[Qgis-developer] Re: python unicode problem in raster metadata

[Martin Dobias]

On Fri, Jan 30, 2009 at 9:47 AM, Volkan Kepoglu <vkepoglu at gmail.com> wrote:
> On 1/29/09, Martin Dobias <wonder.sk at gmail.com> wrote:
>
>> I've hopefully fixed this in r10038 (and without regressions). There
>> were some mismatches in the qgis python support regarding unicode.
>> Please check whether it helped.
>>
>> Martin
>>
>
> Hi, Martin,
>
> How can I check?

Just open the python console in QGIS and give it a try :-)


> After svn checkout, is it needed to make a full recompile of qgis?
> Even for every small change, is it necessary to recompile the qgis? If
> it is necessary, does ccmake understand what has changed and compiles
> only changed part?

Usually you need to do just "make" and "make install" in the
directories where something has been changed. However if there were
changes e.g. in core or gui libraries then you should usually
recompile also other parts of qgis and python bindings to make sure
that nothing gets wrong.

In case you don't know what exactly you're doing, I recommend you to
build the whole source tree.

> Actually I do not want to build the qgis from source code? Is there
> any other way to check for especially when only the python code has
> changed? Does the change of python (or sip part) affect the whole c++
> code of qgis?

Well, you can always wait for a new release, but that may take a while...
In QGIS there are two types of "python" code - one is in python/
subdir which contains bindings to core/gui libraries, the other one is
src/python/ subdir which is a support library for QGIS (i.e. python
console, python plugins). Python support library for QGIS depends on
python bindings to some degree.

Martin