labath added a comment. In http://reviews.llvm.org/D13362#258972, @tfiala wrote:
> I've fixed: > https://llvm.org/bugs/show_bug.cgi?id=25019 > > I think for now I am not interested in trying to tackle the intent of this > change as it unduly complicates the timeout detection logic. > > I am okay with saying: > "If you run a process http://reviews.llvm.org/P1, and that process creates > child processes C1..CN, and shares the stdout/stderr file handles from > http://reviews.llvm.org/P1 to C1..CN, and if http://reviews.llvm.org/P1 > exits, we don't detect the exit until all stdout/stderr handles shared with > C1..CN are closed." That's just a bad test if it is leaving children around. > It will time out. Sounds good to me. This motivates people to write correct tests, which I think is good. My main concern was not leaving those children around after we time out, which I believe you fixed already. http://reviews.llvm.org/D13362 _______________________________________________ lldb-commits mailing list lldb-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
