Hi Jim,
> > - on AIX 4.3.2: compiles fine.
>
> didn't pass tests?
Passed one test, then starts consuming 100% CPU time:
Making check in chgrp
make check-TESTS
PASS: default-no-deref
It's inside the "chgrp/basic" test, the "sleep 1" command. "top" shows this:
PID USERNAME PRI NICE SIZE RES STATE TIME WCPU CPU COMMAND
9024 haible 0 20 340K 408K run 4:24 93.29% 91.15% sleep
"./src/sleep 1" run manually shows the same behaviour.
(Uhh, I already installed these binaries. I'm glad I have no "sleep" commands
in my .profile script :-))
Stacktrace under dbx:
(dbx) run 1
^C
Interrupt in glink.nanosleep at 0x10004a04 ($t1)
0x10004a04 (nanosleep+0x14) 4e800420 bctr
(dbx) where
glink.nanosleep() at 0x10004a04
unnamed block $b12, line 71 in "nanosleep.c"
rpl_nanosleep(??, ??), line 71 in "nanosleep.c"
xnanosleep(??), line 100 in "xnanosleep.c"
main(argc = 0, argv = (nil)), line 148 in "sleep.c"
(dbx) next
stopped in unnamed block $b12 at line 72 in file "" ($t1)
(dbx) next
stopped in unnamed block $b12 at line 71 in file "" ($t1)
(dbx) next
stopped in unnamed block $b12 at line 72 in file "" ($t1)
(dbx) next
stopped in unnamed block $b12 at line 72 in file "" ($t1)
(dbx) next
stopped in unnamed block $b12 at line 71 in file "" ($t1)
(dbx) next
stopped in unnamed block $b12 at line 72 in file "" ($t1)
(dbx) next
stopped in unnamed block $b12 at line 72 in file "" ($t1)
So it's an endless loop in lines 71-72 of nanosleep.c.
Bruno
