Hi!
----
[Please consider this a low-priority bug because it's about old gcc3.x
on old Solaris 11/B145 (some of the issues may still be interesting...
but please please not now). Due lack of bugzilla I file it here in the
list... marked as priority=minor, investigate=after-next-beta]
I build a 64bit built (CC='gcc -std=gnu99 -m64') of ast-ksh.2012-05-18
with the OS-suplied gcc version ("gcc (GCC) 3.4.3
(csl-sol210-3_4-20050802)") on Solaris 11/B145 on an AMD64 machine and
hit a few intersting bugs in "arith.sh":
-- snip --
$ (CC='gcc -std=gnu99 -m64' ./bin/package make) 2>&1 | tee -a buildlog.log
$ (export SHELL=$PWD/arch/sol11.i386/bin/ksh ; $SHELL
./src/cmd/ksh93/tests/shtests src/cmd/ksh93/tests/arith.sh ))
test arith begins at 2012-05-31+03:44:44
arith.sh[333]: let with zero filled fields not working
arith.sh[456]: printf '%g\n' $((Inf)) failed -- expected inf, got -inf
arith.sh[458]: printf '%g\n' Inf failed -- expected inf, got -inf
arith.sh[456]: printf '%g\n' $((Nan)) failed -- expected nan, got -nan
arith.sh[458]: printf '%g\n' Nan failed -- expected nan, got -nan
arith.sh[456]: printf '%g\n' $((1.0/0.0)) failed -- expected
inf, got -inf
arith.sh[458]: printf '%g\n' 1.0/0.0 failed -- expected inf, got -inf
arith.sh[469]: pow(1.0,Inf) != 1.0
arith.sh[486]: $(( -1.0*0)) should be -0
arith.sh[487]: %g of x $x $((x)) for x=-0 should all be -0
arith.sh[488]: %g of y $y $((y)) for y=-0.0 should all be -0
arith.sh[501]: float t=copysign(0.0,0.0) => 0.0 -- expected 0.0
arith.sh[501]: float t=copysign(0.0,0.0) => 0.0 -- expected 0.0
arith.sh[501]: float t=copysign(0.0,0.0) => 0.0 -- expected 0.0
arith.sh[501]: float t=copysign(0.0,-0.0) => 1.0 -- expected 0.0
arith.sh[501]: float t=copysign(1.0,0.0) => 1.0 -- expected 0.0
arith.sh[501]: float t=copysign(1.0,0.0) => 1.0 -- expected 0.0
arith.sh[538]: ((a[++x]++)) should only increment x once
arith.sh[539]: a[1] not incremented
arith.sh[563]: (a[0]+a[1])!=1000
arith.sh[662]: unset compound array variable error with for
loop optimization
test arith failed at 2012-05-31+03:44:44 with exit code 21 [ 200 tests
21 errors ]
-- snip --
These may be compiler bugs in one form or another which need to be
worked around for older platforms... but AFAIK not for the next beta
(please).
----
Bye,
Roland
--
__ . . __
(o.\ \/ /.o) [email protected]
\__\/\/__/ MPEG specialist, C&&JAVA&&Sun&&Unix programmer
/O /==\ O\ TEL +49 641 3992797
(;O/ \/ \O;)
_______________________________________________
ast-developers mailing list
[email protected]
https://mailman.research.att.com/mailman/listinfo/ast-developers
