Martin Mitáš

> As part of using acutest for CI testing, it can be useful to record additional data beyond a simple success / failure. Does it make sense to log out...

Ok, Thanks for the explanation. I can now understand your use case and agree it's worth of adding. (As a side note, perhaps you might achieve the same via `--time`...

Hah, a race condition :-) And you were faster. I see the "output only on a failure" as a sane default for most use cases. I think your use case...

Ouch, I was just looking how the Xunit support is currently implemented in Acutest. Actually it's severely limited in that it's only generated by the main process, and only using...

Wait the patch actually seem to implement the redirection. Nice, so disregard that one point. (Sorry, that happens when making other things in parallel)

Maybe communicates the point with some particular CI engine, but not with others. Some products are really touchy when the XML cannot be validated against their XSD and refuse it...

Hmmm, after some thinking, I would split the problem into few steps: 1. Implement some infrastructure to better send **all** the output from the worker processes to the parent (if...

> any more thoughts on this PR? Nope at the moment. I'm slowly reworking the log infrastructure in a local branch. It just does not progress as fast as I...

Actually, so far I never did anything like that. But yes, I think so. For Unix/Posix platforms, at least. I am also not sure whether tracking `SIGWINCH` is worth the...

Humm, reopening. It is an issue. Having some API for that would be highly useful and lack of it seems as an unintended ommission to me, especially as we have...