placeholder for 0 size flash object

[al_9x]

fx 3.6.6, ns 1.9.9.98rc1

page: http://www.wimbledon.org/en_GB/scores/index.html
object: mip.swf

The div containing the object does not appear to be invisible. Shouldn't a min size placeholder be shown? According to the dom inspector, the placeholder is there but is inside the 0 size dummy object node. Shouldn't the object dummy be removed?

On a simple test page:

Code: Select all

<object width="0" height="0" type="application/x-shockwave-flash" data="http://url1.invalid/">
</object>

the placeholder is also invisible, it is in the dom with a 0 size. If the object size is increased (1x1, 2x2, 3x3) the placeholder starts becoming visible, but is still smaller then the min size. In the test page the dummy object node is gone.

[Giorgio Maone]

This is caused by the changes you recently asked for: now the "real" embedding, i.e. the one which was about to be loaded but was blocked, gets the placeholder.
In this case we've got an outer Active-X style <OBJECT>, which Firefox doesn't even consider as a loading candidate and therefore isn't tracked by NoScript, and an inner <EMBED> which is intercepted and replaced. Older NoScript just search for the outermost object ancestor and replaces it, since this idiom, which dates back to Netscape vs MSIE4 and is quickly disappearing thanks to SWFObject, was very common in the past. Anyway I'm tempted to restore the old "replace the outermost OBJECT ancestor" behavior and see if it can be merged with the new nested object management style.

[al_9x]

Are there not two separate issues here?

1. In the test page, objects are not nested but the placeholder is still 0 size. So even prior to the changes (.92) the placeholder is not visible.
2. the nested issue: an object chain is one logical object, so the placeholder replacing the whole thing (root object), regardless of which was actually blocked by NS, seems logical.

[Giorgio Maone]

Are there not two separate issues here?

Yes they are separated, but #1 is a non-issue:

In the test page, objects are not nested but the placeholder is still 0 size. So even prior to the changes (.92) the placeholder is not visible.

That's because the "minimum size" for placeholders is honored only if the document's surface is bigger than it. Otherwise the { (document's width - objects's left position), (document's height - object's top position) } is used.
Until you produce simplified test cases which contains exclusively embeddings, you'll keep hitting edge cases. Try with:

Code: Select all

<div style="width: 80px;height:80px"><object width="0" height="0" type="application/x-shockwave-flash" data="http://url1.invalid/"></object></div>

[al_9x]

That's because the "minimum size" for placeholders is honored only if the document's surface is bigger than it.

Is that not the case on the wimbledon page? I also tried it in .92, placeholder is not visible.

[Giorgio Maone]

Please check latest development build

[al_9x]

Please check latest development build

1) the root object is now being replaced (so that part is fixed)
2) but though the container div (red border) expands to accommodate the min placeholder height, the actual placeholder has wrong height. The width seems to be ok.

[Giorgio Maone]

I actually get the expected 32x32 placeholder. Do you observe that behavior consistently?

[al_9x]

I actually get the expected 32x32 placeholder. Do you observe that behavior consistently?

Yes, 100% consistent. xp32sp3, fx 3.6.6, ns 1.9.9.98rc2, new profile, apply to whitelisted, allow wimbledon.org

[Giorgio Maone]

Please check latest development build.

[al_9x]

Please check latest development build.

Fixed, I guess all the times that I tried it in rc2 it was outside the viewport.

[al_9x]

I think it's a related issue, the height of the pokkariStorage.swf placeholder on http://redlettermedia.blip.tv/ is less than min

Also something I noticed, the width of the clickable area (the A element) of the placeholder spans the width of the parent element. Can it be bound to just the placeholder?

[al_9x]

Another height problem: http://www.files.tdobson.net/mfs/video/20080501-rms.spx

[Giorgio Maone]

Putting in the backburner, thanks.