code review 98980043: race.bash,race.bat: unset GOROOT_FINAL during tests

race.bash,race.bat: unset GOROOT_FINAL during tests

Just like run.* scripts do.
Fixes race build.

[brainman, 2014-05-07 06:19:26]

Hello dvyukov@google.com (cc: golang-codereviews@googlegroups.com),

I'd like you to review this change to
https://go.googlecode.com/hg/

[brainman, 2014-05-07 06:20:12]

I tested on windows-amd64-race only. It PASSes TestAddr2Line, but fails with:

BenchmarkNewIndexRandom-4	       1	6200775000 ns/op
BenchmarkNewIndexRepeat-4	       1	18565218000 ns/op
BenchmarkSaveRestore-4	==7160==ERROR: Failed to allocate 0x18080000 (403177472)
bytes at 0x03026df00000 (1455)
runtime: newstack framesize=0x0 argsize=0x18 sp=0x22f158 stack=[0x3b8000,
0x3b9fa8]
	morebuf={pc:0x0 sp:0x0 lr:0x0}
	sched={pc:0x40fef0 sp:0x22f160 lr:0x0 ctxt:0x0}
runtime: split stack overflow: 0x22f158 < 0x3b8000
fatal error: runtime: split stack overflow

runtime stack:
runtime.throw(0x749ceb)
	c:/go/src/pkg/runtime/panic.c:520 +0x7e
runtime.newstack()
	c:/go/src/pkg/runtime/stack.c:712 +0x2e3
runtime.morestack()
	c:/go/src/pkg/runtime/asm_amd64.s:228 +0x64

goroutine 22 [stack growth]:
runtime.sighandler(0x22fe10, 0x22f920, 0xc082050000)
	c:/go/src/pkg/runtime/os_windows_amd64.c:41 fp=0x22f168
----- exception handler -----
runtime: unknown pc 0x22fb74 after exception handler
panic during panic
[signal 0xc0000005 code=0x0 addr=0x14 pc=0x42a989]

runtime stack:
runtime.startpanic()
	c:/go/src/pkg/runtime/panic.c:408 +0xfd
runtime.throw(0x747d00)
	c:/go/src/pkg/runtime/panic.c:517 +0x5a
runtime.sigpanic()
	c:/go/src/pkg/runtime/os_windows.c:350 +0x4b
runtime.funcspdelta(0x0, 0x22fb74)
	c:/go/src/pkg/runtime/symtab.goc:209 +0x29
runtime.gentraceback(0x40fef0, 0x22f160, 0x0, 0xc082050000, 0x0, 0x0, 0x64, 0x0,
0x0, 0x22fd00)
	c:/go/src/pkg/runtime/traceback_x86.c:154 +0x2ec
runtime.traceback(0xffffffffffffffff, 0xffffffffffffffff, 0x0, 0xc082050000)
	c:/go/src/pkg/runtime/traceback_x86.c:300 +0x9a
runtime.tracebackothers(0x786840)
	c:/go/src/pkg/runtime/proc.c:317 +0x1c5
runtime.dopanic(0x0)
	c:/go/src/pkg/runtime/panic.c:443 +0x1aa
runtime.throw(0x749ceb)
	c:/go/src/pkg/runtime/panic.c:519 +0x7e
runtime.newstack()
	c:/go/src/pkg/runtime/stack.c:712 +0x2e3
runtime.morestack()
	c:/go/src/pkg/runtime/asm_amd64.s:228 +0x64
exit status 2
FAIL	index/suffixarray	118.422s

but I think it is unrelated.

Alex

[dave_cheney.net, 2014-05-07 06:27:40]

LGTM. Please wait for Dmitry.

[dvyukov, 2014-05-07 06:29:27]

change make.* to run.* in change description

[dvyukov, 2014-05-07 06:30:01]

On 2014/05/07 06:20:12, brainman wrote:
> I tested on windows-amd64-race only. It PASSes TestAddr2Line, but fails with:
> 
> BenchmarkNewIndexRandom-4	       1	6200775000 ns/op
> BenchmarkNewIndexRepeat-4	       1	18565218000 ns/op
> BenchmarkSaveRestore-4	==7160==ERROR: Failed to allocate 0x18080000
(403177472)
> bytes at 0x03026df00000 (1455)
> runtime: newstack framesize=0x0 argsize=0x18 sp=0x22f158 stack=[0x3b8000,
> 0x3b9fa8]
> 	morebuf={pc:0x0 sp:0x0 lr:0x0}
> 	sched={pc:0x40fef0 sp:0x22f160 lr:0x0 ctxt:0x0}
> runtime: split stack overflow: 0x22f158 < 0x3b8000
> fatal error: runtime: split stack overflow
> 
> runtime stack:
> runtime.throw(0x749ceb)
> 	c:/go/src/pkg/runtime/panic.c:520 +0x7e
> runtime.newstack()
> 	c:/go/src/pkg/runtime/stack.c:712 +0x2e3
> runtime.morestack()
> 	c:/go/src/pkg/runtime/asm_amd64.s:228 +0x64
> 
> goroutine 22 [stack growth]:
> runtime.sighandler(0x22fe10, 0x22f920, 0xc082050000)
> 	c:/go/src/pkg/runtime/os_windows_amd64.c:41 fp=0x22f168
> ----- exception handler -----
> runtime: unknown pc 0x22fb74 after exception handler
> panic during panic
> [signal 0xc0000005 code=0x0 addr=0x14 pc=0x42a989]
> 
> runtime stack:
> runtime.startpanic()
> 	c:/go/src/pkg/runtime/panic.c:408 +0xfd
> runtime.throw(0x747d00)
> 	c:/go/src/pkg/runtime/panic.c:517 +0x5a
> runtime.sigpanic()
> 	c:/go/src/pkg/runtime/os_windows.c:350 +0x4b
> runtime.funcspdelta(0x0, 0x22fb74)
> 	c:/go/src/pkg/runtime/symtab.goc:209 +0x29
> runtime.gentraceback(0x40fef0, 0x22f160, 0x0, 0xc082050000, 0x0, 0x0, 0x64,
0x0,
> 0x0, 0x22fd00)
> 	c:/go/src/pkg/runtime/traceback_x86.c:154 +0x2ec
> runtime.traceback(0xffffffffffffffff, 0xffffffffffffffff, 0x0, 0xc082050000)
> 	c:/go/src/pkg/runtime/traceback_x86.c:300 +0x9a
> runtime.tracebackothers(0x786840)
> 	c:/go/src/pkg/runtime/proc.c:317 +0x1c5
> runtime.dopanic(0x0)
> 	c:/go/src/pkg/runtime/panic.c:443 +0x1aa
> runtime.throw(0x749ceb)
> 	c:/go/src/pkg/runtime/panic.c:519 +0x7e
> runtime.newstack()
> 	c:/go/src/pkg/runtime/stack.c:712 +0x2e3
> runtime.morestack()
> 	c:/go/src/pkg/runtime/asm_amd64.s:228 +0x64
> exit status 2
> FAIL	index/suffixarray	118.422s
> 
> but I think it is unrelated.

does it crash w/o this change?

[dvyukov, 2014-05-07 06:30:09]

LGTM

[brainman, 2014-05-07 06:34:30]

*** Submitted as https://code.google.com/p/go/source/detail?r=f96d3354edb6 ***

race.bash,race.bat: unset GOROOT_FINAL during tests

Just like run.* scripts do.
Fixes race build.

LGTM=dave, dvyukov
R=dvyukov, dave
CC=golang-codereviews
https://codereview.appspot.com/98980043

[brainman, 2014-05-07 06:35:26]

On 2014/05/07 06:30:01, dvyukov wrote:
> 
> does it crash w/o this change?

Testing it now ...

Alex

[brainman, 2014-05-07 07:15:13]

On 2014/05/07 06:35:26, brainman wrote:
> On 2014/05/07 06:30:01, dvyukov wrote:
> > 
> > does it crash w/o this change?
> 
> Testing it now ...
> 

Yes, same error message while testing d797b2316601.

Alex

[gobot, 2014-05-07 12:17:19]

This CL appears to have broken the freebsd-arm-cheney builder.
See http://build.golang.org/log/ea2e7bfe7b153483cea6811d0d77a28e93a18219