• remove unnecessary if-clang conditions

• follow kib, take 2

• follow kib's suggestion

@acidburn0.pt_gmail.com: Are you Jason Evans, author of jemalloc?

Wow, that API is unfortunate.

Was this tested with an MTU of 9000?

Thanks for the man page!

It would be nice to include some documentation. The systemd man page seems most useful and canonical, but it's LGPL. Maybe include a link to it in a comment in the generated file?

• wait() for children instead of a sleep race

• rebase

• fixup: make test more robust by opening more files

That's not much of a test plan. Perhaps try sudo?

See D21820. What do you think?

That your fix is better than mine :-)
So feel free to overwrite my changes with yours.

Coincidentally, I have a different fix for this in a work-in-progress tree. My change keeps the current value of kern.compress_user_cores and decompresses the core file. It has the advantage of providing test coverage for kern.compress_user_cores (if the user enables it). I can't find any existing test coverage.

However, the kernel improved a lot since this revision was created and a repeated test has higher chances of showing the difference (if any).

*facepalm* this is really happening? Oy...