http://build.squeak.org/job/SqueakTrunk/574/console shows a typical failure in CI at the moment. The relevant output is:
2013-10-30T22:17:00.132+01:00: Running tests for Tests... !!! Killed command 0 for exceeding allotted time: nice /home/jenkins/workspace/SqueakTrunk/target/cog.r2776/coglinux/bin/squeak -vm-sound-null -vm-display-null "/home/jenkins/workspace/SqueakTrunk/target/PostTestTrunkImage.image" ../tests.st. /home/jenkins/workspace/SqueakTrunk/benchmarks.st spawning command 1 with timeout 600 seconds: nice /home/jenkins/workspace/SqueakTrunk/target/cog.r2776/coglinux/bin/squeak -vm-sound-null -vm-display-null "/home/jenkins/workspace/SqueakTrunk/target/PostTestTrunkImage.image" ../benchmarks.st
2013-10-30T23:01:41.404+01:00: Running tests for ToolBuilder-Kernel...
Note that the "!!!" line happens because the running of the full test suite took OVER 40 MINUTES. My crummy little laptop chewed through the test suite in 10 minutes.
So: on Windows 8 I cannot reproduce the problem. Would someone with a Linux machine please try run the full suite of tests? I strongly suspect we have a dialog popping up, but only on Linux.
(I'm busy working through Pavel's DummyUIManager. I suspect that I'll be copying most of it into my CommandLineUIManager, for submission to trunk. Hopefully I'll get to it in the very near future.)
frank
On 30 October 2013 23:15, Frank Shearar frank.shearar@gmail.com wrote:
http://build.squeak.org/job/SqueakTrunk/574/console shows a typical failure in CI at the moment. The relevant output is:
2013-10-30T22:17:00.132+01:00: Running tests for Tests... !!! Killed command 0 for exceeding allotted time: nice /home/jenkins/workspace/SqueakTrunk/target/cog.r2776/coglinux/bin/squeak -vm-sound-null -vm-display-null "/home/jenkins/workspace/SqueakTrunk/target/PostTestTrunkImage.image" ../tests.st. /home/jenkins/workspace/SqueakTrunk/benchmarks.st spawning command 1 with timeout 600 seconds: nice /home/jenkins/workspace/SqueakTrunk/target/cog.r2776/coglinux/bin/squeak -vm-sound-null -vm-display-null "/home/jenkins/workspace/SqueakTrunk/target/PostTestTrunkImage.image" ../benchmarks.st
2013-10-30T23:01:41.404+01:00: Running tests for ToolBuilder-Kernel...
Note that the "!!!" line happens because the running of the full test suite took OVER 40 MINUTES. My crummy little laptop chewed through the test suite in 10 minutes.
As an aside, this issue causes massive test failures. You can see 62 failures in that build when we should only have something like 5. Note http://build.squeak.org/job/SqueakTrunk/574/testReport/SystemChangeNotificat... which shows a neat symptom of the problem: "RemoteString past end of file" indicates that your changes file has been altered underneath your feet. In the CI, this happens because the changes file has been _deleted_ but the tests are still running.
frank
On 30 October 2013 23:15, Frank Shearar frank.shearar@gmail.com wrote:
http://build.squeak.org/job/SqueakTrunk/574/console shows a typical failure in CI at the moment. The relevant output is:
2013-10-30T22:17:00.132+01:00: Running tests for Tests... !!! Killed command 0 for exceeding allotted time: nice /home/jenkins/workspace/SqueakTrunk/target/cog.r2776/coglinux/bin/squeak -vm-sound-null -vm-display-null "/home/jenkins/workspace/SqueakTrunk/target/PostTestTrunkImage.image" ../tests.st. /home/jenkins/workspace/SqueakTrunk/benchmarks.st spawning command 1 with timeout 600 seconds: nice /home/jenkins/workspace/SqueakTrunk/target/cog.r2776/coglinux/bin/squeak -vm-sound-null -vm-display-null "/home/jenkins/workspace/SqueakTrunk/target/PostTestTrunkImage.image" ../benchmarks.st
2013-10-30T23:01:41.404+01:00: Running tests for ToolBuilder-Kernel...
Note that the "!!!" line happens because the running of the full test suite took OVER 40 MINUTES. My crummy little laptop chewed through the test suite in 10 minutes.
So: on Windows 8 I cannot reproduce the problem. Would someone with a Linux machine please try run the full suite of tests? I strongly suspect we have a dialog popping up, but only on Linux.
I can reproduce the issue with a Cog on Ubuntu. There's no modal dialog. If I cmd-., I can see that MCMethodDefinitionTest is running. It looks like it's looping forever in MCMockPackageInfo >> #changeRecordsForMethod:do:'s
[position notNil & file notNil] whileTrue: [file position: (0 max: position-150). "Skip back to before the preamble" [file position < (position-1)] "then pick it up from the front" whileTrue: [chunk := file nextChunk].
Thoughts?
frank
FWIW, I just tried to run MCMethodDefinitionTest in an updated image and it popped up 3 requests for my initials and failed 3 tests.
Cheers, Bob
On 10/31/13 5:20 AM, Frank Shearar wrote:
On 30 October 2013 23:15, Frank Shearar frank.shearar@gmail.com wrote:
http://build.squeak.org/job/SqueakTrunk/574/console shows a typical failure in CI at the moment. The relevant output is:
2013-10-30T22:17:00.132+01:00: Running tests for Tests... !!! Killed command 0 for exceeding allotted time: nice /home/jenkins/workspace/SqueakTrunk/target/cog.r2776/coglinux/bin/squeak -vm-sound-null -vm-display-null "/home/jenkins/workspace/SqueakTrunk/target/PostTestTrunkImage.image" ../tests.st. /home/jenkins/workspace/SqueakTrunk/benchmarks.st spawning command 1 with timeout 600 seconds: nice /home/jenkins/workspace/SqueakTrunk/target/cog.r2776/coglinux/bin/squeak -vm-sound-null -vm-display-null "/home/jenkins/workspace/SqueakTrunk/target/PostTestTrunkImage.image" ../benchmarks.st
2013-10-30T23:01:41.404+01:00: Running tests for ToolBuilder-Kernel...
Note that the "!!!" line happens because the running of the full test suite took OVER 40 MINUTES. My crummy little laptop chewed through the test suite in 10 minutes.
So: on Windows 8 I cannot reproduce the problem. Would someone with a Linux machine please try run the full suite of tests? I strongly suspect we have a dialog popping up, but only on Linux.
I can reproduce the issue with a Cog on Ubuntu. There's no modal dialog. If I cmd-., I can see that MCMethodDefinitionTest is running. It looks like it's looping forever in MCMockPackageInfo >> #changeRecordsForMethod:do:'s
[position notNil & file notNil] whileTrue: [file position: (0 max: position-150). "Skip back to before
the preamble" [file position < (position-1)] "then pick it up from the front" whileTrue: [chunk := file nextChunk].
Thoughts?
frank
I can run the Monticello tests through the TestRunner, where I have three failures:
MCEnvironmentLoadTest >> #testLoadIntoEnvironment fails, as it should (because noone's implemented the backing behaviour yet).
MCWorkingCopyTest >> #testInfoProxy sometimes passes, and sometimes fails.
Lastly, MCClassDefinitionTest >> #testLoadAndUnload fails with "RemoteString past end of file"
Bob, is that what you see?
frank
On 31 October 2013 10:09, Bob Arning arning315@comcast.net wrote:
FWIW, I just tried to run MCMethodDefinitionTest in an updated image and it popped up 3 requests for my initials and failed 3 tests.
Cheers, Bob
On 10/31/13 5:20 AM, Frank Shearar wrote:
On 30 October 2013 23:15, Frank Shearar frank.shearar@gmail.com wrote:
http://build.squeak.org/job/SqueakTrunk/574/console shows a typical failure in CI at the moment. The relevant output is:
2013-10-30T22:17:00.132+01:00: Running tests for Tests... !!! Killed command 0 for exceeding allotted time: nice /home/jenkins/workspace/SqueakTrunk/target/cog.r2776/coglinux/bin/squeak -vm-sound-null -vm-display-null "/home/jenkins/workspace/SqueakTrunk/target/PostTestTrunkImage.image" ../tests.st. /home/jenkins/workspace/SqueakTrunk/benchmarks.st spawning command 1 with timeout 600 seconds: nice /home/jenkins/workspace/SqueakTrunk/target/cog.r2776/coglinux/bin/squeak -vm-sound-null -vm-display-null "/home/jenkins/workspace/SqueakTrunk/target/PostTestTrunkImage.image" ../benchmarks.st
2013-10-30T23:01:41.404+01:00: Running tests for ToolBuilder-Kernel...
Note that the "!!!" line happens because the running of the full test suite took OVER 40 MINUTES. My crummy little laptop chewed through the test suite in 10 minutes.
So: on Windows 8 I cannot reproduce the problem. Would someone with a Linux machine please try run the full suite of tests? I strongly suspect we have a dialog popping up, but only on Linux.
I can reproduce the issue with a Cog on Ubuntu. There's no modal dialog. If I cmd-., I can see that MCMethodDefinitionTest is running. It looks like it's looping forever in MCMockPackageInfo >> #changeRecordsForMethod:do:'s
[position notNil & file notNil] whileTrue: [file position: (0 max: position-150). "Skip back to before
the preamble" [file position < (position-1)] "then pick it up from the front" whileTrue: [chunk := file nextChunk].
Thoughts?
frank
I get something else. First I get 3 requests for my initials and the following tests show up as *failures*:
'MCMethodDefinitionTest>>#testPartiallyRevertOverrideMethod' 'MCMethodDefinitionTest>>#testRevertOldMethod' 'MCMethodDefinitionTest>>#testRevertOverrideMethod'
If I enter my initials and rerun the tests, I get 1 *error*:
MessageNotUnderstood: MCMockClassA>>one 31 October 2013 8:41:26.694 am
VM: Mac OS - Smalltalk Image: Squeak4.5 [latest update: #13156]
SecurityManager state: Restricted: false FileAccess: true SocketAccess: true Working Dir /Users/bob/squeak/SECOG Trusted Dir /foobar/tooBar/forSqueak/bogus/ Untrusted Dir /Users/bob/Library/Preferences/Squeak/Internet/Untrusted
MCMockClassA(Object)>>doesNotUnderstand: #one Receiver: a MCMockClassA Arguments and temporary variables: aMessage: one exception: MessageNotUnderstood: MCMockClassA>>one resumeValue: nil Receiver's instance variables: ivar: nil
MCMethodDefinitionTest>>testLoadAndUnload Receiver: MCMethodDefinitionTest>>#testLoadAndUnload Arguments and temporary variables: definition: a MCMethodDefinition(MCMockClassA>>one) Receiver's instance variables: testSelector: #testLoadAndUnload timeout: nil navigation: a SystemNavigation isModified: false overrideTimestamp: 'eem 10/9/2010 17:15'
Cheers, Bob
On 10/31/13 8:12 AM, Frank Shearar wrote:
I can run the Monticello tests through the TestRunner, where I have three failures:
MCEnvironmentLoadTest >> #testLoadIntoEnvironment fails, as it should (because noone's implemented the backing behaviour yet).
MCWorkingCopyTest >> #testInfoProxy sometimes passes, and sometimes fails.
Lastly, MCClassDefinitionTest >> #testLoadAndUnload fails with "RemoteString past end of file"
Bob, is that what you see?
frank
On 31 October 2013 10:09, Bob Arning arning315@comcast.net wrote:
FWIW, I just tried to run MCMethodDefinitionTest in an updated image and it popped up 3 requests for my initials and failed 3 tests.
Cheers, Bob
On 10/31/13 5:20 AM, Frank Shearar wrote:
On 30 October 2013 23:15, Frank Shearar frank.shearar@gmail.com wrote:
http://build.squeak.org/job/SqueakTrunk/574/console shows a typical failure in CI at the moment. The relevant output is:
2013-10-30T22:17:00.132+01:00: Running tests for Tests... !!! Killed command 0 for exceeding allotted time: nice /home/jenkins/workspace/SqueakTrunk/target/cog.r2776/coglinux/bin/squeak -vm-sound-null -vm-display-null "/home/jenkins/workspace/SqueakTrunk/target/PostTestTrunkImage.image" ../tests.st. /home/jenkins/workspace/SqueakTrunk/benchmarks.st spawning command 1 with timeout 600 seconds: nice /home/jenkins/workspace/SqueakTrunk/target/cog.r2776/coglinux/bin/squeak -vm-sound-null -vm-display-null "/home/jenkins/workspace/SqueakTrunk/target/PostTestTrunkImage.image" ../benchmarks.st
2013-10-30T23:01:41.404+01:00: Running tests for ToolBuilder-Kernel...
Note that the "!!!" line happens because the running of the full test suite took OVER 40 MINUTES. My crummy little laptop chewed through the test suite in 10 minutes.
So: on Windows 8 I cannot reproduce the problem. Would someone with a Linux machine please try run the full suite of tests? I strongly suspect we have a dialog popping up, but only on Linux.
I can reproduce the issue with a Cog on Ubuntu. There's no modal dialog. If I cmd-., I can see that MCMethodDefinitionTest is running. It looks like it's looping forever in MCMockPackageInfo >> #changeRecordsForMethod:do:'s
[position notNil & file notNil] whileTrue: [file position: (0 max: position-150). "Skip back to before
the preamble" [file position < (position-1)] "then pick it up from the front" whileTrue: [chunk := file nextChunk].
Thoughts?
frank
squeak-dev@lists.squeakfoundation.org