Skip to main content
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [cross-project-issues-dev] Error Reporting Tool Improvement 
Maybe I worded mine badly, but I was after pretty much the same as Ed ;P

(maybe with the addition that I would take the cut-off trace as a base
for all subsequent analysis - blame, duplicate detection, ...)

On 06.05.2015 14:56, Marcel Bruch wrote:
> Thanks for the feedback Ed. I understand your critique with the current blame logic. Unfortunately cutting off / ignoring frames below Display.runAsync or Display.runDeferred can have the opposite effect (missing the project that scheduled the ui runnable). I remember stacktraces that contained the necessary information below these frames. I’ll have to investigate further to find out how often we will be wrong with one or the other approach.
> 
> Carsten’s feature request is different from your request. While I’m (ATM) not convinced cutting off frames from a trace wont remove necessary information (some times), I can imaging to produce reasonable results by improving the blame function as you suggested. I created bug 466576 [1] to track your request specifically. I’ll post updates to this bug.
> 
> Thanks,
> Marcel
> 
> [1] https://bugs.eclipse.org/bugs/show_bug.cgi?id=466576 <https://bugs.eclipse.org/bugs/show_bug.cgi?id=466576>
> 
> 
>> On 06 May 2015, at 14:00, Carsten Reckord <reckord@xxxxxxxx> wrote:
>>
>> For cutting off the stack frames at Display.readAndDispatch() there's
>> already https://bugs.eclipse.org/bugs/show_bug.cgi?id=451359.
>>
>> On 06.05.2015 13:54, Ed Merks wrote:
>>> Marcel,
>>>
>>> The error reporting tool is extremely cool and very useful!
>>>
>>> One aspect that could likely be improved significantly is the computation of the 
>>> "blamed" projects.  In particular I'm having to triage many reports such as this 
>>> one:
>>>
>>> https://dev.eclipse.org/recommenders/committers/confess/#/problems/54f9b544e4b0a38aecd742e2/details
>>>
>>> No doubt that's because Oomph is on the stack:
>>>
>>>
>>>
>>> I would like to suggest that the "blame" logic should consider that 
>>> Display.readAndDispatch processes events of arbitrary origin.  As such, nothing 
>>> below that point in the stack is likely contributing to any problem above that 
>>> point in the stack.
>>>
>>> Given that Oomph often starts a "background" dialog to perform tasks on startup, 
>>> every possible problem that happens elsewhere in the IDE ends up in Oomph's 
>>> triage bucket,  so the list I must triage is large to the point where I'm not 
>>> sure it can ever be managed:
>>>
>>> https://dev.eclipse.org/recommenders/committers/confess/#/problems/?projects=oomph&kinds=NORMAL&kinds=FREEZE&categories=UNCONFIRMED&page=1&size=100&sort=createdOn,desc
>>>
>>> The duplicates are a bit annoying too, but you have some nice support for that 
>>> already:
>>>
>>> https://dev.eclipse.org/recommenders/committers/confess/#/problems/5522970ce4b026254ee014bf/similar
>>>
>>> Kudos for creating this great technology!
>>>
>>> Cheers,
>>> Ed
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>> _______________________________________________
>>> cross-project-issues-dev mailing list
>>> cross-project-issues-dev@xxxxxxxxxxx
>>> To change your delivery options, retrieve your password, or unsubscribe from this list, visit
>>> https://dev.eclipse.org/mailman/listinfo/cross-project-issues-dev
>>>
>>
>> _______________________________________________
>> cross-project-issues-dev mailing list
>> cross-project-issues-dev@xxxxxxxxxxx
>> To change your delivery options, retrieve your password, or unsubscribe from this list, visit
>> https://dev.eclipse.org/mailman/listinfo/cross-project-issues-dev
> 
> 
> 
> _______________________________________________
> cross-project-issues-dev mailing list
> cross-project-issues-dev@xxxxxxxxxxx
> To change your delivery options, retrieve your password, or unsubscribe from this list, visit
> https://dev.eclipse.org/mailman/listinfo/cross-project-issues-dev
> 

-- 
Yatta Solutions GmbH
- Carsten Reckord -

  t  +49 (0)69 2475666-33
  f  +49 (0)69 2475668-0
  e  reckord@xxxxxxxx

Anschrift Office Kassel
  Universitätsplatz 12
  34127 Kassel

Anschrift Office Frankfurt a.M.
  Mainzer Landstraße 50
  60325 Frankfurt a.M.

Sitz, Handelsregister:
  Sitz der Gesellschaft: Kassel
  Amtsgericht Kassel, HRB 14720
  USt-IdNr DE263191529

Geschäftsführung:
  Johannes Jacop
  Dr. Christian Schneider

Kontakt Geschäftsstelle:
  t  +49 (0)69 2475666-0
  f  +49 (0)69 2475668-0
  e  info@xxxxxxxx

Back to the top