Re: [QGIS-Developer] 3.4 backports
On Tue, 17 Sep 2019 at 09:28, Nathan Woodrow wrote: > > Is it easy to disable the auto backport and have it just create the pull > request and tag someone? That way it can still be semi-automatic but with > some human checks That's what it does -- it creates the PR, but the PR can't be merged without someone (other than the original developer) approving it. I'm seeing a lot of backports which the bot can't auto create over the last week, where devs have manually backported... yet the validity of the backport is questionable and I'd hate to see wasted effort going into manual backports which won't be accepted. Nyall > > On Tue, Sep 17, 2019 at 9:06 AM Nyall Dawson wrote: >> >> Hi all, >> >> Just a quick plea that during this freeze period please think >> carefully about the suitability of backporting fixes to 3.4. At this >> stage in its lifecycle 3.4 should only receive absolutely mission >> critical fixes (such as crash fixes or data corruption issues). >> >> The consequence of regressions in 3.4 caused by backporting minor bug >> fixes much outweighs the benefits of fixing these bugs. While the >> backport label makes 3.4 backports easy to do, it does not guarantee >> their safety or suitability. >> >> Thanks! >> Nyall >> ___ >> QGIS-Developer mailing list >> QGIS-Developer@lists.osgeo.org >> List info: https://lists.osgeo.org/mailman/listinfo/qgis-developer >> Unsubscribe: https://lists.osgeo.org/mailman/listinfo/qgis-developer ___ QGIS-Developer mailing list QGIS-Developer@lists.osgeo.org List info: https://lists.osgeo.org/mailman/listinfo/qgis-developer Unsubscribe: https://lists.osgeo.org/mailman/listinfo/qgis-developer
Re: [QGIS-Developer] 3.4 backports
Is it easy to disable the auto backport and have it just create the pull request and tag someone? That way it can still be semi-automatic but with some human checks On Tue, Sep 17, 2019 at 9:06 AM Nyall Dawson wrote: > Hi all, > > Just a quick plea that during this freeze period please think > carefully about the suitability of backporting fixes to 3.4. At this > stage in its lifecycle 3.4 should only receive absolutely mission > critical fixes (such as crash fixes or data corruption issues). > > The consequence of regressions in 3.4 caused by backporting minor bug > fixes much outweighs the benefits of fixing these bugs. While the > backport label makes 3.4 backports easy to do, it does not guarantee > their safety or suitability. > > Thanks! > Nyall > ___ > QGIS-Developer mailing list > QGIS-Developer@lists.osgeo.org > List info: https://lists.osgeo.org/mailman/listinfo/qgis-developer > Unsubscribe: https://lists.osgeo.org/mailman/listinfo/qgis-developer ___ QGIS-Developer mailing list QGIS-Developer@lists.osgeo.org List info: https://lists.osgeo.org/mailman/listinfo/qgis-developer Unsubscribe: https://lists.osgeo.org/mailman/listinfo/qgis-developer
[QGIS-Developer] 3.4 backports
Hi all, Just a quick plea that during this freeze period please think carefully about the suitability of backporting fixes to 3.4. At this stage in its lifecycle 3.4 should only receive absolutely mission critical fixes (such as crash fixes or data corruption issues). The consequence of regressions in 3.4 caused by backporting minor bug fixes much outweighs the benefits of fixing these bugs. While the backport label makes 3.4 backports easy to do, it does not guarantee their safety or suitability. Thanks! Nyall ___ QGIS-Developer mailing list QGIS-Developer@lists.osgeo.org List info: https://lists.osgeo.org/mailman/listinfo/qgis-developer Unsubscribe: https://lists.osgeo.org/mailman/listinfo/qgis-developer
Re: [QGIS-Developer] windows build qsci dll issues
On Tue, 17 Sep 2019 at 06:43, Alexis R.L. wrote: > > Greetings, > > I've been trying to build on windows and I'm having issues with python. I > have set the python exe in cmake to the python 3.7 in osgeo apps (and not the > python3 in bin since that was causing issues), the pythonpath and pythonhome > also point to the python in osgeo/bin but when it comes time to execute "from > PyQt5.Qsci import..." I get a dll error. > > I have used pip to download qscintilla and the sip bindings again, but this > did not change anything. > > Has anyone else had that issue? Where is this dll so that I can add it to my > Path environment? I can't help with this particular issue. But for reference, here's my "just works" from-scratch Windows build process, using Qt Creator. No warranty, zero support! If it works for you it works ;) Rough notes only... - I install MS Visual Studio community 2015, just the c++ language related parts (using http://download.microsoft.com/download/D/2/3/D23F4D0F-BA2D-4600-8725-6CCECEA05196/vs_community_ENU.exe) - I install cmake (64 bit) and add that to the windows path - I download ninja and shove ninja.exe in c:\program files\cmake\bin (yep - can't get this to work otherwise!) - I use cygwin (64 bit) to install flex and bison - I use osgeo4w 64 bit, installed to the default location. I use this to install qgis-dev-deps - I personally install WSL and ubuntu for windows, and use that for git and stuff. It won't affect things, but I find it easier to use this to get sipify etc up and running because I can just use ubuntu bash commands... In qt creator kits: - Qt versions: manually add C:\OSGeo4W64\apps\Qt5\bin\qmake.exe - Compilers: should autodetect the msvc stuff, e.g. Microsoft Visual C++ Compiler 14.0 (x86) under C:\Program Files (x86)\Microsoft Visual Studio 14.0\VC\vcvarsall.bat x86 . Should be about 14 entries here, automatically populated - Debuggers: Should autodetect C:\Program Files (x86)\Windows Kits\10\Debuggers\x86\cdb.exe and C:\Program Files (x86)\Windows Kits\10\Debuggers\x64\cdb.exe - Cmake: should autodetect C:\Program Files\CMake\bin\cmake.exe , otherwise manually add it Kits: - manually add a new kit: - compiler: set both c and c++ to Microsoft Visual C++ Compiler 14.0 (x86_amd64) (make sure it's the 64 bit one!) - debugger and qt version: set to above ones - cmake - set to system cmake, as above. Change generator to "ninja" - Click "Change" next to environment, and insert: GDAL_DATA=c:\osgeo4w64\share\gdal GDAL_DRIVER_PATH=c:\osgeo4w64\bin\gdalplugins PATH=c:\program files\cmake\bin;C:\cygwin64\bin;C:\OSGeo4W64\bin;C:\OSGeo4W64\apps\Python37;C:\OSGeo4W64\apps\Python37\Scripts;c:\osgeo4w64\apps\qt5\bin;C:\OSGeo4W64\apps\Python37\sip\PyQt5\;${PATH} PYTHONHOME=C:\OSGeo4W64\apps\Python37 PYTHONPATH= QGIS_INSTALL_PATH=c:\osgeo4w64\apps\qgis-dev QML2_IMPORT_PATH=C:\OSGeo4W64\apps\Qt5\qml QT_PLUGIN_PATH=C:\OSGeo4W64\apps\qgis-dev\qtplugins;C:\OSGeo4W64\apps\Qt5\plugins Phew. Now open the qgis project. In the cmake config, set CMAKE_BUILD_TYPE to "RelWithDebInfo". Make sure your build environment looks good, i.e. has the above entries. Build! Cross fingers and close eyes... Nyall ___ QGIS-Developer mailing list QGIS-Developer@lists.osgeo.org List info: https://lists.osgeo.org/mailman/listinfo/qgis-developer Unsubscribe: https://lists.osgeo.org/mailman/listinfo/qgis-developer
[QGIS-Developer] windows build qsci dll issues
Greetings, I've been trying to build on windows and I'm having issues with python. I have set the python exe in cmake to the python 3.7 in osgeo apps (and not the python3 in bin since that was causing issues), the pythonpath and pythonhome also point to the python in osgeo/bin but when it comes time to execute "from PyQt5.Qsci import..." I get a dll error. I have used pip to download qscintilla and the sip bindings again, but this did not change anything. Has anyone else had that issue? Where is this dll so that I can add it to my Path environment? Thanks Alex ___ QGIS-Developer mailing list QGIS-Developer@lists.osgeo.org List info: https://lists.osgeo.org/mailman/listinfo/qgis-developer Unsubscribe: https://lists.osgeo.org/mailman/listinfo/qgis-developer
[QGIS-Developer] Plugin [1526] LAStools approval notification.
Plugin LAStools approval by pcav. The plugin version "[1526] LAStools 1.4" is now approved Link: http://plugins.qgis.org/plugins/LAStools/ ___ QGIS-Developer mailing list QGIS-Developer@lists.osgeo.org List info: https://lists.osgeo.org/mailman/listinfo/qgis-developer Unsubscribe: https://lists.osgeo.org/mailman/listinfo/qgis-developer
Re: [QGIS-Developer] Documentation wish list for "hackfesters"
Hi all, The hackfest has gone for weeks, but QGIS is still there as well as writers and developers. So I wonder if someone had considered the opportunity to give a hand on any of the identified issues. *NB: It could not be for the "Spare Time Inc" since we still have unallocated funds (and not "founds" - damn keyboard).* Regards, Harrissou Le ven. 23 août 2019 à 18:21, DelazJ a écrit : > Hi all, > > With the QGIS Community Meeting beginning, a bunch of developers and > writers will meet to brainstorm and improve QGIS. Here are some suggestions > of topics (among many others) some documentation contributors would be > REALLY interested in having a fix for. We need a developer to address and > ease our "life" (sounds tragical enough? :D ). Well. > > In short: > 1/ *Improve the automatic creation of screenshots* [1] > 2/ *Redirect docs links *and ensure that changing links in docs does *not > break the help button* in application [3] > 3/ *Document/improve the doctest architecture for PyQGIS Cookbook *(kind > of [4]) > > For the details: > 1/ QGIS is moving fast. Dialogs are changing at every release, making > screenshots in the docs obsolete. There was a plan to avoid writers > manually take each screenshot, and instead generate them in series, > browsing tabs of a specified dialog eg. A work had been initiated in the > QgsAppScreenshots class [1]. It could be nice to get more news on it. > Incidentally, with such a feature, we can expect more translated > screenshots (which may not be good for our server but helpful for > non-English readers and translators). > > 2/ Because features are getting bigger, we sometimes need to reorganize > the chapters in documentation in order to match how things behave in the > application and also avoid very lng chapters. Moving sections or > renaming them (to match their actual scope) breaks the Help button > associated to the feature, for at least three releases (the next-to-be LTR > and the two previous versions since they point to the same doc). To fix > that, doc writers have to update the links in the code and this change only > applies to current release(s), not for EOLed one (eg 3.6 could have more > not working links than 3.8 or master - we have a script for checks [2]). > We need to find a more viable way to address this issue and redirect > requested links to their actual place in the corresponding doc. > This issue is a long standing one that has been reported/discussed a > couple of times already (eg [3] has some suggestions) > > 3/ During the La Coruna hackfest, a doctest infrastructure has been set up > to ensure code samples provided for PyQGIS Cookbook are still working. > Months later, it looks to me like it's less used than it should: > - partly, because we did not make it mandatory, > - partly because it does not apply (or we did not figure out how?) in > some cases. Eg it seems it does not know about iface [4]. Also there are > test projects the system should rely on but nobody uses it and afaics > there's no example someone not used to this could follow. > It could be nice clarifying the system and help writers adopt it. > > [1] https://github.com/qgis/QGIS/blob/master/src/app/qgsappscreenshots.h > [2] https://github.com/qgis/QGIS/blob/master/scripts/chkdoclink.sh > [3] > https://github.com/qgis/QGIS-Documentation/pull/2258#issuecomment-352246124 > [4] https://github.com/qgis/QGIS-Documentation/issues/3776 > > Et voilà... Sorry to have been long. Looking forward... > And for information, docs still have unallocated founds... > > Happy hackfest, > Harrissou > ___ QGIS-Developer mailing list QGIS-Developer@lists.osgeo.org List info: https://lists.osgeo.org/mailman/listinfo/qgis-developer Unsubscribe: https://lists.osgeo.org/mailman/listinfo/qgis-developer
[QGIS-Developer] Plugin [1135] History viewer for a PostgreSQL base with audit triggers approval notification.
Plugin History viewer for a PostgreSQL base with audit triggers approval by pcav. The plugin version "[1135] History viewer for a PostgreSQL base with audit triggers 2.1" is now approved Link: http://plugins.qgis.org/plugins/pg_history_viewer/ ___ QGIS-Developer mailing list QGIS-Developer@lists.osgeo.org List info: https://lists.osgeo.org/mailman/listinfo/qgis-developer Unsubscribe: https://lists.osgeo.org/mailman/listinfo/qgis-developer
Re: [QGIS-Developer] Upgrading to V3
Hi, I can't really answer your question but you should use https://qgis.org/api/classQgsMapLayerComboBox.html instead of creating your own combobox. (if possible). Etienne Le dim. 15 sept. 2019 à 10:11, Maaza Mekuria a écrit : > QGIS Layer List upgrade to V3 C++ plugin - Geographic Information Systems > Stack Exchange body,.top-bar{margin-top:1.9em} > > I was upgrading a c++ plugin that I had used in the past and almost > eliminated all the the usual compiler errors due to version changes > according for the current version (3.6). The plugin used to collect current > map canvas layers in a combo list box for selection. I am getting two > errors that I can't figure out what to do with. This code compiled without > problem pre-2.0. > > The first happens to be about QGSMapLayer that can't be instantiated > because it is turned as an abstract class and I am getting an error > [Severity Code Description Project File Line Source Suppression State Error > C2259 'QgsMapLayer': cannot instantiate abstract class tarjan_sccplugin > C:\OSGeo4W64\apps\Qt5\include\QtCore\qlist.h 214 Build ] > > This is portion of the code generating the error. Where cmbNetSourceLayer > is a combo-box . > > void tarjan_sccGui::populateLayers( QString const& url ) { > cmbNetSourceLayer->clear(); > > QString layerName; QString layerType; > > // fill list of layersQMap mapLayers = > QgsProject::instance()->mapLayers();QMap::iterator > layer_it = mapLayers.begin(); > for (; layer_it != mapLayers.end(); ++layer_it){ > QgsVectorLayer* vl = dynamic_cast(layer_it.value()); > if (!vl) > continue; > else > > cmbNetSourceLayer->addItem(vl->name() );} > > The second error appears to be in explicit declaration for the plugin. > Something has changed in how the plugins are initiated. [Severity Code > Description Project File Line Source Suppression State Error (active) > invalid explicit instantiation declaration tarjan_sccplugin > d:\GIS\v3\Quantum-GIS\src\core\qgsoptionalexpression.h 82 IntelliSense ] > > No idea what must be done to overcome this error. The cited module > appears to be an optional data module. > > > Any help is appreciated. > ___ > QGIS-Developer mailing list > QGIS-Developer@lists.osgeo.org > List info: https://lists.osgeo.org/mailman/listinfo/qgis-developer > Unsubscribe: https://lists.osgeo.org/mailman/listinfo/qgis-developer ___ QGIS-Developer mailing list QGIS-Developer@lists.osgeo.org List info: https://lists.osgeo.org/mailman/listinfo/qgis-developer Unsubscribe: https://lists.osgeo.org/mailman/listinfo/qgis-developer