1024329 - Absolute DIV positioning bug flickers/blinks fails to match desktop FireFox

Status: RESOLVED WORKSFORME

(Core :: Graphics: Layers, defect)

Description

[Richard Maher]

User Agent: Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/35.0.1916.114 Safari/537.36

Steps to reproduce:

go to https://hypocampus.application.uwa.edu.au 

The closest library will be highlighted with a throbbing (opacity of black overlay increases/decrease) This is UWA campus in Perth WA. Other librarys are also highlighted with InfoBubbles. 


Actual results:

The Marco Polo images renders halfway down the div and flickers.

BTW. If you try to get the Version info of FF from running it and going through settings/about FireFox then FF crashes on Kit Kat :-( App Manager is the only way to get version info.


Expected results:

Same as FireFox Desktop and Chrome and IE and Opera where the image is stable.

Comment 1

[Aaron Train [:aaronmt]]

Seems like the URL is unavailable

Comment 2

[Kartikaya Gupta (email:[e-mail])]

I also cannot access either of the two URLs. Closing as incomplete for now, please reopen if you have updated URLs (or even better, a reduced test case) that demonstrates the problem.

Comment 3

[Richard Maher]

Hi Aaron,

Sorry about that, I left off an S. Please try this URL: -

http://hypocampus.applications.uwa.edu.au/ffbug.html

Try your desktop FF first then Android to see the difference.

If you want to see a 2nd bug then drop the ffbug.html and that will go to the default page of civitas.html which has the mozilla App install code active. Android will flounder around and then ask you to allow someone other than Google Play install to the phone. When you have approved that then the Web App will install and be registered as an App but when you try to run it the PERMISSION_DENIED on the geolocation is all you can get. The manifest has the permission requested, the Android App Manager shows the permission, and the user is never promted for geolocation access.

Thanks again.

Cheers Richard Maher

Comment 4

[Kartikaya Gupta (email:[e-mail])]

When I load the new URL and allow sharing of location data the site invariably times out within a few seconds saying the location data request timed out. Do you have a reduced test case that just shows the DIV positioning bug without the geolocation request?

Comment 5

[Richard Maher]

Hi Kartikaya, I have fixed that issue and the geolocation watch should no longer error on timeout. I know it's a bit complex and a whole lot of noisy code that has nothing to do with the bug but the relevant bits are in  the ready() function of the Pulsator objects: -

myLoc.InfoWindow.setContent("<div unselectable='on' draggable='true' ondragstart='return stopEvent(event)'" + " class='wrapper' id='W" + myLoc.DivId + "'><div class='polo' id='" + 
myLoc.DivId + "'></div><div id='T" + myLoc.DivId + "' class='textHolder'>" + 
bubbleContent[messageId] + "</div></div>");

I'm just sticking some content in a Google Maps InfoWindow. The classes in play are "wrapper", "polo", and "textHolder".

Cheers Richard

Comment 6

[Kartikaya Gupta (email:[e-mail])]

I'm still seeing the timeout. Please attach a reduced test case.

Comment 7

[Richard Maher]

Ok, I will attach two files. One small html and one JPEG background image that illustrates the problem.

The SSCCE can also be accessed via http://hypocampus.applications.uwa.edu.au/blink.html 

This doesn't exactly replicate the behaviour I'm seeing but it's pretty close.

Comment 8

[Richard Maher]

Attachment: HTML CSS and Javascript needed to reproduce the problem

Comment 9

[Richard Maher]

Attachment: Background image I'm using

Comment 10

[Kartikaya Gupta (email:[e-mail])]

Thanks, that lets me reproduce the problem.

Comment 11

[Kartikaya Gupta (email:[e-mail])]

Attachment: Reduced test case

I reduced the test case a little further. Seems like a layerization problem of some sort.

Comment 12

[Richard Maher]

Seems to have been fixed in Nightly?

Comment 13

[Aaron Train [:aaronmt]]

Re-open please if this is reproducible again.

Comment 14

[Richard Maher]

Here's the crazy thing, the code works as expected if it is activated as an installed web app via nightly but if you go into nightly and run it via URL then the problem persists/is observable.

Comment 15

[Gregory Pappas [:gregp]]

Unable to reproduce