From: McLean, M. <Mike.McLean@PearsonTC.com> - 2003-05-19 13:35:32
|
>From: Bjoern Fernhomberg [mailto:li...@fe...] > >McLean, Mike wrote: >> I can confirm the problem. I reported this on this development >> email list a >> few weeks ago, but didn't submit a patch. (Since my patch was an >> "emergency" >> patch that simply commented out the log call). >> >> I think the more general fix would be to add null pointer protection in >> logger (ala the "included" libatalk/compat/snprintf.c); however that could >> be difficult given the va_args stuff. > >Right. I'm not sure there are no other places in the source where a NULL >might eventually get passed, so we should remove the cause and not the >symptoms :-) > >Anybody willing to work on this ? >Otherwise we could use plp_vsnprintf if we're building on Solaris. >IIRC this problem is Solaris vnsprintf specific, other platforms seem to >handle this just fine. >Comments ? plp_vsnprintf is the default from configure in our Solaris 2.5.1 installation and it works. The problem only occurs (for us) in Solaris 2.6 and above when configure recognized the Sun-snprintf and therefore defined HAVE_SNPRINTF. The simplest solution may be to always undefined HAVE_SNPRINTF on Solaris. **************************************************************************** This email may contain confidential material. If you were not an intended recipient, please notify the sender and delete all copies. We may monitor email to and from our network. **************************************************************************** |