[Qgis-developer] Fwd: Threads in sextante

Etienne Tourigny etourigny.dev at gmail.com
Sun May 13 14:36:16 PDT 2012


Victor,

I agree on all your points.

May I suggest to have a unique interface/approach that other existing
interfaces can benefit from (i.e. standalone gdaltools and ftools).

It seems that QProcess is an interesting candidate (and QThread when
not using external app.), because it works in both c++ and in pyqt
(ref. gdaltools) - plus there is already code inside qgis that uses
it.

my 2 centavos
Etienne

On Sun, May 13, 2012 at 6:13 PM, Victor Olaya <volayaf at gmail.com> wrote:
> Here is my opinion on all that, in short:
>
> -The main idea of using threads is not now to execute multiplep
> algorithms at the same time, so we can remove a lot of complexity
> (like 2 processes writing the same file, as it was mentioned)
>
> -Implementing a canceling mechanism is going to be the greatest
> improvement once we have the threading stuff
>
> -So, for me, this is what should be done:
>
>  1)implementing the threading part
>  2) implement the cancelling mechanism (a responsive cancel button, etc)
>  3) modifying algorithm to respond to cancellin events.
>
> I think that would be more than enough for a first part, and it would
> be a great improvement for SEXTANTE
>
> Regards
>
> p.s. Just to let you know, I am leaving for Honduras tomorrow, and
> will be there working for 2 weeks, so I not might be very active on
> the lists during the next 15 days or so
>
> 2012/5/13 Marco Hugentobler <marco.hugentobler at sourcepole.ch>:
>> Hi Alex
>>
>>
>>>Also I think about rewritting fTools in C++ and integrating its
>>>algorithms into QGIS
>>>analysis library. This makes tools a bit faster and will allow to use
>>>algorithms in
>>>plugins
>>
>> That would be a great enhancement and leverage the analysis library as a
>> whole.
>>
>> Regards,
>> Marco
>>
>>
>>
>>
>> On 13.05.2012 09:23, Alexander Bruy wrote:
>>>
>>> Hi Etienne,
>>>
>>> 2012/5/12 Etienne Tourigny<etourigny.dev at gmail.com>:
>>>>
>>>> I stand corrected then - I guess I was confusing it with ftools, which
>>>> is blocking and does not spawn a process.
>>>> Any way ftools could be adapted to use the same approach as gdaltools?
>>>
>>> most fTools functions already implemented as threaded, but there are some
>>> minor issues. I work on improving processes handing in fTools, and several
>>> tools already don't block main QGIS window, has non-modal dialogs and
>>> allows
>>> to cancel running process.
>>>
>>> Also I think about rewritting fTools in C++ and integrating its
>>> algorithms into QGIS
>>> analysis library. This makes tools a bit faster and will allow to use
>>> algorithms in
>>> plugins
>>>
>>
>>
>> --
>> Dr. Marco Hugentobler
>> Sourcepole -  Linux&  Open Source Solutions
>> Churerstrasse 22, CH-8808 Pfäffikon SZ, Switzerland
>> marco.hugentobler at sourcepole.ch http://www.sourcepole.ch
>> Technical Advisor QGIS Project Steering Committee
>>
>>
>> _______________________________________________
>> Qgis-developer mailing list
>> Qgis-developer at lists.osgeo.org
>> http://lists.osgeo.org/mailman/listinfo/qgis-developer
> _______________________________________________
> Qgis-developer mailing list
> Qgis-developer at lists.osgeo.org
> http://lists.osgeo.org/mailman/listinfo/qgis-developer


More information about the Qgis-developer mailing list