[980205]: Patch for wrong mangling (was: Re: [980129]: bug in COFF's name mangling?)

[Manfred Hollstein]

On Wed, 4 February 1998, 16:33:46, Manfred.Hollstein@ks.sel.alcatel.de wrote:

 > On Tue, 3 February 1998, 14:25:34, Manfred.Hollstein@ks.sel.alcatel.de wrote:
 > 
 >  > Hi,
 >  > 
 >  > I'm just trying to build egcs on my various systems. Solaris and SunOS
 >  > are  finished and work ok, but  on m88k-motorola-sysv3  (a COFF system
 >  > with SDB debugging info)   I'm  running into trouble   while  building
 >  > libstdc++, namely the bigstmp-complx stuff. I don't know, if this is a
 >  > common problem or something m88k specific, the build  on my other COFF
 >  > system (m68k-motorola-sysv) usually takes 3.5 days (and nights) and is
 >  > still running.
 >  > 
 >  > Initially GNU as was complaining about `C_EFCN out of scope'; this one
 >  > could be avoided  by using Motorola's native assembler  - but this one
 >  > then failed later on on some GNU as extensions used for *comio.o :-(
 >  > 
 >  > Anyway, looking at fcomplex.o on this system shows _two_ names for one
 >  > address:
 >  > 
 >  > $ nm --dem fcomplex.o | grep __doami
 >  > 00001030 T __doami(complex<float> *, complex<float> const &)
 >  > 00001030 T complex<float> & __doami<float>(complex<float> *, complex<float> const &)
 >  > 
 >  > while e.g. on SunOS I'll only get _one_ function name:
 >  > 
 >  > $ nm --dem fcomplex.o | grep __doami
 >  > 00001300 T complex<float> & __doami<float>(complex<float> *, complex<float> const &)

This is not only a problem with COFF/sdb, but also e.g. on Solaris and SunOS:

$ strings -a fcomplex.o | grep __doami
__doami__FPt7complex1ZfRCt7complex1Zf:F(13,3)
__doami__H1Zf_Pt7complex1ZX01RCt7complex1ZX01_Rt7complex1ZX01

As you can see, the STABS entry for the particular function uses a differently
mangled name than the symbol itself.

Doing this with an fcomplex.o from egcs-1.0's libstdc++.a gives:
__doami__H1Zf_Pt7complex1ZX01RCt7complex1ZX01_Rt7complex1ZX01:F(13,3)
__doami__H1Zf_Pt7complex1ZX01RCt7complex1ZX01_Rt7complex1ZX01

The append patch fixes it for me in both debugging worlds. But, since I'm far
away from being a C++ frontend expert, Mark Mitchell or Jason Merrill,
could you please check, whether the patch is ok or the bug is caused elsewhere?
BTW, you can reproduce the problem with the attached `bug.cc'.

Mon Feb  9 12:07:56 1998  Manfred Hollstein  <manfred@s-direktnet.de>

	* decl.c (grokdeclarator): Use DECL_TEMPLATE_INSTANTIATION as well
	when deciding to override DECL_ASSEMBLER_NAME.

diff -c -p egcs-980205.orig/gcc/cp/decl.c egcs-980205/gcc/cp/decl.c
*** egcs-980205.orig/gcc/cp/decl.c	Tue Feb  3 23:52:39 1998
--- egcs-980205/gcc/cp/decl.c	Mon Feb  9 12:07:31 1998
*************** grokdeclarator (declarator, declspecs, d
*** 9930,9938 ****
  	  return NULL_TREE;
  
  	if (ctype == NULL_TREE && DECL_LANGUAGE (decl) != lang_c
! 	    && (!DECL_TEMPLATE_SPECIALIZATION (decl) ||
! 		name_mangling_version < 1)) 
! 	  DECL_ASSEMBLER_NAME (decl) = current_namespace_id (declarator);
  	
  	if (staticp == 1)
  	  {
--- 9930,9939 ----
  	  return NULL_TREE;
  
  	if (ctype == NULL_TREE && DECL_LANGUAGE (decl) != lang_c
! 	    && (! (DECL_TEMPLATE_SPECIALIZATION (decl) ||
! 		   DECL_TEMPLATE_INSTANTIATION (decl)) ||
! 		name_mangling_version < 1))
!   	  DECL_ASSEMBLER_NAME (decl) = current_namespace_id (declarator);
  	
  	if (staticp == 1)
  	  {