QWAT
QWAT copied to clipboard
qwat
[QGIS] Snapping on invisible feature
In QGIS, we can snap on invisible features even if the snapping options have been defined as "snap on all visible layers". The snapping options don't take into accoount the layer style.
Will this be fixed for QGIS 3.0 ?
Démo GIF : https://share.oslandia.net/public/b91c8e
Hi @dsavary let's split it in two phases:
-
do you have qgis 2.99 installed? That could let you test directly in qgis nightlies. You can have them using the osgeo4w installer, and updating is only a 5-6 click job (just run the osgeo4w/bin/setup.bat and choose advanced installer)
-
before fixing what appears to you as a bug, it must be discussed and see if it really is a bug or a feature for all users. When it appears strange to you, it might be a useful feature to other. When you say layer style is not taken into account, are you talking about invisible feature that are:
- unchecked classes ?
- unclassified object (ie object not in any category or range class and not falling in a Else Filter)
- fully transparent ?
- not visible because of a blending mode effect ?
- classified but set with "no symbol" style
For unchecked classes, that would make sense not to have a snapping index. For the others, we'll need to discuss this largely with the community.
I tested with QGIS 2.99 3 week ago. The bug is still not fixed. I talk about invisible feature that are unchecked classes . The snapping options check only if the layer are checked. Perhaps, it's not a bug, but a new feature request .
Ok, this case is probably an issue. Let's raise the issue on the mailing list.
thread opened here http://osgeo-org.1560.x6.nabble.com/issue-with-snapping-on-unchecked-legend-classes-td5322685.html
For Lausanne, this issue is important. In the priority, it should be fix for QGIS 3.0 , ideally for QGIS 2.18 LTR if it's still possible. I don't understand the use cases where you would snap the new feature to invisible features on the map ...
ticket created upstream https://issues.qgis.org/issues/16838
My guess is that the whole snapping workflow deserves something consistent. We have quite a bunch of good issues and feature requests in that area
@dsavary backporting a change to 2.18 seems not possible according to @wonder-sk, it's a little bit more than only a bugfix. @qwat-psc: what is the priority of this issue when compared to all other issues with snapping? I have in mind the following needs:
- improve and fix the new snapping dialog (help to display the dock, intensive test ) #228
- add snapping priority and interactive selection (visual snapping - part 2)
- allow select all and deselect all in advanced snapping options
- layer finder filter (for projects with lots of layers)
IMHO, Snapping on invisible features is a feature.
Removing it completely would be wrong. I suggest to set it optionnal in the same way that snapping on intersections is optionnal.
Default could be set to visible only.
@wonder-sk Hi Martin, I have 2 days available in the end of 2017 budget to fix that in QGIS 3 or 3.2. Would you be available to work on that?
@wonder-sk Hi in case you had some time left, We are out of budget for 2017. Would you be interested in that for 2018 ?
Sorry for the delayed response... Sure we can do it in 2018 if you are interested in this feature.
@wonder-sk well, let's keep that in standby that until I get a clearance for the 2018 actions. Any time estimate on that task?
We have an external contributor funding that fix. Work starting soon, thanks!
Planned for 3.2. 3.1 is the current master development version.
2.18 was investigated by Martin and needed API changes >> NoGO
@lbartoletti hub cleanup, I assign you to the task you are already working on
The main bug was fixed one year ago. I fixed a related bug for tracing.
FYI, there is a last related bug https://github.com/qgis/QGIS/issues/26733
Do you want me to take care of it?
Do you want me to take care of it?
are you talking to me? ;-) Please go ahaed and close qgis/QGIS#26733 and https://github.com/qwat/QWAT/issues/193 if everything is OK: