[Qgis-developer] Re: [work] [Qgis-community-team] Re: Python plugins and C++ plugins: using different terms

Tim Sutton tim at linfiniti.com
Mon Sep 1 14:24:25 EDT 2008


Hi

> Monday 01 of September 2008 14:56:02 Mateusz Loskot napisał(a):
>> Just my 5 low-level eurocents to this issue.
>> I think that plugins developers writing in C++ language should provide
>>
>> 1. At least two compile-time tests (preprocessor directive #ifdef):

>>
>> a) Supported platform - on what platforms a plugin can be built.
>> IOW, a plugin is neither built nor installed on unsupported platforms.

QGIS core plugins should run on all platforms. 3rd party plugins run
on platforms defined by their developer and the development of those
plugins will happen out of QGIS source tree so no need for use to
worry about compile time tests I think.



>>
>> b) Minimal QGIS version required to use a plugin

Agreed we should have this. Even if the plugin loads on older versions
subtle changes / bug fixes may prevent it producing results as
expected.


>>
>> 2. One run-time check to confirm minimal required version of QGIS.
>> IMO, it would be nice to add a small reflection to plugin base class -
>> QgsPlugin:
>
> Hi Mateusz,
> Your improvements are helpful, of course, but they aren't sufficient to
> implement the automatic installation. The main problem remains: some plugins
> stop working after QGIS upgrade. It's the reason for my opinion, that
> simplifying of the installation without providing stable and up-to-date
> plugin repositories may cause many troubles in the future and isn't good for
> QGIS reputation ;)


This should be a thing of the past with QGIS and its stable api.QGIS
should be backwards compatible for all 1.x plugins but newer plugins
may specify requiring e.g. QGIS 1.1 if it has had things added to its
API.

Not providing a stable API till this point, allthough inconvenient for
plugin writers, was a conscious decision on our part so that we did
not need to be held accountable for an API that was not sustainable
for a reasonable length of time. In the long run I believe this was a
good decision.


Regards

Tim


>
> We have a lot of work on another features, so for me the best choice is to
> polish "python plugin distribution system" ;) and tend to include more
> important of C++ plugins to the QGIS package. For other ones, authors can
> provide installers (also with possibility to determine QGIS version)
>
> Regards!
> B.
> _______________________________________________
> Qgis-developer mailing list
> Qgis-developer at lists.osgeo.org
> http://lists.osgeo.org/mailman/listinfo/qgis-developer
>



-- 
Tim Sutton - QGIS Project Steering Committee Member (Release Manager)
==============================================
Visit http://linfiniti.com to find out about:
 * QGIS programming services
 * Mapserver and PostGIS based hosting plans
 * FOSS Consulting Services
Skype: timlinux Irc: timlinux on #qgis at freenode.net
==============================================


More information about the Qgis-developer mailing list