This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
Re: dbxout.c stabs output problem
- To: Jim Wilson <wilson at cygnus dot com>
- Subject: Re: dbxout.c stabs output problem
- From: Henry Sobotka <sobotka at axess dot com>
- Date: Sat, 25 Jul 1998 14:11:41 -0700
- CC: law at cygnus dot com, egcs-bugs at cygnus dot com, ian at cygnus dot com
- References: <199807240215.TAA09630@rtl.cygnus.com>
Thank you very much for the clarifications and patch.
It might help others in the future if a couple sentences were added to
stabs.texinfo to document the two changes we've identified.
1) Given:
typedef int foo;
typedef int bar;
typedef foo *pfoo;
and:
.stabs "int:t1=r1;-2147483648;2147483647;",128,0,0,0
Old stabs:
.stabs "foo:t1",128,0,1,0
.stabs "bar:t1",128,0,2,0
.stabs "pfoo:t20=*1",128,0,3,0
New stabs:
.stabs "foo:t20=1",128,0,1,0
.stabs "bar:t21=1",128,0,2,0
.stabs "pfoo:t22=23=*20",128,0,3,0
2) With C++, the use of a void last argument to signal non-varargs.
As well as fix: "The single '#' stab ends at the next ';', not the next
'#' as stated by stabs.texinfo." [Ian Taylor]
Also, I think Ian's detailed breakdown of the C++ stabs for the longlong
example would be a valuable addition to the manual as it's more
comprehensive than anything in the C++ section, and shows how internally
a simple struct becomes a class with basic assignment and copy
constructors.
If no one there has the time and/or inclination to do it, I'd be glad to
submit a patch.
Henry