610935 - The display style problem of div insides a table

Status: RESOLVED FIXED

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

Description

[Yousheng Li [:yli]]

The html code like this:
<div style="width:300px;background:yellow;height:50px;">
    <table width="150%" cellspacing="0" cellpadding="0" border="0">
        <tr><td style="background:blue;">parent div float=left</td></tr>
    </table>
</div>

In Firefox, the max width of table is the parent div's width.

But in IE, Chrome, Safri and Opera, the width of table is larger than it's parent div.

This is a bug for Firefox?

Comment 1

[Brian Polidoro]

FWIW it works if you use CSS.

Comment 2

[Boris Zbarsky [:bzbarsky]]

Jonas, nsAttrValue::ParseSpecialIntValue has:

1077   if (aCanBePercent && (isPercent || tmp.RFindChar('%') >= 0)) {
1078     if (val > 100) {
1079       val = 100;
1080     }

Why, exactly?

Comment 3

[Boris Zbarsky [:bzbarsky]]

OK, this code used to live in nsHTMLValue, and before that in nsGenericHTMLElement.

The original blame is revision 1.1 of nsGenericHTMLElement.cpp:

  1.1 <[e-mail]> 1998-09-02 16:52
  Spanked

Not helpful.

Do any other browsers do this for any %-valued attributes?  If not, can we just nuke lines 1078-1080?


Comment 4

[Jonas Sicking (:sicking) No longer reading bugmail consistently]

Seems like i just carried that forward. If other browsers don't do it we should nuke it too I agree.

Comment 5

[Boris Zbarsky [:bzbarsky]]

Tested with image width too, and the behavior is the same; no weird clamping in other browsers.

I think we should remove the clamping.  Ms2ger, want to do the honors?

Comment 6

[:Ms2ger (he/him; ⌚ UTC+1/+2)]

Certainly.

Comment 7

[:Ms2ger (he/him; ⌚ UTC+1/+2)]

Attachment: Patch v1

Comment 8

[:Ms2ger (he/him; ⌚ UTC+1/+2)]

Attachment: Patch for checkin

Comment 9

[Dão Gottwald [:dao]]

http://hg.mozilla.org/mozilla-central/rev/35dfd6a0ace7