Cristian Cadar

Thanks, @kren1 . let's reopen this.

@andrewvaughanj thanks for the useful info! I'm afraid that change doesn't help (I've tried it quickly), but even if it did, I think we need to understand better what goes...

Great to see several workarounds here. @MartinNowack I guess it would be useful to understand whether this is an LLVM issue, and if so, to report it. We should also...

@MartinNowack thanks for this. I have finally looked at this patch, which is a bit tricky (in particular, I find the ObjectState class to be poorly documented, I'm currently doing...

@kren1 thanks for contributing this, it's great to have it as a PR here. I guess the question now is how feasible is to merge it, or parts of it,...

This is indeed the right solution, thanks, @251. Most likely the best thing would be to have the current `read` model be a special case of `pread`. For now I...

It is a bug in KLEE, only that @paulmar already knew about it. Needs to be fixed, thanks for the report!

I ran into this bug again. It is essentially a duplicate of #189 A simplified example of the above, which still shows the bug (checked with LLVM 8): ``` #include...

Good catch. We should definitely implement one of these solutions. I guess the setitimer is indeed more flexible. Do you want to write a patch implementing one of them?

@251 I believe this very old issue was solved by the new time API you contributed a couple of years back. This being said, I still see a few `alarm`...