it's not critical in prototyping code, but I wouldn't ship anything with that failure in place. However, something odd is taking place since it appears as though the exceptions aren't being thrown as expected. If you are relying on these to catch such cases as above then nothing good will happen.
On 10 October 2016 at 07:34, Michał Frątczak <m.fratc...@badibadi.com> wrote: > So I managed to build pyilmbase (CentOS 6.5 + Red Hat DTS 2.1 (g++ v4.8.2)), > and I run tests from file pyImathTest.py > > All of them pass except for testFpExceptions() which crashes with > following error: > > terminate called after throwing an instance of 'Iex_2_2::InvalidFpOpExc' > what(): Invalid floating-point operation. > Aborted (core dumped) > > > The crash happend on line 8038: > > f = sqrt (-1) > > It will pass if I remove preceding code fragment, starting at line 8023: > > # The overflow exception is not masked properly on Windows -- see > # Platform/PlatformFpu.cpp and note the non-implementation of > # setFpExceptions(). > # > if sys.platform != 'win32' and sys.platform != 'darwin': > try: > v = V3d (1e200, 1e200, 1e200) > v = v * v * v > print v > except iex.MathExc: > pass > else: > assert 0 > > > Is this smth I should worry about? Any hints appreciated. > > > thanks, > > -michal > > _______________________________________________ > Openexr-devel mailing list > Openexr-devel@nongnu.org > https://lists.nongnu.org/mailman/listinfo/openexr-devel > >
_______________________________________________ Openexr-devel mailing list Openexr-devel@nongnu.org https://lists.nongnu.org/mailman/listinfo/openexr-devel
