[Qgis-user] (no subject)

[Alister Hood]

> Date: Tue, 13 Mar 2012 16:40:11 +0100
> From: Paolo Cavallini <cavallini at faunalia.it>
> Subject: Re: [Qgis-user] (no subject)
> To: qgis-user at lists.osgeo.org
> Message-ID: <4F5F6A5B.60100 at faunalia.it>
> Content-Type: text/plain; charset=ISO-8859-1
> 
> Il 12/03/2012 22:23, Alister Hood ha scritto:
> 
> > 1. It is common for QGIS plugins to have dependencies which are not
> > readily available for Windows (you may need to put in the effort to
> build them yourself).
> 
> Not so common: can you provide examples?

Sure, I guess it all depends on what you consider to be common.  I certainly haven't seen this problem with _most_ plugins, but it is more than just one or two.

Recent examples I can think of are the Contour plugin (I think its dependencies are available now, after the switch to Python 2.7 in OSGeo4W), and the plugin(s?) that require Rpy or whatever it is that isn't available since the switch to Python 2.7.  And for a long time before that on Windows some of the plugins that used R (I'm not sure if they're the same ones which can't be used now) could only be used with quite an old version of R.
I remember several other examples, but the details don't spring to mind.

There's also the point that on Windows if a plugin has an unusual dependency you have to actually go looking for a Windows package, rather than just installing it from your distro's package manager.  And I've tried a number of plugins with python dependencies that you have to install by manually extracting files into the right place from a Windows installer package.  This isn't particularly user friendly ;)

> > 2. Plugins sometimes have Windows related problems.  I guess some of
> > the developers don't do much testing for Windows.
> 
> I do not think this is a widespread phenomenon.

It might not be.  But if you think people don't generally report bugs on windows, who knows?

> > 4. There seems to be a reasonable history of issues with QGIS that
> > only appear on Windows, particularly problems with printing.  Again, I
> > think this is due to the developers not using windows.
> 
> I would say many Win users do not report and follow bugs, and they are not
> used helping (by coding or by sponsoring) to fix them.

Yes, I guess that's the other contributing factor.

> > 5. FWIW I get a reasonable number of crashes when I try to run QGIS on
> > 64 bit Windows 7, and none on 32bit Windows XP.  I do always run a
> > self-compiled (on XP) master version of QGIS, not an official build,
> > but I don't really see why that would affect running on 64bit Windows
> > 7.  As you have seen though, some people report no trouble on Windows 7.
> 
> I think these are local (compilation) problems. I use regularly qgis on a
> variety of win machines for courses, and I do not encounter major
> problems.

Actually, in the past when I used the OSGeo4W packages I had even more crashes on that Windows 7 machine, so I'm wondering if there's something wrong with the machine...