This one time, at band camp, Petr Salinger said:
> In the provided kdump, the child calls rfork,
> in linux it is named clone. The rest of child dump
> looks like thread manager, it is needed to also know
> trace of 7.
>
> >5 ruby1.8 CALL sigprocmask(SIG_SETMASK,0xbfbe3af0,0xbfbe3af0)
>
(the short version - the parent forks and waitpids, and never gets a
SIGCHLD or a return from waitpid. This appears to be because the child
never exec's after the fork).
Please could test patches from my previous mails:
Date: Fri, 6 Aug 2010 13:31:14 +0200 (CEST)
Date: Fri, 6 Aug 2010 18:45:06
On Fri, 06 Aug 2010, Petr Salinger wrote:
With my DSA hat on I have to say that I'm not entirely happy with what
we have so far.
The biggest pain for us currently is that puppet just does not work
reliably on kfreebsd (both i386 and amd64). This may be a ruby bug, but
>>>
With my DSA hat on I have to say that I'm not entirely happy with what
we have so far.
The biggest pain for us currently is that puppet just does not work
reliably on kfreebsd (both i386 and amd64). This may be a ruby bug, but
it's still really, really annoying.
Which version of ruby are you u
4 matches
Mail list logo
