njw@doc.imperial.ac.uk (11/26/90)
(People run screaming into the night after reading the subject line... :-) Has anyone got this to work? Is it a workable idea to start with? Reason for madness: So I can debug ATK programs. Platform: hp9000s300, hpux7.0, gcc1.37.1, ATK 6+bit Things start breaking when building class. Firstly, the makefiles assume you will be using /lib/libc.a, etc. Not the case with gcc & hpux (due to robotussin'd libs). After hacking this about, I then get into difficulties with it making its "transitive closure awk script" (Gawd!) The awk script churned for 3 hours (I later found out), but produced no output. HoHum. So. Is it even feasible to try this? Or does the dynamic loading/linking *need* to run "native"? If I cannot use gcc to build, then how do I debug ATK programs? I have been going under the assumption that the standard debuggers for hpux do not understand the dynamic loading of Andrew. ([Shame] yes, I need a post-mortem debugger to test my programs... they don't *ALL* work perfectly first time :-) Nick, Enjoying the wonders of hpux.... ______________________________________________________________________________ Nick Williams. njw@doc.ic.ac.uk ... Dept of Computing, Imperial College, London SW7 2B7. UK njw@athena.mit.edu ... Project Athena, M.I.T., Cambridge, MA 02139. USA Any opinions or views you may find hiding in this message are mine, and not policy, intent, ideas, twinklings of eyes, or anything at all related with my current organisation, unless specifically noted as such.
wdc@ATHENA.MIT.EDU (Bill Cattey) (11/28/90)
Suggestion: Don't use robotussin'ed gcc. Use the new stuff from ohio state. It is a radically different coff encapsulation, and it seems to coexist nicely with both gdb and dbx on the PS/2. That's how we're trying to get it to work on the PS/2. We hope to have a working gcc/gas/gdb/atk suite as soon as NFS stays up long enough for us to complete the build and fix the little bugs in gdb. I must strongly emphasize that things are only SEEMING to work so far. I have been unable to keep the PS/2 up long enough for rigorous testing. -wdc