You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@daffodil.apache.org by Mike Beckerle <mb...@tresys.com> on 2018/02/20 23:34:57 UTC
can the failure be repeated
I can't get DAFFODIL-1908 to fail locally. I have some ideas to try, but I can only verify they don't break anything.
If I give you a git hash/branch can you set the tests to run against that?
Alternatively we have to find a way to drive it into this failure state somehow.
But I want to fix what appear to be some easy threading inconsistencies in code that is my best guess related to the issue.
RE: can the failure be repeated
Posted by Dave Thompson <dt...@tresys.com>.
Mike,
Yes I should be able to run the tests on the test machine with your branch, but unless we are trying to get this into v2.1.0 I'm going to wait until we are finished with v2.1.0.
The tests passed the last 5 times I ran them after failing twice so I will need to run the test numerous times for any kind of assurance.
Dave
From: Mike Beckerle
Sent: Tuesday, February 20, 2018 6:49 PM
To: Dave Thompson <dt...@tresys.com>; dev@daffodil.apache.org
Subject: Re: can the failure be repeated
I fixed two possible thread issues in the DaffodilXMLLoader.
Commit is b8229d16fe7c36fbe9c7695e34d0385535c9bb34
This is in my fork repo, not the primary repo. I.e., clone https://github.com/mbeckerle/incubator-daffodil.git
Branch is daffodil-1908-intermittent
Can we try this on the big test machine? I'm curious whether this makes a difference.
________________________________
From: Mike Beckerle <mb...@tresys.com>>
Sent: Tuesday, February 20, 2018 6:34:57 PM
To: Dave Thompson
Cc: dev@daffodil.apache.org<ma...@daffodil.apache.org>
Subject: can the failure be repeated
[This sender failed our fraud detection checks and may not be who they appear to be. Learn about spoofing at http://aka.ms/LearnAboutSpoofing]
I can't get DAFFODIL-1908 to fail locally. I have some ideas to try, but I can only verify they don't break anything.
If I give you a git hash/branch can you set the tests to run against that?
Alternatively we have to find a way to drive it into this failure state somehow.
But I want to fix what appear to be some easy threading inconsistencies in code that is my best guess related to the issue.
Re: can the failure be repeated
Posted by Mike Beckerle <mb...@tresys.com>.
I fixed two possible thread issues in the DaffodilXMLLoader.
Commit is b8229d16fe7c36fbe9c7695e34d0385535c9bb34
This is in my fork repo, not the primary repo. I.e., clone https://github.com/mbeckerle/incubator-daffodil.git
Branch is daffodil-1908-intermittent
Can we try this on the big test machine? I'm curious whether this makes a difference.
________________________________
From: Mike Beckerle <mb...@tresys.com>
Sent: Tuesday, February 20, 2018 6:34:57 PM
To: Dave Thompson
Cc: dev@daffodil.apache.org
Subject: can the failure be repeated
[This sender failed our fraud detection checks and may not be who they appear to be. Learn about spoofing at http://aka.ms/LearnAboutSpoofing]
I can't get DAFFODIL-1908 to fail locally. I have some ideas to try, but I can only verify they don't break anything.
If I give you a git hash/branch can you set the tests to run against that?
Alternatively we have to find a way to drive it into this failure state somehow.
But I want to fix what appear to be some easy threading inconsistencies in code that is my best guess related to the issue.