356665 - If table element has margin styled, then the offsetLeft, offsetTop, offsetWidth and offsetHeight of this table element will be wrong.

Status: NEW

(Core :: Layout, defect)

Description

[Topper]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.1; zh-TW; rv:1.8.0.7) Gecko/20060909 Firefox/1.5.0.7
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.1; zh-TW; rv:1.8.0.7) Gecko/20060909 Firefox/1.5.0.7

If table element has margin styled, then we got
offsetWidth = correct value + margin-left + margin-right,
offsetHeight = correct value + margin-top + margin-bottom,
offsetLeft = correct value + margin-left,
offsetTop = correct value + margin-top.
The BoxObject of this table has same problems.


Reproducible: Always

Steps to Reproduce:
1. Create a TABLE element, with margin styled.
2. Check offsetLeft, offsetWidth, ...
3.

Actual Results:  
offsetWidth = correct value + margin-left + margin-right,
offsetHeight = correct value + margin-top + margin-bottom,
offsetLeft = correct value + margin-left,
offsetTop = correct value + margin-top.

Comment 1

[Topper]

Attachment: show bugs

Comment 2

[Dave]

Attachment: Shows erroneous values for table offset widths when the margin is set.

Alignment was effected through the following css style:
table.alignCenter
{
margin-left:auto;
margin-right:auto;
}
This method was chosen to pass W3C validation.  However the bug is there as well when you use the align=center attribute for the table

Comment 3

[John P Baker]

If bug 258255 was in the correct component then this would be a duplicate.

Comment 4

[John J. Barton]

This bug causes errors in Firebug inspect for tables with margin.
http://code.google.com/p/fbug/issues/detail?id=289

Comment 5

[John J. Barton]

I want to mark this 'firebug wanted' but the UI won't let me. So I am cc beltzner.

Comment 6

[John J. Barton]

Just tested this in FF3.0b4pre and the problem still exists but the offset is different.

Comment 7

[Justin Dolske [:Dolske]]

Could we get a testcase that reports a binary pass/fail result [instead of just a gob of numbers] for the dimensions being tested? That would help for verifying the bug, and eventual conversion to a mochitest for the test suite.

Comment 8

[John J. Barton]

Also probably causes:
 Issue 68:  	 Inspect outline located too much to the left or right of element
http://code.google.com/p/fbug/issues/detail?id=68
 Issue 123:  	 "position: absolute" on "body" element confuses Firebug
http://code.google.com/p/fbug/issues/detail?id=123
Moving up on firebug list.

Comment 9

[John J. Barton]

I'll try marking the firebug-p2 bugs as wanted for 1.9.2

Comment 10

[Henrik Skupin [:whimboo][⌚️UTC+2]]

Also happens on OS X. A good testcase can be found with attachment 337022 [details].

Comment 11

[Robert O'Callahan (:roc) (email my personal email if necessary)]

John, please use getBoundingClientRect/getClientRects in Firebug instead of the offset* properties.

Comment 12

[John J. Barton]

Ok thanks, I thought that these functions might be good alternatives but I didn't understand the details enough to know for sure. This will encourage me.

Comment 13

[Robert O'Callahan (:roc) (email my personal email if necessary)]

https://developer.mozilla.org/En/DOM:element.getBoundingClientRect
http://ejohn.org/blog/getboundingclientrect-is-awesome/

Comment 14

[John J. Barton]

Mike Radcliffe took up the challenge and used getBoundingClientRect to slay at least 6 Firebug bugs related to offsets.  (In Firebug 1.4a24).

Comment 15

[Martijn Wargers (dead)]

I guess bug 562005 is related to this bug, right?