----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/61799/#review184060 -----------------------------------------------------------
Fix it, then Ship it! 3rdparty/libprocess/src/subprocess_posix.hpp Line 194 (original), 195 (patched) <https://reviews.apache.org/r/61799/#comment260061> Hmm. I thought about this some more and I think we should retain the `strerror` here. 1. We still used it in other async-signal-safe situations. 2. In glibc it only mallocs when you give it an unknown error number. 3. In practice on Linux, malloc after fork is safe. 4. The error message is pretty useful. Whay do you think? - James Peach On Aug. 21, 2017, 8:54 p.m., Andrei Budnik wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/61799/ > ----------------------------------------------------------- > > (Updated Aug. 21, 2017, 8:54 p.m.) > > > Review request for mesos, Alexander Rukletsov, Alexander Rojas, Benjamin > Bannier, Benjamin Mahler, and James Peach. > > > Bugs: MESOS-7791 > https://issues.apache.org/jira/browse/MESOS-7791 > > > Repository: mesos > > > Description > ------- > > Using ABORT causes coredump when path to an executable is invalid. > In addition, using of _EXIT allows us to get rid of string > concatenation, thereby making `childMain` more async-signal safe. > > > Diffs > ----- > > 3rdparty/libprocess/src/subprocess_posix.hpp > b478beccff22e2ffa08b9b91cfd5b9c6ada9b697 > > > Diff: https://reviews.apache.org/r/61799/diff/3/ > > > Testing > ------- > > sudo make check (mac os x, fedora 25) > internal CI > > > Thanks, > > Andrei Budnik > >
