----- Original Message ----- From: "Gary R. Van Sickle" <[EMAIL PROTECTED]>
> > Hmm, strange. > > > > Is that consistent regardless of the souce type - > > local/installfromnet/download only? > > The SIGSEVs are, but where they come from varies: > > install from net and download only: > choose.cc line 541: "if (!pkg->Categories.number ())" > > pkg == 0x330009 as I look at it right now, seems "un-pointerlike", but obviously > not NULL. Looking at in the local variable window shows a lot of 0x00's, name > and installed_from prts == 0x00, and... gdbtk just crashed so I can't tell you > more ;-). Ah, this is what I was cursing about before. I think I'll be warming up my gdb debugging skills shortly, if I can't figure out what I'm doing that is killing gdb/gcc. (It may be cruddy c++ output?). > local: > gdb locks up in the same manner it did with the endless loop you fixed before. > I don't know if that's what it is or not, but that's what it behaves like. If > not running under gdb, it'll GPF. It's not that. I'm tracking this down (sortof) at the moment. Rob