On 03/27/2013 10:18 PM, Jeremy Harris wrote: > On 03/27/2013 07:53 PM, Todd Lyons wrote: >> #12 0x0000000000487edb in smtp_notquit_exit () >> #13 0x0000000000474b87 in receive_bomb_out () >> #14 0x0000000000474bdf in ?? () >> #15 <signal handler called> >> #16 0x00007f2fb03a4bd8 in *__GI___poll (fds=0x7fff2f62b540, nfds=1, >> timeout=<value optimized out>) at ../sysdeps/unix/sysv/linux/poll.c:83 >> #17 0x00007f2fb1d7f8de in send_dg (statp=0x7f2fb0640300, buf=<value >> optimized out>, buflen=<value optimized out>, buf2=0x0, buflen2=<value >> optimized out>, >> ans=0x7fff2f62c110 "?큂", anssiz=1024, ansp=0x7fff2f62c970, ansp2=0x0, >> nansp2=0x0, resplen2=0x0) at res_send.c:1062 >> #18 __libc_res_nsend (statp=0x7f2fb0640300, buf=<value optimized out>, >> buflen=<value optimized out>, buf2=0x0, buflen2=<value optimized out>, >> ans=0x7fff2f62c110 "?큂", >> anssiz=1024, ansp=0x7fff2f62c970, ansp2=0x0, nansp2=0x0, resplen2=0x0) >> at res_send.c:553 >> #19 0x00007f2fb1d7da95 in *__GI___libc_res_nquery >> (statp=0x7f2fb0640300, name=<value optimized out>, class=<value >> optimized out>, type=<value optimized out>, > >> Thoughts? Am I missing a bigger picture? > > Traditionally (and I'm talking about the 1980's here) about the *only* > thing one could portably do in a signal-handler was to set an int-sized > flag and return. Things may have progressed a little since then but > attempting a dns-lookup does seem a trifle unwise...
well I don't think the rules have changed in reality - imho doing anything non-trivial from a signal handler (and a dns lookup is very far from trivial) is doomed and simply won't work. Stefan -- ## List details at https://lists.exim.org/mailman/listinfo/exim-dev Exim details at http://www.exim.org/ ##
