From: GARY GENDEL (ggendel_at_sarnoff.com)
Date: Wed Jun 18 2003 - 14:57:43 BST
Thanks Flemming. That patch definately helps. However, ups doesn't seem to catch low level exceptions. For example, I have a program with a Floating exception. When the exception occurs, it dumps core and exits. Ups didn't catch it. Gary Flemming Madsen wrote: > On rereading this I come to think of: > > I have had some success with simply disabling Ups's sanity checks > Apparantly it can perfectly well survive a couple of > > Fatal internal error: bad name in parse_name (aborting) ... > > with the following: > > --- lib/libukcprog/panic.c.orig 2002-10-29 13:54:38.000000000 +0100 > +++ lib/libukcprog/panic.c 2003-04-24 09:52:20.710989000 +0200 > @@ -50,5 +50,5 @@ > > fprintf(stderr, "Fatal internal error: %s (aborting) ...\n", > message); > fflush(stderr); > - /*abort();*/ > +/* abort(); */ > } > > > > > GARY GENDEL wrote: > >> Greg, >> >> I have successfully built ups using Sun's workshop compiler version >> 7.0 with Flemming's patch (included, but you'll need some simple hand >> edits because the patch was for an earlier version of UPS). It works >> great with C code, but C++ code causes it to crash when I try to >> interrogate a variable (by clicking the code). I have similar results >> when debugging Sun C++ or Gnu g++ compiled code. I'm afraid it looks >> like my favorite debugger may no longer support recent Sun platforms. :^( >> >> Gary >> >> Greg Wagner (TTR) wrote: >> >>> Hello, >>> >>> We recently upgraded several of our machines to SUN OS 2.8 so I down >>> loaded and built ups-3.35 last week. All I did to make it was: >>> >>> % configure >& configure.log >>> >>> % make ups >& make.log >>> >>> When I ran ups it dumped core. >> >> >> >> >> [snip] >> >> > > >
This archive was generated by hypermail 2.1.4 : Fri Jun 20 2003 - 17:40:39 BST