Steven Chamberlain, on Tue 23 Feb 2016 11:45:11 +0000, wrote:
> Mathieu Malaterre wrote:
> > Previous patch disabled the functionality completely. We should use
> > Steven's patch instead. Re-opening issue (Thanks Steven!)
>
> My code was specifically for kfreebsd; it doesn't help hurd, but maybe
> it has some equivalent?
We haven't written the equivalent bits, no. So both patches make sense:
checking the presence of headers, and using whatever is available,
or fallback to disabling the functionality (and Hurd will get it
automatically when it adds the header support).
> Also, I'd prefer that ilmbase builds first on kfreebsd (with this
> feature disabled) and migrates to testing, before we try this.
Note that it'll need a symbol update to get built. Also it seems the
testsuite has issues with comparing floats:
on hurd-i386 and kfreebsd-i386:
lt-ImathTest: testBoxAlgo.cpp:892: void {anonymous}::boxMatrixTransform():
Assertion `b21 == b2' failed.
on m68k:
lt-ImathTest: testBox.cpp:253: void {anonymous}::testExtendByPoint(const char*)
[with T = Imath_2_2::Vec2<float>]: Assertion `b.min == p && b.max == p' failed.
It should probably use approximated comparison, shouldn't it?
Samuel
