These are chat archives for openseadragon/openseadragon

21st
Jan 2015
manuels
@manuels
Jan 21 2015 09:05
Hi, I´m wondering if it's possible to overlay several tiles that are partially transparent
Ian Gilman
@iangilman
Jan 21 2015 17:26
@ckuetbach @msalsbery #574 is merged.
@manuels what do you mean by tiles?
You can have whatever transparency you want on overlays.
Mark Salsbery
@msalsbery
Jan 21 2015 17:31
@iangilman Thanks! On the others I knew there would be small conflicts (that’s why I often end up doing more than one fix in a patch -everything’s so inter-related)... 576 can be done last…I reckon I’ll make the fixes and as each one gets merged I’ll merge master back into the next one, fix the conflicts, rinse, repeat?
Christian Kütbach
@ckuetbach
Jan 21 2015 17:33
@iangilman thanks a lot. I will make a pullrequest soon, to enable some kind of IE Versiondetection fallback for embedded IE browser (There will be no sideeffect, because it only tries another regex, if no version could be detected)
Ian Gilman
@iangilman
Jan 21 2015 17:33
@msalsbery works for me if it works for you!
@ckuetbach awesome, thank you!
Christian Kütbach
@ckuetbach
Jan 21 2015 17:35
@iangilman Is #508 still an issue in 1.2.x? As far as I remember, you mentioned, that the cleanup and reference strip code will be rewritten for 1.2
Ian Gilman
@iangilman
Jan 21 2015 17:43
Looks like it still is. Also, the planned cleanup was for #446, which will probably land after 1.2.1. I've updated #508 with some additional info.
Mark Salsbery
@msalsbery
Jan 21 2015 17:53
@iangilman I think 575 is ready, as soon as Travis CI gets around to it...
Ian Gilman
@iangilman
Jan 21 2015 17:53
Excellent.
Mark Salsbery
@msalsbery
Jan 21 2015 17:55
I need to revisit the focus on click issue...makes the viewer jump to the top of the page in the site-build sample pages
Ian Gilman
@iangilman
Jan 21 2015 17:55
Weird.
#575 is merged!
Mark Salsbery
@msalsbery
Jan 21 2015 19:20
@iangilman Awesome, thanks! Merged master into 576 and resolved conflicts - that one's ready as well
Mark Salsbery
@msalsbery
Jan 21 2015 19:29
@iangilman We're explicitly scrolling the viewer canvas to the top of the page in viewer.js onCanvasFocus() (around line 2193). I'm not sure why that code never got called before, but now that setting focus works it gets called, and not to good effect. Do we want that behavior?
Mark Salsbery
@msalsbery
Jan 21 2015 19:35
I guess to scroll the viewer into view when tabbing through a page. But that's bad when clicking on a viewer...
Mark Salsbery
@msalsbery
Jan 21 2015 19:53
@iangilman I found it redundant and removed it. New PR #577. All browsers I tested handle scrolling the viewer into view when it's tabbed to.
@iangilman Now I'm thinking hard-wiring the tabIndex to 0 is bad...needs to be an option in case the user is using tabIndex on other page elements. I'll add that to the PR.
Mark Salsbery
@msalsbery
Jan 21 2015 20:53
@iangilman Added tabIndex option to Viewer. PR #577 ready
Ian Gilman
@iangilman
Jan 21 2015 22:45
All sounds good!
@msalsbery remind me why we have tabIndex at all? Because otherwise you wouldn't be able to tab into it, right?
What happens if we have two OpenSeadragons on the page and they both have tabIndex = 0?
Mark Salsbery
@msalsbery
Jan 21 2015 22:48
Correct, because it’s a div which is an element that isn’t an automatic tab stop.
You can see the behavior on the site-build page for navigator - has lots of viewers. 0 is the default if tabIndex is not specified. Source order is used (the element order in the DOM)
Ian Gilman
@iangilman
Jan 21 2015 22:49
Cool. And according to your doc comment, looks like 0 is fine to have multiple of.
Mark Salsbery
@msalsbery
Jan 21 2015 22:50
Yes it’s the default
Ian Gilman
@iangilman
Jan 21 2015 22:50
Sounds good.
Mark Salsbery
@msalsbery
Jan 21 2015 22:50
default in HTML is what I’m referring to
Also the default proposed in the PR
With that PR the navigator page works well. You can tab top to bottom, and as each viewer gets focus you can keyboard navigate it
By the way, multiple elements with the same tabIndex behavior is undefined, but the general implementation is to use source order for matching indexes
Ian Gilman
@iangilman
Jan 21 2015 23:11
Cool.