[Webkit-unassigned] [Bug 42815] getBoundingClientRect Broken for SVG Elements

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=42815





--- Comment #22 from Cosmin Truta <ctruta at chromium.org>  2010-09-29 08:00:18 PST ---
(In reply to comment #21)
> Good work Cosmin.
> Note, that we've never tested the combination of the zoom property on SVG elements at all. I wasn't even aware that it actually works.
> 

> It's definately non-standard for SVG - what happens if you manually zoom into your standalone SVG file - does getBoundingClientRect still work? Does it already work?

It wasn't working in the past, but it does works now with this patch :-)
I was referring to this when I mentioned the off-by-one errors that occur under arbitrary zooming.

> We really need a svg specific test in svg/zoom/page. Can you come up with one?

I tried, but I couldn't. Could you please give me a few more hints?

The results of getBounding... do not show up refreshed until the browser window is explicitly reloaded. The same is happening when running svg/zoom/resources/testPageZoom.js: the pixel dump does show up zoomed, but the text output acts as if zooming did not happen. (It's harder to tell that now, after patching the source, since the coordinates do indeed stay the same; but even before, with the bug inside, testPageZoom.js alone was not sufficient to display the wrong coordinates after zooming.)

You can see this effect by loading the layout test from this patch, in an unpatched WebKit. Zoom it manually, and see that the updated (wrong) coordinates don't appear until you hit the reload button.

So I can't test this in DumpRT, unless I'm able to do somehow trigger a page reload from JavaScript. But I don't know how (or if) I could do that.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.