Gavin Sherry <swm(at)linuxworld(dot)com(dot)au> writes:
> I had a colleague test for me on his IRIX (6.5?) MIPSPRO machine. The
> following regression tests fail:
> http://treehou.se/~finn/regression.diffs
Hm, looks like strtod() has some bizarre behavior for input "infinity"
on that machine (NOT just failing to recognize it, but doing something
actively incorrect --- we've seen such bugs on Solaris).
I had a success report from someone else on Friday, so maybe this
problem is confined to specific IRIX versions?
regards, tom lane