with `-cpu max` it should work fine. I'll do some more digging to find out why it fails for `cortex-a53` in the first place,...