Performer2.0 and Purify?

New Message Reply Date view Thread view Subject view Author view

Daniel C. Williams X-2453 (dcw++at++sarnoff.com)
Tue, 27 Feb 1996 10:26:45 -0500 (EST)


Can anyone report on experiences using Purify and Performer 2.0?

I'm running Irix 5.3 and using IrisGL on a 3-pipe Onyx RE2.

I've created an executable that is mostly linked with DSOs; I was
forced to link statically with libpf_iv.a and call pfdLoadFile_iv
directly instead of going through the loader DSO lookup because that
gave me a runtime error that purify couldn't locate the purified
version of the DSO that Performer tried to load.

Anyway, here's the runtime error I now get:

ZPW: Zero page write:
  * This is occurring while in:
        _lmalloc [amalloc.c:684]
        _amalloc [amalloc.c:76]
        pfMemory::operator new(unsigned int,unsigned int,void*) [pfMemory.C:94]
        pfMemory::malloc(unsigned int,void*) [pfMemory.C:369]
        pfMalloc [cMemory.C:31]
        pfdNewShare [pfdShare.c:77]
        pfdConvertFrom_iv [pfiv.C:2199]
        loadIv(char*) [pfiv.C:382]
        pfdLoadFile_iv [pfiv.C:171]
        pfGeometry::pfGeometry(RWCString&,const char*,const int) [pfGeometry.c++:31]
        DisplayWorker_i::createGeometry(const char*,const char*,CORBA_Environment&) [dwgeometry.c++:25]
        DisplayWorkerDisplayWorker_i::createGeometry(const char*,const char*,CORBA_Environment&) [DisplayWorker_i.h:97]
  * Writing 4 bytes to 0xc

Of course, this causes an immediate core dump.

It seems like more than just a coincidence that the problem is
occuring in the very library I had to treat specially to get it
to run in the first place.

It happens even when I force all stages into a single process.

I'd be grateful for any hints or clues.

Dan

-- 
Daniel Williams, Consultant to: David Sarnoff Research Center
Voice: (609) 734-2153    Email: dcw++at++sarnoff.com, dan++at++sass.com

New Message Reply Date view Thread view Subject view Author view

This archive was generated by hypermail 2.0b2 on Mon Aug 10 1998 - 17:52:27 PDT

This message has been cleansed for anti-spam protection. Replace '++at++' in any mail addresses with the '@' symbol.