[Webkit-unassigned] [Bug 129565] Remove filterRes parameter from filters

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Tue Dec 8 09:24:58 PST 2015


https://bugs.webkit.org/show_bug.cgi?id=129565

Felix Hassert <felix.hassert+webkit at sevenval.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |felix.hassert+webkit at sevenv
                   |                            |al.com

--- Comment #2 from Felix Hassert <felix.hassert+webkit at sevenval.com> ---
I stumbled upon very similar problems with SVG Filters that are rendered in Safari in viewport pixels, disregarding device pixels on 'retina' devices. This was reported in Bug 124914 and in a somewhat related discussion in Bug 93471.

I am also using filterRes to get sharp images on retina images. I have a test-page attached.

The HTML is:

<img src="filtered.svg" width="400">

and filtered.svg has a width of 800px for clients with DPR=2. as soon as a filter is used in the SVG it appears blurry in the browser.

i can fix that by repeating the (desired) dimensions as filterRes in the svg:

<svg ....  width="800" height="531" ...>
.. 
  <filter filterRes="800 531" id="noop">
...

I don't know of any other workaround to make filters apply on a canvas regarding the dpr. If there is a better and more compliant way, please let me know.

However, as long as I need this workaround, I would like to discourage you to drop filterRes.

That said, I think that it is a bug in webkit, that the actual canvas size for filters is (presumably) derived from the targeted viewport area. Other browsers (Chrome and Firefox) that have already dropped filterRes support do not show this behavior. Please use the attached test-page svg-filter.tgz to verify it.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.webkit.org/pipermail/webkit-unassigned/attachments/20151208/74a8ac4f/attachment.html>


More information about the webkit-unassigned mailing list