To: awandelt AT rs1 DOT thch DOT uni-bonn DOT de (Andreas Wandelt)
Cc: djgpp AT sun DOT soe DOT clarkson DOT edu
Subject: Complex class in the c++ part of 1.11
Date: Tue, 04 Jan 94 10:02:33 PST
From: Jonathan Ryshpan <jon AT halsp DOT hitachi DOT com>

>Someone reported a bug concerning the Complex class. I am getting the
>same errors for both the Complex and the String class, so that at the
>moment I cannot do any c++ with djgpp1.11 :-( . I know that the same
>problem does exist under linux with the gcc, and that it DID exist on
>our RS6k workstation. On the latter, it went away after a libgpp update
>(libgpp 2.5.2). Under linux, though, this didn't help.

>The error messages i get concern missing pragmas and such, so i assume
>it is a library problem, and not one of the compiler.

Gcc-2.5.5 running on a Unix PC (ISC 3.01) can't perform complex
arithmetic.  The test program tComplex.cc compiles without errors; 
but tComplex crashes executing the statement:

	Complex e = pow(a, 2);

This is very likely because of the following problem (taken from the
file PROBLEMS of the Gcc-2.5.5 sources):

	124. On the 386, bad code results from foo (bar()) when bar
	returns a double, because the pseudo used fails to get preferenced
	into an fp reg because of the distinction between regs 8 and 9.

I don't have any problems with the String class.

Good Luck:

Jonathan Ryshpan                <jon AT halsp DOT hitachi DOT com>

	===============> I write for myself ONLY. <===============