[Qgis-developer] Discussion on browser and file extensions

[Etienne Tourigny]

On Fri, May 25, 2012 at 12:41 PM, Giuseppe Sucameli
<brush.tyler at gmail.com> wrote:
> Hi Etienne,
>
> On Fri, May 25, 2012 at 5:08 PM, Etienne Tourigny
> <etourigny.dev at gmail.com> wrote:
>>> BTW we need another method, right? Something like
>>> QgsLayerItem::layerName()
>>
>> This is the approach in my patches, only it is also supported in
>> QgsDataItem (see below why).
>
> IMHO it isn't.
>
> In your patch the fileName method is used to get the name
> displayed in the browser (see QgsBrowserModel::data() line 153),
>
> instead in my proposal the QgsLayerItem::layerName() would be used to get
> the name of the layer to load (not the filename).
>
> You should keep the QgsDataItem::name() for filename (data shown
> in the browser) and add a new method for QgsLayerItem::layerName().

ah ok, so name() would return the filename, whereas layerName() would
return the filename without extension, to be used in the drag&drop
operations. Sounds reasonable.
Would need to modify mime data stuff also, in QgsMimeDataUtils::Uri

what about file-based datacollection items? I guess only the layers
can get dropped to other targets, therefore no need to change
anything?

>
> BTW, I'm wondering if Radim meant the same.
>
>> Also, this could use uri() like you suggested instead of path(), not
>> sure I understand the distinction
>
> URI is the string used to load a layer, path is a way to identify a
> qgsdataitem in the browser tree.

ah ok, thanks

>
> Cheers.
>
>> QString QgsDataItem::fileName() const
>> {
>>  QFileInfo fileInfo( path() );
>>  return fileInfo.isFile() ? fileInfo.fileName() : name();
>> }
>>
>> Again, this works in the default case, but the function should be
>> overloaded for efficiency in the case of gdal ang ogr /vsizip files
>> (like I did in my patches)
>>
>> Regards,
>> Etienne
>>
>>>
>>>> QgsDataItem is base for various data sources, most of them are not
>>>> file based.
>>>
>>> I agree, having a fileName() method in the QgsDataItem class
>>> is non-sense.
>>
>> Well unless mistaken, you could have a QgsDataCollectionItem which is
>> file-based, such as multi-layer shape or qml?
>>
>>>
>>> Cheers.
>>>
>>> --
>>> Giuseppe Sucameli - Faunalia
>> _______________________________________________
>> Qgis-developer mailing list
>> Qgis-developer at lists.osgeo.org
>> http://lists.osgeo.org/mailman/listinfo/qgis-developer
>
>
>
> --
> Giuseppe Sucameli