Re: Infinite loop with GCC 4.3

[Trevor Daniels]

Han-Wen Nienhuys wrote Monday, May 26, 2008 2:49 PM
> 2008/5/26 Joe Neeman <address@hidden>:
> > The position of the gcc folks is that this is Intel's fault and so they
> > will not fix it. This leaves us with a few options:
> > 1) Change Real to be a long double instead of a double (on x86 only?)
> > 2) Compile with -msse on x86
> > 3) Find some way of forcing a truncation (as suggested in comment #88 of
> > the above link)
> > 4) Anything else?
>
> comment #60 sounds promising too. We'd have to have a small bit of
> system dependent code, but it is localized, so I can live with it.

Comment #60 suggests setting the x87 control word to x027f (rather
than the default x037f) which instructs the FPU to perform its
calculations in double precision rather than double-extended
precision.  This seems both the easiest and safest way of avoiding
the problem, as double-extended precision is hardly likely to be 
required for LilyPond's calculations.   This suggestion should
work as long as the FPU is not re-initialised again, since this would 
reset the control word back to the default value.  I don't know when this 
happens, but it might be at the start of a compilation unit, which means
the control word would need to be set in every compilation unit.

Trevor
 
> --
> Han-Wen Nienhuys - address@hidden - http://www.xs4all.nl/~hanwen
>
>
> _______________________________________________
> lilypond-devel mailing list
> address@hidden
> http://lists.gnu.org/mailman/listinfo/lilypond-devel