Re: [Qgis-developer] Testing infrastructure (was: QGIS 2 64bits, is it stable ?)

2013-10-04 Thread Hien TRAN-QUANG
To get back to the first question (stability of QGIS 2 64bits/ Win),
my feelings are that a lot of bugs had been corrected before the
release announcement, but package phase did not go to the same level
of control and showed many regression. For instance, French IGN
projections bug in master was a GRASS one, and that bug came back with
the released version (but that does'nt seem to be Grass's fault this
time). Old or removed packages had been inserted in the installer in
the first version. And so on.

My self-compiled-from-release QGIS brick  is strong. Other bricks are
separately strong too. But the wall made of those bricks isn't.

Hien TRAN-QUANG



 -- Message transféré --
 From: Sandro Santilli s...@keybit.net
 To: qgis-developer@lists.osgeo.org
 Cc:
 Date: Fri, 4 Oct 2013 17:50:05 +0200
 Subject: [Qgis-developer] Testing infrastructure (was: QGIS 2 64bits, is it 
 stable ?)
 On Fri, Oct 04, 2013 at 04:19:38PM +0200, Paolo Cavallini wrote:
 -BEGIN PGP SIGNED MESSAGE-
 Hash: SHA1

 Il 04/10/2013 14:33, Régis Haubourg ha scritto:

  What about starting the infrastructure for unit test and commit validation
  and testing it now?

 AFAIK the infrastructure is already in place and running.

 No reference ?

 Any documentation for developers about how to create and run tests ?

 All I could find on the website is this interesting page:
 http://hub.qgis.org/projects/quantum-gis/wiki/Developers_Manual#Unit-Testing

 As of November 2007 we require all new features going into trunk to be
  accompanied with a unit test

 Is the document still valid ?

 Any page reporting test run status ?

 I think these are all part of an infrastructure (among other aspects)

  I must say I could be very interested in a commercial support for this, if
  this is sticking to community version and not a separate branch.

 If you set up a list of requirements, I'm sure there will be people willing 
 to
 undertake the task.

 Yes please !

 But probably also coming up with a list of requirements would be part of
 such project, as it will require finding an agreement among developers to
 cope with policies that will need to be defined (I guess).

 What about funding the new QA manager for working on such a plan ?

 --strk;

 PS: make check output against * master 6d5cf6f :
  82% tests passed, 14 tests failed out of 78
  Total Test time (real) = 471.88 sec
  The following tests FAILED:
   4 - qgis_diagramtest (Failed)
  14 - qgis_rendererstest (Failed)
  25 - qgis_composermaptest (Failed)
  30 - qgis_stylev2test (Failed)
  31 - qgis_composerhtmltest (Failed)
  33 - qgis_composerscalebartest (OTHER_FAULT)
  45 - qgis_wcsprovidertest (Failed)
  51 - PyQgsVectorLayer (Failed)
  52 - PyQgsRasterLayer (Failed)
  56 - PyQgsDelimitedTextProvider (Failed)
  64 - PyQgsComposerMap (Failed)
  72 - PyQgsPalLabelingBase (Failed)
  73 - PyQgsPalLabelingCanvas (Failed)
  78 - PyQgsAppStartup (Failed)





 -- Message transféré --
 From: Jürgen E. Fischer j...@norbit.de
 To: qgis-developer@lists.osgeo.org
 Cc:
 Date: Fri, 4 Oct 2013 20:03:44 +0200
 Subject: Re: [Qgis-developer] Testing infrastructure (was: QGIS 2 64bits, is 
 it stable ?)
 Hi Sandro,

 On Fri, 04. Oct 2013 at 17:50:05 +0200, Sandro Santilli wrote:
 On Fri, Oct 04, 2013 at 04:19:38PM +0200, Paolo Cavallini wrote:
  AFAIK the infrastructure is already in place and running.

 No reference ?

 The nightly builds run tests.  And AFAIK 
 http://jenkins.linfiniti.com/job/QGIS/
 builds QGIS and runs tests after each commit.

 Any page reporting test run status?

 In jenkins and for the nightlies on:

 http://dash.orfeo-toolbox.org/index.php?project=QGIS

 We might move that to our new server at some point.

 All I could find on the website is this interesting page:
 http://hub.qgis.org/projects/quantum-gis/wiki/Developers_Manual#Unit-Testing

 As of November 2007 we require all new features going into trunk to be
  accompanied with a unit test

 Is the document still valid ?

 Never heard of that - or successfully repressed it.

 Any page reporting test run status ?

 s.a.

 PS: make check output against * master 6d5cf6f :
  82% tests passed, 14 tests failed out of 78
  Total Test time (real) = 471.88 sec
  The following tests FAILED:
   4 - qgis_diagramtest (Failed)
  14 - qgis_rendererstest (Failed)
  25 - qgis_composermaptest (Failed)
  30 - qgis_stylev2test (Failed)
  31 - qgis_composerhtmltest (Failed)
  33 - qgis_composerscalebartest (OTHER_FAULT)
  45 - qgis_wcsprovidertest (Failed)
  51 - PyQgsVectorLayer (Failed)
  52 - PyQgsRasterLayer (Failed)
  56 - PyQgsDelimitedTextProvider (Failed)
  64 - PyQgsComposerMap (Failed)
  72 - PyQgsPalLabelingBase (Failed)
  73 - PyQgsPalLabelingCanvas (Failed)
  78 - PyQgsAppStartup

Re: [Qgis-developer] Bug 7941 : workaround ?

2013-08-02 Thread Hien TRAN-QUANG
Hello,

Can you check that you have a file named ntf_r93.gsb in the
OSgeo4W/share/proj folder ? That file is a datum shift grid provided by
french IGN and is called by the  +nadgrids=ntf_r93.gsb parameter in the
proj.4 call.


Hien.



 -- Message transféré --
 From: Benoît Laurent blaur...@heurisis.eu
 To: QGIS Developer List qgis-developer@lists.osgeo.org
 Cc:
 Date: Fri, 02 Aug 2013 14:47:20 +0200
 Subject: [Qgis-developer] Bug 7941 : workaround ?
 Hello,

 I am facing bug 7941 
 (http://hub.qgis.org/issues/**7941http://hub.qgis.org/issues/7941)
 which has not been corrected yet.

 I have many shapefiles in Lambert 93 and a raster file (.tif) in Lambert I
 which I would fit together. Because of the bug, I can't. Is there a
 workaround ? I tried to change the coordinates in the .tfw file but I does
 not work properly.

 Any help would be appreciated.

 Benoīt


 ___
 Qgis-developer mailing list
 Qgis-developer@lists.osgeo.org
 http://lists.osgeo.org/mailman/listinfo/qgis-developer


___
Qgis-developer mailing list
Qgis-developer@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/qgis-developer