Nicholas Clark wrote: > > Was this message about this patch: > > [ 11191] By: jhi on 2001/07/07 15:26:16 > Log: Subject: Re: [PATCH ext/POSIX/POSIX.pm] Re: sigaction.t under QNX > From: Norton Allen <allen@huarp.harvard.edu> > Date: Fri, 6 Jul 2001 12:01:14 -0400 (edt) > Message-Id: <200107061601.MAA00597@bottesini.harvard.edu> > Branch: perl > ! ext/POSIX/POSIX.xs > > ? > > Because I've just realised it's causing a segv on ARM Linux: > > $ ./perl ext/POSIX/sigaction.t > 1..18 > Segmentation fault > > SEGV cured by backing out 11191 and its two PTH_ co-patches only. > Not a solution, as QNX needs them. Yes, plus the bug addressed by the patches is not at all QNX-specific. QNX was just the first OS on which it tripped a test. > [quite a messy SEGV, as gdb says: > Program received signal SIGSEGV, Segmentation fault. > 0xc94d98 in ?? () > > and refuses to play ball at all in giving anything away about where this > is. Even attempting to force an early return from the function isn't working] Ouch. > I've no idea where, and I'm feeling somewhat shattered, so I'm not about > to commence tracing at assembly language level to find this one. > However, unless Jarkko's IPAQ is up to building perl, I guess I'm alone on > this one. (and Jarkko's probably busier than I am) :-( I wish I could help. -Norton