I just had 'sleep 2' coredump on me...
After investigation I found it explode inside the 'pow' maths function
Looking for where it found 'pow' I find it using the one declared by /lib/libmp.so.1
When I looked at the neighbouring Sol2.6 I find that the function seems to have been renamed to mp_pow so it looks like a hiccup on Suns part for 2.5.1.
If I prepended -lm to the linking stage everything works fine.
The root of the problem is that configure found a pow() so didn't use libm, but didn't notice the pow() was broken.
I spiked the configure script with a #error in the relevant code fragment and configure fell back to the pow in the maths lib and everything worked.
I don't know how you would deal with this, and I'm no C whizz. It looks like a special case is needed in configure.
Also, while I'm here, the 'make check' for stty fails when checking 'stty size' for '0 0' after running 'stty rows 0 cols 0' as Solaris spat an IO error back.
The Solaris stty does exactly the same, so it doesn't look like a bug in the program, just the test. The HPUX gave '0 0' happily.
I'll be happy to help if you need any more info - I'll keep the compile tree on ice for now.
Faithfully,
DW
PS: Incidentally - I'm happily compiling three solarii versions and a hpux in the same source tree at the same time curtesy of the GNU configure process and merging the results seamlessly thanks to --exec-prefix, and its worked every time!!
God Bless you all!!!
