Bug 155059

Summary: Rendering slippy maps (i.e. tiled images) broken on retina/HiDPI displays
Product: WebKit Reporter: Michael Gratton <mike>
Component: WebKitGTKAssignee: Nobody <webkit-unassigned>
Status: RESOLVED DUPLICATE    
Severity: Normal CC: bugs-noreply, mcatanzaro, otaylor, yoon
Priority: P2    
Version: WebKit Local Build   
Hardware: Unspecified   
OS: Unspecified   
Attachments:
Description Flags
Screenshot of the problem none

Michael Gratton
Reported 2016-03-04 20:04:53 PST
Created attachment 273067 [details] Screenshot of the problem Rendering slippy maps (e.g. OpenStreetMap https://www.openstreetmap.org/) is broken in WebKitGTK+ 2.11.90 under Epiphany 2.19.91. Tiles are rendered with rouble the x,y space between them, see attachment. Maybe a JS issue? Wasn't a problem under WebKitGTK+ 2.10.x.
Attachments
Screenshot of the problem (238.72 KB, image/png)
2016-03-04 20:04 PST, Michael Gratton 		no flags
Details
View All Add attachment proposed patch, testcase, etc.
Michael Catanzaro
Comment 1 2016-03-05 05:27:51 PST
Does this break GNOME Maps?
Michael Gratton
Comment 2 2016-03-05 17:42:09 PST
(In reply to comment #1) > Does this break GNOME Maps? No it's working fine, but Maps uses libchamplain (which uses clutter), not webkitgtk, AFAIK.
Michael Catanzaro
Comment 3 2016-03-06 06:13:21 PST
Hm, cannot reproduce in trunk... can you say the exact distro package version (e.g. webkitgtk4-2.11.90-1.fc24) of WebKit that you are using?
Michael Gratton
Comment 4 2016-03-06 14:09:44 PST
Oh okay. This is using a .deb built using the same Ubuntu packaging config as for 2.10.x, but upgraded to build the 2.11.91 source. Looking at debian/control, it doesn't seem to have been built with any odd flags, and the patches applied look harmless enough. NB I checked OSM with MiniBrowser, and it's still a problem there, so not an issue with Ephy. Any suggestions? Would it be useful to see the build's cmakeconfig.h?
Michael Gratton
Comment 5 2016-03-06 14:13:44 PST
Oh, just worked out what it is - this is on a MacbookPro with a high DPI retina display and WindowScalingFactor set to 2. Sorry, should have thought of that first.
Michael Catanzaro
Comment 6 2016-03-06 15:31:18 PST
OK great, CCing the folks who worked on high DPI. Can you confirm again that this is indeed a regression from 2.10, it works in 2.10 on the same laptop?
Michael Gratton
Comment 7 2016-03-06 15:42:28 PST
Definitely a regression - downgrading the webkit packages only to 2.10 makes the problem go away. I should note that the problem also stops with 2.11 if WindowScalingFactor is set to 1, but then the desktop becomes unusable.
Michael Gratton
Comment 8 2016-03-19 23:41:51 PDT
This is no longer a problem in 2.11.92, so maybe was caused by the render tile limit? Closing, will re-open if I see it again.
Michael Catanzaro
Comment 9 2016-03-20 08:13:25 PDT
Probably, yes. *** This bug has been marked as a duplicate of bug 155426 ***
Note You need to log in before you can comment on or make changes to this bug.