* Stefan Seifert -- Saturday 03 December 2005 01:24:
> as discussed already on IRC, there seems to be another gcc 4.0.2/SUSE
> 1.0 problem:
> engine sounds on the 737, Concorde and every other plane that uses
> thrust_lb[0] as base for the engine sound calculation don't work on this
> platform.
It turned indeed out to be yet another one of these ugly aliasing
bugs with gcc 4.0.2 ([1] *AND* [2]!). In fastmath.hxx apparently
reinterpret_cast doesn't work appropriately. Could explain why Alex'
clean patch didn't work. Now the question is: should fgfs work
around a broken gcc release, when there's hope that the next version
will be fixed? Or is it not a bug, but something that will be in
all gcc versions for the next time (I've not found any bug reports
about it yet!). Anyway, attached is a patch that makes the
turbines whine again, and some of us whine less.
m.
[1] 4.0.2 20050901 (prerelease) (SUSE Linux)
[2] 4.0.2 20050808 (prerelease) (Ubuntu 4.0.1-4ubuntu9))
diff -u -p -r1.4 fastmath.hxx
--- fastmath.hxx 8 May 2004 12:58:29 -0000 1.4
+++ fastmath.hxx 3 Dec 2005 14:37:33 -0000
@@ -54,16 +54,18 @@ void fast_BSR(float &x, register unsigne
inline float fast_log2 (float val)
{
- int * const exp_ptr = reinterpret_cast <int *> (&val);
- int x = *exp_ptr;
- const int log_2 = ((x >> 23) & 255) - 128;
- x &= ~(255 << 23);
- x += 127 << 23;
- *exp_ptr = x;
+ union {
+ float f;
+ int i;
+ } v;
+ v.f = val;
+ const int log_2 = ((v.i >> 23) & 255) - 128;
+ v.i &= ~(255 << 23);
+ v.i += 127 << 23;
- val = ((-1.0f/3) * val + 2) * val - 2.0f/3; // (1)
+ v.f = ((-1.0f/3) * v.f + 2) * v.f - 2.0f/3; // (1)
- return (val + log_2);
+ return (v.f + log_2);
}
_______________________________________________
Flightgear-devel mailing list
Flightgear-devel@flightgear.org
http://mail.flightgear.org/mailman/listinfo/flightgear-devel
2f585eeea02e2c79d7b1d8c4963bae2d
