request with the needed changes is not possible.
Mark Johnson, Berlin Germany
___
Qgis-developer mailing list
Qgis-developer@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/qgis-developer
I have opened an issue at:
https://hub.qgis.org/issues/12510
the I suggest that plugins that use a hardcoded spatialite-amalgamation be
depreciated since they can damage a db created with a later spatialite
version.
Mark Johnson, Berlin Germany
er
or submit it now, completing it within the next 7-10 days.
There is also the question: keep it short or (what I would prefer)
extensive?
Mark Johnson, Berlin Germany
___
Qgis-developer mailing list
Qgis-developer@lists.osgeo.org
List info: http://lis
>
> Hey Mark,
>
> A QEP should be opened in order for review to take place on any design
> changes that might need descussion.
>
The 'design changes' will be more on how the Georeferencer will work, not
QGIS 3.0 itsself.
> Also please note that 2.x is not receiving any new features as we are now
>
ng correct results
- when Topological error exists
-- and should be corrected
Mark Johnson
___
Qgis-developer mailing list
Qgis-developer@lists.osgeo.org
List info: http://lists.osgeo.org/mailman/listinfo/qgis-developer
Unsubscribe: http://lists.osgeo.org/mailman/listinfo/qgis-developer
>
> thanks for your feedback. Shouldn't this be reported by ST_IsValid also?
>
Yes, it should.
SELECT
ST_GeomFromText('POLYGON((7 7, 77 77, 35 35,7 7))') AS bad_poly,
ST_IsValid(ST_GeomFromText('POLYGON((7 7, 77 77, 35 35,7 7))')) AS
bad_poly_valid,
ST_IsValidDetail(ST_GeomFromText('POLYGON((7
>
> let me be more clear: there are NO invalid geoms in the DB, as reported
> by st_isvalid. what is strange:
> * invalidities are reported in qgis log (this may be due to
> simplification, thanks Alvaro)
>
>
Then attept an 'ST_Simplify(geom,tolerance)' on these geometries to see if
these errors oc
ndering using MapUnit
- retrieve the 1 meter value
- multiply by the meter value set in the MapUnit property and use that
result
This would be an API break that should be done now.
Mark Johnson, Berlin Germany
___
Qgis-developer mailing list
Q
To add to the last note:
If API break is not desired, a 'MetersUnit' could be used
- that would implement the the 'transformation' to MapUnits
That might even be better to avoid any breaks and be clear for everyone.
Mark Joh
> Are you referring to symbology here? Ie adding a new "map length (metres)"
Yes, where the combo-box show
- millimeter
- pixels
- Map units
one could add
- Meters to Map units
Presently definded in QgsSymbolV2 and calulated in
QgsSymbolLayerV2Utils::convertToMapUnits
and
QgsSymbolLayerV2Utils::c
case of
degrees
- the use the librttopo functionality as a part of the 'QgsGeos' would be
the most effective way to deal with this
This would be a new dependency, but would also offer many other
functionalities (such as Topology).
Mark Johnson, Berlin Germany
__
try this out and see if it brings a reasonable result.
Mark Johnson, Berlin Germany
___
Qgis-developer mailing list
Qgis-developer@lists.osgeo.org
List info: https://lists.osgeo.org/mailman/listinfo/qgis-developer
Unsubscribe: https://lists.osgeo.or
>> I'm confused - what is missing?
The topic here is that when using MapUnits
- where the MapUnits are degrees
that a constant value (valid around the world) is not possible.
If you set this to 5 (intended as meters on the map)
- the further you zoom in, the circle will get bigger always showing
ly installing Qwt
Building QScintilla for PyQt5
Building PyQt5
QGIS3 ccmake
I would be willing to start this with my experiences.
Mark Johnson, Berlin Germany
___
Qgis-developer mailing list
Qgis-developer@lists.osgeo.org
List info: https://lis
>> Can you clarify on which systems you are working and which dependencies
cause headaches?
In my case I am working on an old machine, where many things have been
installed in the past and cause conflicts. So until a new machine can be
bought I am attempting to get this done in a custom application
>> You should instead make your contributions to the "developers guide" -
source is at https://github.com/qgis/QGIS-Documentation/tree/release_2.
18/source/docs/developers_guide.
I assume you mean the wiki area, which is nice since I have experience in
that type of wiki
--> https://github.com/geop
>> No, I mean in the .rst files. Send the changes via a PR. Then it'll be
added to the developers guide at
Well, I have no idea what :
* .rst files are
* how to make then or
* how to view them
The wiki pages I can do on the fly while working this out.
Can you live with a wiki version ?
- if not I
initial
release of QGIS 3.
The present master version does not read tables that have more than one
geometry correctly.
Mark Johnson, Berlin Germany
___
Qgis-developer mailing list
Qgis-developer@lists.osgeo.org
List info: https://lists.osgeo.org/mailman
>> I see no use of OGR_F_GetGeomFieldRef() so it seems
that you select among the various different geometry fields by relying on
the
"layer_name(geometry_field_name)" syntax
That is caused by the fact that the initial list (QgsOgrProvider::subLayers())
returns only the "layer_name(geometry_field_n
>
> In the provider we could decide to allocate layer ids diffently than OGR .
>
>
>> Example if you have a dataset with
>
> - layer1 (geom1, geom2)
>
> - layer2 (geom)
>
>
>> Then you would map that to :
>
> - qgis_ogr_provider_layer_id = 1: layer1, geom1
>
> - qgis_ogr_provider_layer_id = 2: laye
>
> Application will crash with a Lookup error?
Yes. And that's perfectly fine. That's only an issue that can happen for us
>> developers playing
>
> with different versions, and we know how to solve those issues.
>
>
The proposed solution for this constellation for-sees
- at application start: r
>
> Please keep in mind that's there's some non-negotiables here:
>
One way or another this matter needs to be resolved before an initial
> release of QGIS 3.
This topic is only about QGIS 3.0 only.
Mark
___
Qgis-developer mailing list
Qgis-developer@
There are different issues you've raised and that we might want handle or
> not:
>
> 1) make sure that a URI points to the same layer if the datasource evolves
> (new layer)
>
> 2) be able to deal with the corner cases of layers with same name
>
>
>
> 1) involves that we cannot always rely on the i
>
> Isn't this about PR #3697 - which is opened against 2.18? If #3697 is
> abandoned, can I close to clean up the PR queue?
Since the problem is not resolved, no.
The present QGIS 3.0 cannot table with more than geometry.
- at that time there was no QGIS 3
I changed that to 3.0 this morning.
M
>
>
> 1. No extra code will be allowed in master to handle GDAL < 2.0. QGIS
> 3.0 has a hard GDAL >= 2.0 requirement, and no code will be accepted
> to work around this requirement.
The original code did many tasks being coded multiple times
- these common tasks are written once in a extra functi
>
> The layer name can be stored in subLayers() for conveniency, since it has
>> usually a short life. And for subLayers() / askUserForOGRSublayers() make
>> the layer id what is used.
>
> subLayers is not used when read from a project, only when a new source is
added.
>
>> Not sure to what you ar
>
> Great - looking forward to seeing you push this to the PR! It'll make
> the remaining changes much easier to review.
Thats the way it was done when submitted..
Mark
___
Qgis-developer mailing list
Qgis-developer@lists.osgeo.org
List info: https://l
>
> Yet I still see lots of code in that PR like:
>
> if defined(GDAL_VERSION_NUM) && GDAL_VERSION_NUM >=
> GDAL_COMPUTE_VERSION(2,0,0)
>
> That's what I'm referring to. All this conditional code needs removal
> (unless it's testing for something in GDAL > 2.0 ).
Mosty in the static functions
- Q
>
> Ok - I find this thread very confusing. Does this acknowledgement mean
>
> that you WILL remove these ifdefs from the PR?
>
> Yes, indeed.
>
>> Here's the thing: I *want* your work to land in QGIS. It's important
>
> stuff, and needs to be addressed. But in its current form it won't be
>
> mer
>
> So if we want to address both we need the layer_id and layer_name. I'd say
> when you build a OGR URI (mostly in QgisApp::askUserForOGRSublayers()),
> then look at the uniqueness of the layer names. If there's layer name
> unicity, then use only the layer name. Otherwise fallback to uniquely th
>
> So if both a layerid and layername, you would prever layerid over
> layername ?
no, the other way around
Default: layername, only when a duplicate layername exists, should the id
be used
Checking for duplicate names is done in subLayers()
- if layername is unique, id will be set to -1 and sen
>
> That will *not* work for other formats that can have multiple geometry
> fields, like GML.
Could you send me such a GML to test this?
Have you ever seen or have a GML with a duplicate layer-name?
but it would probably be cleaner to have a dedicated URI parameter for the
> geometry column name
>
> Good - I think it will get outdated quickly and history shows that such
> content does not get well maintained. If we had to do anything along these
> lines, it could be to have a small chapter which lists all the QGIS
> dependencies along with a pointer to the upstream build instructions for
>
> On Ubuntu 17.04, which is based on Qt5.7.1, output from
>
> `./output/bin/qgis`
>
>
>> `Couldn't load PyQt.
>
> Python support will be disabled. (a really big deal)
>
>
>> Traceback (most recent call last):
>
>File "", line 1, in
>
> ImportError: /usr/lib/x86_64-linux-gnu/libQt5Core.so.5: v
>
> > Can you live with a wiki version ?
>
> > - if not I will stop it
>
>
>> Hi Mark,
>
>
>> if you insist ( and I see you do :-) ):
>
> Well, nobody said it should be stopped
>
>> Then PLEASE create a system in which you make it clear
>
> WHICH version of QGIS you are trying to build on
>
> each
>
> Python support will be disabled.
>
I assume that these errors came from QGIS3-Dialog: correct?
That would mean that QGIS3 has found all of the libQt5* it needs
- and thus all paths are correct
The used python3 calls
- /usr/local/lib/python3.5/dist-packages/PyQt5/QtCore.so
so the pythonpath sho
I noticed that although many of the classes 'V2' have been renamed, the
class 'QgsComposerTableV2' (together with QgsComposerTextTableV2) has not.
(Just in case this was overlooked).
Mark
___
Qgis-developer mailing list
Qgis-developer@lists.osgeo.org
Li
>
> When I am using the georeferencer plug in, the software is crashing and
> the "the application stopped unexpectedly" message appears. Does anyone
> knows why?
Here it would be important to learn, when using the georeferencer, what the
working conditions are:
How many gcp-points are being use
r.
At the moment only String / Int / Double are supported.
Mark Johnson
___
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/lis
>> What makes you think that?
Based on the implementation of the QgsSpatiaLiteProvider (which only
supports TEXT, INTEGER and DOUBLE) and the documentation of QgsField.
>> Constructor. Constructs a new QgsField object.
* \param type Field variant type, currently supported: String / Int /
Doub
Good, then I will work those samples into QgsSpatiaLiteProvider.
For Spatialite 4.5.0 QgsSpatiaLiteProvider needs to be worked on, which I
am doing now.
Mark
___
QGIS-Developer mailing list
QGIS-Developer@lists.osgeo.org
List info: https://lists.osgeo.or
>> what are your thoughts on
http://osgeo-org.1560.x6.nabble.com/Future-of-spatialite-provider-td5311979.
html
?
For gpkg: should remain solely with OGR.
For VectorLayers (SpatialTables and SpatialViews) is more problematic,
since OGR does not support writable SpatialViews, which I use heavily.
>> What GDAL version? You need GDAL >= 2.1 for M values
Also the QGIS running must be compiled with GDAL >= 2.1 for this to work.
Mark
___
QGIS-Developer mailing list
QGIS-Developer@lists.osgeo.org
List info: https://lists.osgeo.org/mailman/listinfo/qgis
>> QGIS seems to need to draw the entire background in it's head, then
displays it all at once
Yes, this is a problem (it seems) with all of the the Raster rendering.
With a GeoTiff during Panning, the Background blanks out and pops back,
while the Vectors don't.
Hard on the eyes when searching fo
t::ISODate
)));
(QDateTime can store lower Dates, but they cannot be displayed as an
ISODate)
If somebody could give a hint as to where the 'Feature Attributes" Dialog
is, I could try this out tomorrow.
Mark Johnson
___
QGIS-Developer mailing
atabase
- or where you wish to *transform* to another projection (i.e.
ST_Transform(geometry,3068)
With "SELECT InsertEpsgSrid(3068)"
- you would add an 'EPSG:3068' entry to 'spatial_ref_sys'
http://www.gaia-gis.it/gaia-sins/spatialite-sql-latest.html#p16
Mark Johns
umed, that since Grass also has both a Raster and Vector provider
that this would not cause a problem.
Some sort of extra step seems to be needed that I have overlooked.
Mark Johnson, Berlin Germany
___
QGIS-Developer mailing list
QGIS-Developer@lists.osge
>
> I'm not sure if you are aware of it, but GDAL 2.2 SQLite driver supports
> rasterlite2 datasets
> (in read-only), provided it is built against librasterlite2.
>
> Even
Yes, I know.
This a part of the 'overhaul' of the QgsSpatiaLiteProvider in preparation
for Spatialite 4.5.0 which will contai
QgsSpatiaLiteProvider and adding a QgsRasterLite2Provider is needed.
https://groups.google.com/forum/#!topic/spatialite-users/xC65HkmUbPw
The pull request for these changes has not yet been reviewed.
Mark Johnson, Berlin Germany
___
QGIS-Developer mailing list
QGIS
My assumption is that it a Provider would be a major source to gather the
needed information an set QgsLayerMetadata.
I am doing this now for the new QgsSpatiaLiteProvider.
QgsDataProvider does not contain (at present) a mMetadata; member as
QgsMapLayer does.
So when setDataProvider runs in QgsVe
, whereas the
proposed new version (for QGis 3.2) does.
Mark Johnson, Berlin Germany
___
QGIS-Developer mailing list
QGIS-Developer@lists.osgeo.org
List info: https://lists.osgeo.org/mailman/listinfo/qgis-developer
Unsubscribe: https://lists.osgeo.org
lite connection it is very easy to create an empty pure GPKG
Database (106.5 Kb) with:
SELECT gpkgCreateBaseTables();
It would contain only 3 SRS (-1,0 and 4326) in 'gpkg_spatial_ref_sys' with
all of the other GPKG admin-tables (including 'gpkg_extensions&#x
to new dependencies that have been added
Otherwise I would be willing to cooperate with anybody that could get this
working as a part of QGIS.
Mark Johnson, Berlin Germany
___
Qgis-developer mailing list
Qgis-developer@lists.osgeo.org
List info: http://lists.osg
s EPSG:25832. The shape file has the following .prj file:
ETRS_1989_UTM_Zone_32N
- with a 'Z' in 'Zone' as apposed to the small 'z' in 'zone'
If QGIS is only checking this, then: 'z' != 'Z'
- checking with a 'toLower' could re
on is based on the code of 2 years ago, but will be
updated when the integration has been completed.
Mark Johnson, Berlin Germany
___
Qgis-developer mailing list
Qgis-developer@lists.osgeo.org
List info: http://lists.osgeo.org/mailman/listinfo/qgis-developer
Unsubscribe: http://lists.osgeo.org/mailman/listinfo/qgis-developer
eopackage specification had not occurred to me.
Mark Johnson
___
Qgis-developer mailing list
Qgis-developer@lists.osgeo.org
List info: http://lists.osgeo.org/mailman/listinfo/qgis-developer
Unsubscribe: http://lists.osgeo.org/mailman/listinfo/qgis-developer
>
> Will it easily be possible to extract such a photo from the spatialite file?
>
> No, once created, the database itself could retrieve all the data without
spatialite
- the exception being that the geometry could not be read
The position(s) are stored in extra fields
The image itself is a simple
centrating on the map while moving,
so this 'blinking' is truly a strain on the eyes.
---
The above solution seems to have resolved the first issue, since it went on
to 54% where it failed while linking qgis.
(Undefined reference
to `QgsRelationManagerDialog::setLayers(QList const&
g++ 4.9.1
After looking for the cause of the link problem, I came to the conclusion
that my (rather old) git copy has gotten a bit 'confused'.
I have made a new copy and have started again (without adding '-std=c++11')
- it is now at 40%, well see how fart it goes.
___
reate these Styles and export them would then a great help for
the Database designer.
Mark Johnson, Berlin Germany
___
Qgis-developer mailing list
Qgis-developer@lists.osgeo.org
List info: http://lists.osgeo.org/mailman/listinfo/qgis-developer
Unsubs
oal of the Grant Programme '9) Correction of QgsOgrProvider
implementaion of GDAL 2.0' was to resolve this problem.
Mark Johnson
___
Qgis-developer mailing list
Qgis-developer@lists.osgeo.org
List info: http://lists.osgeo.org/mailman/lis
Sorry, I misread the request.
In this case, if I have now understood correctly, you are opening the
source before retrieving the layer.
That would then be everything up to the first '|'.
For a file-base source the full-filename with path.
For a Database-System such as PostGis the Connection-Str
Be warned though that when using gdal 2.*
- GetLayer with the layerid will not (always) work as it did with gdal 1.*
Mark
___
Qgis-developer mailing list
Qgis-developer@lists.osgeo.org
List info: http://lists.osgeo.org/mailman/listinfo/qgis-developer
Uns
= OGR_DS_GetLayer( ogrDataSource, iLayerIndex );
or with the layer-name
ogrLayer = OGR_DS_GetLayerByName( ogrDataSource, TO8( sLayerName ) );
when all tasks are completed, the Datasource is closed.
Inside QgsOgrProvider all of this is done as needed, but when using the
OGR-API directly each
yName( ogrDataSource, TO8( sLayerName ) );
This should work for both PostgreSQL and spatialite based on your samples.
Mark Johnson
___
Qgis-developer mailing list
Qgis-developer@lists.osgeo.org
List info: http://lists.osgeo.org/mailman/listinfo/qgis-developer
Unsubscribe: http://lists.osgeo.org/mailman/listinfo/qgis-developer
>
> But the above is NOT an ogrDataSource
yes, but you are using the OGR-API directly
- so you must use the OGR-Syntax
i.e. you must 'translate' the syntax used in QgsSpatiaLiteProvider to
something OGR will understand.
___
Qgis-developer mailing list
Q
>
> > Running the automated test is done via:
> >
> > ctest -V -R ProcessingToolsTest
I have never used this before, from where should it be called?
Mark
___
Qgis-developer mailing list
Qgis-developer@lists.osgeo.org
List info: http://lists.osgeo.org
Sorry, but I often have problems with python ...
Are you supposed to configure something beforehand?
This is what I get:
UpdateCTestConfiguration from
:.../QGIS/build_release-2_16/DartConfiguration.tcl
Parse Config
file:/home/mj10777/000_links/gis_software/qgis_git/mj10777/QGIS/build_release-2_1
>
> Didn't you use "ctest" ?
Yes, the first list of errors was from that
> PYTHONPATH to include ${BUILDDIR}/output/python/plugins.
Have done that and now am building the latest branch
> using latest tree
> in release-2_16 branch
Before you wrote branch 'master_2'
- that is what I checked ou
Would this halt all updates in the 2 branch?
- I was hoping to submit the needed changes for gdal 2.* in QgsOgrProvider
in the next week or so
Mark Johnson
___
Qgis-developer mailing list
Qgis-developer@lists.osgeo.org
List info: http://lists.osgeo.org
question:
How can the srid of a geometry retrieved from a QgsVectorLayer be
determined?
Mark Johnson
___
Qgis-developer mailing list
Qgis-developer@lists.osgeo.org
List info: http://lists.osgeo.org/mailman/listinfo/qgis-developer
Unsubscribe: http://lists.osgeo.org/mailman/listinfo/qgis-developer
should be removed.
Reasons:
As gdal evolves, it will become more difficult to 'emulate' OGR functions
that do not exist in gdal 1.*
- the inclusion of which would cause a 'symbol lookup error'
A proper support for M-Values will probably become more important in the
future than it is now.
---
(sorry if this is a bit long winded, but one has to start somewhere)
Mark Johnson, Berlin Germany
___
Qgis-developer mailing list
Qgis-developer@lists.osgeo.org
List info: http://lists.osgeo.org/mailman/listinfo/qgis-developer
Unsubscribe: http://lists.osgeo.org/mailman/listinfo/qgis-developer
>
> I think you should target master only for your changes (and consider only
> GDAL
> >= 2 case). It is unlikely they are appropriate for backporting, or that
> should be considered in a later stage depending on the impact they have.
Since on my older machine, getting master to run properly is pr
ng
the extra functionality of Spatialite 5.0.
Mark Johnson, Berlin Germany
___
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
uture developments.
It is also likely that Sandro will take a closer look after the present
development phase is over.
Mark Johnson, Berlin Germany
___
QGIS-Developer mailing list
QGIS-Developer@lists.osgeo.org
List info: https://lists.osgeo.
>> Determining if the function gaiaStatisticsInvalidate exists in the
C:/OSGeo4W/lib/spatialite_i.lib failed with the following output:
gaiaStatisticsInvalidate has existed since Spatialite 4.2, so a version
before that must of been found.
Could you show a list of files in C:/OSGeo4W/lib/ that co
>> C:\OSGeo4W/lib\spatialite.lib
>> C;\OSGeo4W\lib\spatialite_i.lib
This naming convention seems to be normal for OSGeo4W.
I found a site that contains the 32 bit Spatialite 4.3 version:
http://www.norbit.de/~jef/osgeo4w/x86/release/spatialite/
http://www.norbit.de/~jef/osgeo4w/x86/release/spati
>> Determining if the function gaiaStatisticsInvalidate exists in the
C:/OSGeo4W/lib/spatialite_i.lib failed
yes, the first one says it cannot find gaiaStatisticsInvalidate.
But if the file size is the same it should be found in spatialite_i.lib,
then it looks as if the cmake routine 'CheckFunctio
That is very likely the cause.
At the moment I am working on getting Spatialite/RasterLite2 working on
android with arm(32), arm64 X86_64.
https://github.com/geopaparazzi/libjsqlite-spatialite-android/wiki/Android_4.5.0.mk
___
QGIS-Developer mailing li
, QgsMapLayer/QgsVectorLayer
classes but I fail to see what is being done differently.
If an hint could be given where this done, I would be grateful.
Mark Johnson
___
QGIS-Developer mailing list
QGIS-Developer@lists.osgeo.org
List info: https
ssume that 'CachingEnabled' causes this
effect (Raster not being 'refreshed' but just re-using the existing image,
while redrawing the changed geometry).
BTW: what is the cause of the 'blank' screen effect when the raster is
being refreshed?
Mark Johnson
A possible cause is that the given path is not relative to the diretory
from which QGIS was started from.
Possibly the value of :
QGIS_PREFIX_PATH
found in Options/System/Current environment variables
Mark
___
QGIS-Developer mailing list
QGIS-Developer
Am Do., 26. Juli 2018 um 15:32 Uhr schrieb Mark Johnson <
mj10...@googlemail.com>:
> A possible cause is that the given path is not relative to the diretory
> from which QGIS was started from.
>
> Possibly the value of :
> QGIS_PREFIX_PATH
>
Sorry, I meant:
PWD
>> Executing PRAGMA foreign_keys = ON
This is also done for Spatialite in the SpatialiteProvider, so this is not
a general GeoPackage issue.
The need to keep this on is important, since the TRIGGERS enforces the
'sanity' of the data (such as same srid and geometry-type rules).
Also the Database-D
>> We use a direct port from GRASS:
This may also be caused by:
/*
* From original rttopo documentation:
* Tested against:
* http://mascot.gdbc.gov.bc.ca/mascot/util1b.html
* and
* http://www.ga.gov.au/nmd/geodesy/datums/vincenty_direct.jsp
*/
QgsPointXY QgsDistanceArea::computeSphero
85 matches
Mail list logo
