Re: Internal Compiler error with C++

[Mike Stump <mrs at windriver dot com>]

> Date: Wed, 20 Oct 1999 11:18:03 +0200
> From: "Olivier Bonningues" <bonnin@nortelnetworks.com>

> We use the cygnus C++ compiler for PowerPC in the Tornado
> environnement of WindRiver System compagny.

This is confusing.  What is less confusing is, `we use the Tornado 2.0
C++ compiler from WindRiver'.  Cygnus also supplies compilers, but the
one you are using didn't come from them (not their bug).

> Do you know this problem ?

Nope, not before now.

> Is it fixed and solved in a new release of the compiler ?

Hum...  There isn't a newer release from WindRiver, so no, not from
them.  There is a newer release of the compiler from the FSF, called
gcc-2.95, I suspect it doesn't have this bug, so yes, if you consider
FSF compilers.

> Is there a workaround ?

Yes:

  if (coding.putName("ABCD")) {
    if (RTDataWrapper_char( pv->Tei ).put( coding )) {
      if (RTDataWrapper_char( pv->Cell ).put( coding )) {
        return 1;
      }
    }
  }
  return 0;

will cause it to work, there may be less drastic ways to make it work
as well.  I just tried the simplest form of it.  I'd suspect that
using the work around will be easier than trying to get and install
gcc 2.95.  If you can't use this work around and don't want to build
2.95 and if you haven't filed an spr yet, you may want to file an spr
with WindRiver.

> It will be very hepful to us to have any information about this problem.

Sure, happy to help.