colonel@gloria.UUCP (Col. G. L. Sicherman) (07/06/85)
Has anybody seen this one before? Run this through f77 -O under 4.2bsd:
implicit integer (a-z)
common / people / one, four, two, eight, ten
one = 1
four = 4
two = 2
eight = 8
five = or(one,four)
a = 1
b = 2
c = 3
d = 4
e = 5
f = 6
ten = or(two,eight)
print *, ten
stop
end
The answer will be "5". Looks like trouble with common subexpressions
that aren't common.
--
Col. G. L. Sicherman
...{rocksvax|decvax}!sunybcs!colonelwoods@hao.UUCP (Greg Woods) (07/09/85)
> Has anybody seen this one before? Run this through f77 -O under 4.2bsd: > .... Looks like trouble with common subexpressions that aren't that common The common expression bugs in the optimizer are well-documented and have been discussed before in this newsgroup. Donn Seeley's bug fixes (which have also been well-documented in this newsgroup) repair many of these bugs. I suggest you install them. Your program works fine on our system, optimized, *with* the bug fixes. --Greg P.S. I think Donn's net address is utah-cs!donn (correct me if I'm wrong, Donn) -- {ucbvax!hplabs | allegra!nbires | decvax!noao | harpo!seismo | ihnp4!noao} !hao!woods CSNET: woods@NCAR ARPA: woods%ncar@CSNET-RELAY