test/nilptr/arrayindex1.go spins forever on darwin/snow leopard #265
Description
Before filing a bug, please check whether it has been fixed since
the latest release: run "hg pull -u" and retry what you did to
reproduce the problem. Thanks.
What steps will reproduce the problem?
1. Run test/nilptr/arrayindex1.go on Darwin/Snow Leopard
2. Watch test spin forever.
3.
What is the expected output? What do you see instead?
Output is:
SIGBUS: bus error
Faulting address: 0x0
pc: 0x1c4b
main·main+0x4b /Users/dga/go/test/nilptr/arrayindex1.go:30
main·main()
mainstart+0xf /Users/dga/go/src/pkg/runtime/386/asm.s:81
mainstart()
goexit /Users/dga/go/src/pkg/runtime/proc.c:135
goexit()
eax 0x0
ebx 0x16814
ecx 0x20047000
edx 0x1
edi 0x9
esi 0x20047050
ebp 0x0
esp 0x20048fc4
eip 0x1c4b
eflags 0x10246
cs 0x17
fs 0x0
gs 0x3f
at this point, the process does *not* exit; it keeps spinning.
What is your $GOOS? $GOARCH?
GOARCH=386
GOOS=darwin
Which revision are you sync'ed to? (hg log -l 1)
changeset: 4148:b56be100e580
tag: tip
user: Rob Pike <r@golang.org>
date: Wed Nov 18 19:58:32 2009 -0800
summary: replace custom findByte with bytes.IndexByte
Please provide any additional information below.
This has been happening for a few hg syncs now.
Attaching to the spinning process with gdb shows:
0x000053af in ?? ()
(gdb) info reg
eax 0x1 1
ecx 0xa4d4 42196
edx 0x0 0
ebx 0x1f 31
esp 0x20055ab4 0x20055ab4
ebp 0x0 0x0
esi 0x60 96
edi 0x2 2
eip 0x53af 0x53af
eflags 0x10206 66054
cs 0x17 23
ss 0x1f 31
ds 0x1f 31
es 0x1f 31
fs 0x0 0
gs 0x1f 31
6nm suggests that the point where it's spinning is inside sigtramp:
53a3 T sigtramp
During the spinning, activity monitor shows the go process taking half the
CPU and kernel task 0 taking the other half. I'm not entirely sure how to
interpret that unless it's repeatedly segfaulting.