-finline core dumps egcs-1.0.1...

[Mike Simons]

  Have any known -inline bugs been fixed in the last month?

from  mysql-3.21.23-beta/sql
file  sql_yacc.cc
on    SunOS mars 5.5 Generic_103093-02 sun4u sparc SUNW,Ultra-1
with  gcc version egcs-2.90.23 980102 (egcs-1.0.1 release)

  compiling '-O9' core dumps the compiler...
  compiling '-O9 -fno-inline' produces something... (yet to be tested)

  Adding -save-temps shows that it dies while producing assembly output:
{standard input}: Assembler messages:
{standard input}:526: Warning: end of file not at end of a line; \
  newline inserted
{standard input}:995: Error: Unknown pseudo-op:  `.st'
c++: Internal compiler error: program cc1plus got fatal signal 11

  the .yy file is about 1700 lines.
  the .cc file is about 6700 lines, and includes about 30 other headers. 
  the .ii file is about 23000 lines.

  I took a shot at reducing the .cc and/or .ii file down to something
more reasonable, but nothing small is going to come out of it soon.  =)
I *might* be able to trim the .ii file down to 5000 lines in two hours
and still produce the bug...


In addition to what I've said already what would be most useful?

Does anyone want a copy of those core files?

Would a back trace and local variable dumps from of the core file help?

-- 

    Thanks,
      Mike Simons
      Science Applications International Corporation
      msimons@saic1.com                  703-925-5674