513541 - <img>.complete should return false for invalid images

Status: NEW

(Core :: DOM: Core & HTML, defect, P5)

Description

[:Gavin Sharp [email: [e-mail]]]

See http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2009-August/022452.html .

>> The "and it is a valid image" part seems to contradict the current 
>> behavior of Safari 3.0.2, Opera 9.22, Firefox 2.0 and Firefox trunk (all 
>> tested on Windows). These browser all give values of true for an invalid 
>> image's "complete" property (see attached testcase or 
>> http://people.mozilla.org/~gavin/test/img-complete.html ). Internet 
>> Explorer 7 does seem to return false.

> My findings match yours. I have left the spec as is, for compatibility
> with IE, and because it seems the most logical.

Comment 1

[:Gavin Sharp [email: [e-mail]]]

Also:

> It appears this behavior was explicitly chosen in Mozilla, in bug
> 190561 (https://bugzilla.mozilla.org/show_bug.cgi?id=190561). I think
> the arguments given in that bug might merit reconsideration; detection
> of image existence is currently possible by other means, and I'm not
> sure how much weight "the Rhino Book" should have in influencing
> current DOM specs.

Comment 2

[:Gavin Sharp [email: [e-mail]]]

Attachment: testcase

Expected: Image.complete is false
Actual (trunk): Image.complete is true

Comment 3

[:Gavin Sharp [email: [e-mail]]]

I also filed https://bugs.webkit.org/show_bug.cgi?id=28832 .

Comment 4

[Olli Pettay [:smaug][[e-mail]]]

(In reply to comment #0)
> See http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2009-August/022452.html
> > My findings match yours. I have left the spec as is, for compatibility
> > with IE, and because it seems the most logical.

Just making we're ok with Hixie's decision and not with https://bugzilla.mozilla.org/show_bug.cgi?id=190561 ?

Comment 5

[:Gavin Sharp [email: [e-mail]]]

From my message:

> It appears this behavior was explicitly chosen in Mozilla, in bug 190561 
> (https://bugzilla.mozilla.org/show_bug.cgi?id=190561). I think the 
> arguments given in that bug might merit reconsideration; detection of 
> image existence is currently possible by other means, and I'm not sure 
> how much weight "the Rhino Book" should have in influencing current DOM 
> specs.

Comment 6

[:Gavin Sharp [email: [e-mail]]]

(In reply to comment #5)
> > detection of image existence is currently possible by other means

Though I don't really remember what I was referring to, here... It seems to be the strongest argument for keeping our current behavior.

Comment 7

[:Gavin Sharp [email: [e-mail]]]

I was referring to detecting the image's effect on page layout:
http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2009-August/022481.html
http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2009-September/022589.html

Given that, there's no reason not to make this change as far as I can tell.

Comment 8

[linux.user.since.2002]

For clarification, is 'invalid' a reference to the HTML5 'Broken State'?  Less/More/Other?
See also bug 694839, which was filed specifically with regard to IMG.complete for Broken IMGs.

Thanks!

Comment 9

[Simon Pieters [:zcorpan]]

I think this bug should be RESOLVED INVALID given what the spec says and there is interop for this as far as I can tell.

http://software.hixie.ch/utilities/js/live-dom-viewer/saved/3064

Comment 10

[Firefox Bug Husbandry Bot]

https://bugzilla.mozilla.org/show_bug.cgi?id=1472046

Move all DOM bugs that haven't been updated in more than 3 years and has no one currently assigned to P5.

If you have questions, please contact :mdaly.

Comment 11

[Mitra Ardron]

I'm just noting that current behavior of img.complete is different from in Chrome. When src is empty it appears to mark the image as complete immediately, which is not the expected behavior if the intention is to set 'src' from some other code.

This caused a nasty incompatibility for us, since the resizing code fired based on complete=true and left us with a 0x0 box to put the image in !