shym
shym
For some reason, `enabled_if` seemed to work locally but is not enough to tell `dune` not to compile the test in CI, so I updated the PR with a workaround...
(Runs on `trunk` will break, it was updated to 5.2, `ppxlib` must follow suite)
I finally got round to update that PR with your suggestions. It also takes care now to preserve the original backtrace.
Well, that’s quite a success... ~~I don’t get the same output when running the new internal tests locally and I don’t understand where this can come from.~~ I finally saw...
`sed` is indeed the path I was following :smile: I might have been a bit radical in my filtering (removing all progress lines), but I expect this is good enough...
> The PR now installs a handler on every `cmd` interpreted in a central hot path of both `Lin` and `STM`. This can potentially affect our ability to trigger parallel...
Proposition of an alternate design for that functionality: maybe we could install the catch-all exception handler only at the very start of the execution of the whole sequence of commands...
I just updated this PR with the latest proposal: using an inline `try with` at the various calling sites, sharing only the code that it is run when an exception...
I’ve added the suggested improvements because I’d like to use them to work on my cygwin branch. What I’m undecided about: - whether a time-out should appear as a test...
I wonder whether some recent modifications of the compiler might not be making some expected-to-fail tests require more runs to appear. To try and start testing that hypothesis, I just...