Bug 3690 - testClone(frysk.util.TestFStack)j
Summary: testClone(frysk.util.TestFStack)j
Status: RESOLVED FIXED
Alias: None
Product: frysk
Classification: Unclassified
Component: general (show other bugs)
Version: unspecified
: P2 normal
Target Milestone: ---
Assignee: Nurdin Premji
URL:
Keywords:
Depends on: 3702
Blocks: 2244
  Show dependency treegraph
 
Reported: 2006-12-11 03:54 UTC by Andrew Cagney
Modified: 2006-12-20 21:09 UTC (History)
2 users (show)

See Also:
Host:
Target:
Build:
Last reconfirmed:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Andrew Cagney 2006-12-11 03:54:36 UTC
8) testClone(frysk.util.TestFStack)junit.framework.AssertionFailedError: Task #23053
#0 0x3a3402 in __kernel_vsyscall ()
#1 0xca2966 in nanosleep ()
#2 0xca278b in sleep ()
#3 0x8048d5d in ../../frysk/frysk-core/frysk/pkglibdir/funit-threads.c main ():
line #177
#4 0xc2d4e4 in __libc_start_main ()
#5 0x8048801 in _start ()
did not match: (Task #\d+\n#0 0x[\da-f]+ in __kernel_vsyscall \(\)\n#1
0x[\da-f]+ in __nanosleep_nocancel \(\)\n#2 0x[\da-f]+ in sleep 

note the nanosleep vs __nanosleep_nocancel.

For FStack it might be better to focus on the UI interface, and the broad format
of the output, and not be too worried about the exacting details of the stack
backtrace.  Assume here that libunwind and its bindings have unit tests covering
that.

Of course, if those lack coverage then the thing to do is expand those tests at
that level.
Comment 1 Andrew Cagney 2006-12-11 04:05:00 UTC
2006-12-10  Andrew Cagney  <cagney@redhat.com>

	* TestFStack.java (testCloen): Note broken - 3690.

Comment 2 Nurdin Premji 2006-12-11 21:38:54 UTC
Fixed

2006-12-11
frysk-core/frysk/util/CL
	* TestFStack.java: Stripped down regex's to just the common text between threads.
	(testClone): Re-enabled
	(testStressMultiThreadedDetach): Ditto.
Comment 3 Jan Kratochvil 2006-12-11 21:52:11 UTC
Please provide the `Value' and `Size' fields of `readelf -a libc.so.6' of the
symbols involved, therefore `nanosleep' and `__nanosleep_nocancel'. They look as
overlapping while I was improving the related libunwind resolving functionality
it would be good to know if there really isn't some bug.
Comment 4 Andrew Cagney 2006-12-11 22:04:02 UTC
Two bugs here; TestJUnit was to pedantic; a potential problem with libunwind.

Nurdin, can you create a new libunwind bug, blocking this, and fill in as much
information as possible.
Comment 5 Nurdin Premji 2006-12-20 21:09:08 UTC
Closing, 3702 deals with the libunwind problem.