[Qgis-psc] outstanding issues

[Werner Macho]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi @all

@Tim, I tried again to prepare a "better" testcase. The attributes are
completely irrelevant (so I deleted the area)

And I added the results from
Vector -> Geoprocessing Tools -> Intersect from QGIS 2.8, 2.10 and maste
r.

I hope the difference in the results is now clearly visible (We have a
intersection (which is correct) in 2.8 but nothing in 2.10 and master ..

So 2.8 shows the intersecting polygon while from 2.10 on it is
"silently" discarded (because the returned result is a geometry
collection (which is correct) but from 2.10 on it does not show the
polygon anymore.

I cannot tell exactly what else is affected by this - but I tested at
least union which also gives wrong results - so I imagine difference
and cut would probably have the same results.

Just load the testcase project in QGIS 2.8 - do an intersection
between g1 and g2 - and than load it in QGIS 2.12 and try the same again
.

Hope this one is better explained now

regards
Werner

On 20/11/15 06:06, Tim Sutton wrote:
> Hi
> 
> @werner can you please replace. Y little summary table,  removing 
> irrelevant items and adding relevant items?
> 
> I was hoping someone could clearly lay out what is broken, and in
> which versions.
> 
> Thanks!
> 
> Regards
> 
> Tim
> 
> On Nov 20, 2015 12:34 AM, "Werner Macho" <werner.macho at gmail.com 
> <mailto:werner.macho at gmail.com>> wrote:
> 
> Hi!
> 
> I tested 2.8.* 2.10 2.12 and master .. And with the attached
> testcase the correct result was only visible to the user with 2.8.*
> .. Hope the testproject will get through the list ..
> 
> Tested the intersect tool in QGIS(ftools) and processing. Same
> result. And I am pretty sure that this happens elsewhere too. So as
> Luigi said - the users are currently getting wrong results (or lets
> summarize it as "half the truth" :) ) presented. Since 2.8.3 at
> least the intersect tool is affected and QGIS should communicate
> that .. IMHO I would prefer a crash than getting results I cannot
> rely on.
> 
> Hope my testfiles are useful.
> 
> regards Werner
> 
> On 19/11/15 23:13, Luigi Pirelli wrote:
>> Hi Nyall
> 
>> as reported by Giovanni in the first thread mail and confirmed
>> by Werner during Hackmeeting... I don't know if they tested in
>> master (I suppose they did)
> 
>> Giovanni and Werner have a test data set to stress FTools in the 
>> new geometry engine context. I don't know if they added this
>> test data set in a issue. Thre focus of the thread, imho is 1)
>> what to do with the current distributed ftool (How awaring users
>> of problem) 2) plan a solution to fix it
> 
>> cheers Luigi Pirelli
> 
>> *********************************************************************
*
>
>> 
****************************
> 
> 
> * Boundless QGIS Support/Development: lpirelli AT boundlessgeo DOT
> com
>> * LinkedIn: https://www.linkedin.com/in/luigipirelli * 
>> Stackexchange: 
>> http://gis.stackexchange.com/users/19667/luigi-pirelli * GitHub: 
>> https://github.com/luipir * Mastering QGIS: 
>> https://www.packtpub.com/application-development/mastering-qgis 
>> *********************************************************************
*
>
>> 
****************************
> 
> 
> 
>> On 19 November 2015 at 23:04, Nyall Dawson
>> <nyall.dawson at gmail.com
> <mailto:nyall.dawson at gmail.com>>
>> wrote:
>>> On 20 November 2015 at 07:49, Tim Sutton <tim at kartoza.com
> <mailto:tim at kartoza.com>>
>>> wrote:
>>> 
>>>>> ftools OK              reshape/split tools ok
>>>> QGIS 2.8.3          Y*                                   N
>>>> QGIS 2.10           N                                    N
>>>> QGIS 2.12 N                                    N
>>> 
>>> I wasn't aware of this... can you summarise (or post a link
>>> to) the issues in the reshape/split tools? There was a lot of
>>> issues with this fixed in 2.12, and I'm not aware of any
>>> outstanding (reported) ones.
>>> 
>>>> 
>>>> 1) make small band-aid fixes to fix immediate issues 2) try
>>>> to sponsor e.g. Nyall or Jürgen to completely rewrite in C++
>>>> and have python bindings added so that they can be used from 
>>>> processing 3) try to sponsor e.g. Victor to implement as 
>>>> processing algs and then just wrap them as actions in the 
>>>> vector menu
>>>> 
>>> 
>>> I'd also like to point out that reimplementing these in c++ 
>>> would allow them to be reused by other (non-processing)
>>> plugins too. There's also likely a considerable speed boost
>>> possible from this approach.
>>> 
>>> Nyall _______________________________________________ Qgis-psc 
>>> mailing list Qgis-psc at lists.osgeo.org
> <mailto:Qgis-psc at lists.osgeo.org>
>>> http://lists.osgeo.org/mailman/listinfo/qgis-psc
> 
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2

iEYEARECAAYFAlZOtsQACgkQDAH1YiCxBglNogCghjyYGV0fy96+V6+q5akNgA7p
sOUAniyvxuIhq2BCDajDVYgiskM4XqXH
=pVdZ
-----END PGP SIGNATURE-----
-------------- next part --------------
A non-text attachment was scrubbed...
Name: testcase.zip
Type: application/x-zip-compressed
Size: 11941 bytes
Desc: not available
URL: <http://lists.osgeo.org/pipermail/qgis-psc/attachments/20151120/b7d591e5/attachment.bin>