[logback-dev] Nebula Grid migration

ceki ceki at qos.ch
Mon Jul 16 13:29:39 CEST 2012


Hi Christian,

In a branch called 'nebula', I've migrated logback-beagle code to use
a nebula Grid instead of a plain-old Table without adding any new
features. There have been a few regressions as a result.

1) slower performance

This should be mitigated by a slower update cycle but in batches. I'd
like to move to producer/consumer model where the grid is updated at
most every 500 ms in batches.

2) the alignment of the WARN ERROR flag is different

This will be fixed as we move to multiple columns.

3) It is no longer possible to select rows by scrolling outside the
Table/Grid window. However, it is possible to select rows by moving
the arrow keys on the keyword.

The next step is to display multiple columns with stack traces
spanning the entire row. I'll keep the list posted.

Cheers,

-- 
Ceki
http://tinyurl.com/proLogback

On 16.07.2012 10:18, Christian Trutz wrote:
> Hi Ceki,
>
> +1 for using multiple columns and for Nebula grid table (although I've
> been using it ever ...)
>
> Christian
>
> 2012/7/15 ceki <ceki at qos.ch>:
>>
>> Displaying each field in a separate columns is nice idea but I forgot
>> about stack trace lines.  Stack trace lines have to be displayed on
>> cells spanning multiple (if not all) columns. The regular SWT Table
>> does not offer this capability. Fortunately, the Nebula Grid widget
>> [1] does.  I could easily install it via its p2 repository.
>>
>> Cheers,
>>
>> [1] http://www.eclipse.org/nebula/widgets/grid/grid.php
>>
>>
>>
>> On 15.07.2012 16:09, ceki wrote:
>>>
>>> On 13.07.2012 22:53, Christian Trutz wrote:
>>>>
>>>> Hi together,
>>>>
>>>> here my comments on Logback Beagle Release 1.0.0:
>>>>
>>>> 4. Introduce LabelProvider and ContentProvider for virtual table
>>>> (BeagleView)
>>>> 5. ITableItemStub#getText() and #getImage() --> LabelProvider
>>>> 7. Remove ITableItemStubs (incl. subclasses) --> ContentProvider
>>>
>>>
>>> Currently, logback-beagle displays its output in table with a single
>>> column. The user experience is likely to be improved by displaying
>>> output on multiple columns, say, one column per event field. With
>>> multiple columns the user would be able to resize columns which could
>>> improve readability of the logs.
>>>
>>> WDYT?
>>>
>>>> Regards Christian
>>
>>
>>
>> --
>> Ceki
>> http://tinyurl.com/proLogback
>>
>>
>
>






More information about the logback-dev mailing list