This is the mail archive of the mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH] PR libstdc++/13045 / C++ demangler, floating values and function pointer type return type postfix fix.

On Tue, Nov 25, 2003 at 09:44:38AM +0100, Gerald Pfeifer wrote:
> On Mon, 24 Nov 2003, Ian Lance Taylor wrote:
> > 1) Obviously, it is an ABI change, which means that code compiled with
> >    gcc 3.3.1 and 3.3.2 may not link with code compiled with gcc 3.4   
> >    and higher (versions of gcc before 3.3 are broken anyhow; support  
> >    for gcc 3.3 itself is available with -fabi-version=1).
> But GCC 3.4 will change the (C++) ABI anyway, so is this really an issue?

This is about changing the ABI right?
Changing the way floats are *mangled*?

Imho, for that to work you will need a way to specify
every significant bit in floating points (so the original
target form can in principle be reproduced).
That means indeed a form for SNaN, QNaN, +Inf, -Inf,
+0, -0, normalized and unnormalized representations
of values in a certain range, and of course the
number of bits in the exponent (with or without sign)
and the fraction (with or without sign).
And that is only for the IEEE format, other formats
might introduce more possibilities that I am not aware
of.  One concern with IEEE is that in the future
NaN's might be extended to define more error types.

Carlo Wood <>

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]