[webkit-reviews] review canceled: [Bug 28832] <img>.complete should return false for invalid images : [Attachment 159380] Patch

[bugzilla-daemon at webkit.org]

Jussi Kukkonen <jussi.kukkonen at intel.com> has canceled Jussi Kukkonen
<jussi.kukkonen at intel.com>'s request for review:
Bug 28832: <img>.complete should return false for invalid images
https://bugs.webkit.org/show_bug.cgi?id=28832

Attachment 159380: Patch
https://bugs.webkit.org/attachment.cgi?id=159380&action=review

------- Additional Comments from Jussi Kukkonen <jussi.kukkonen at intel.com>
Removing r?, this is not complete. Patch is still a work in progress, but if I
could get a confirmation from someone more experienced (that this is a good
direction to go and has a chance of being accepted) that would be awesome...

The patch does this:
 - Return true if no src or it is empty string
 - Return false if image is not loaded yet or is not decodable
 - Otherwise return true (image is loaded and decoded _or_
   image is loaded but no decode has been attempted)

The gotcha is that the completeness state now changes more over the lifetime of
the image. So the worst case scenario would be an image that is first assigned
a source url with a borken image and is then set visible.
1. image has no src -> complete=true
2. script sets src, but is not fetched yet -> complete=false
3. image is loaded (but is still not visible) -> complete=true
4. script sets image visible, decoding fails ->
 complete = false

Some things I already know to be missing:
* missing support for some graphics ports
* tests need to be updated