Re: Name mangling once again

[Alexandre Petit-Bianco <apbianco at cygnus dot com>]

Martin Kahlert writes:

> I reported a problem with unresolved symbols on Jan, 31 and
> Alexandre Petit-Bianco told me, that it went away after updating his
> tree. I completely checked out the gcc tree today (Feb, 2) but the
> problem is not solved at all.

Actually there were other problems too. Bryce hacked on it and checked
something in (your upgrade most likely grabbed it.)

> _ZN4VHDL2VG3VCG8VCG_Node28class$VHDL$VG$VCG$VCGA_TitleE
> come from?

I don't know, but you're right, it's strange:

  apbianco@kazmo[~]: c++filt 
  _ZN4VHDL2VG3VCG8VCG_Node28class$VHDL$VG$VCG$VCGA_TitleE
  VHDL::VG::VCG::VCG_Node::class$VHDL$VG$VCG$VCGA_Title

I checked the way we mangle the class field and it seems to be fine,
i.e we always terminate the string properly, both in mangle.c and
jvgenmain.c. Is there a way I could get my hands on the bytecodes
you're trying to compile?

./A