238409 - Regression in attributes parsing (cellpadding)

Status: RESOLVED FIXED

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

Description

[Wladimir Palant]

Build 20040311 parses <table cellpadding> as <table cellpadding="0">, the
previous behavior was to ignore the attribute (which resulted in something like
<table cellpadding="2">). This seems to be a regression from 232706, the patch
was checked in on Mar 3 18:06.

Comment 1

[Wladimir Palant]

Attachment: Testcase, the last table is rendered differently

I meant cellspacing, not cellpadding in my report (cellpadding has the same
parsing problem but it is 0 by default).

Comment 2

[Wladimir Palant]

Attachment: Proposed patch

Comment 3

[Johnny Stenback (:jst)]

Comment on attachment 144669 [details] [diff] [review]
Proposed patch

sr=jst, and requesting 1.7 approval for this regression.

Comment 4

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

This is a pretty bad regression. For many attributes we now treat an invalid
value as '0' whereas we used to ignore the attribute entierly. This affects both
how we render the element and the DOM.

The fix is very simple and should be very safe.

Comment 5

[Asa Dotzler [:asa]]

Comment on attachment 144669 [details] [diff] [review]
Proposed patch

a=asa (on behalf of drivers) for checkin to 1.7

Comment 6

[Johnny Stenback (:jst)]

Fix checked in.

Comment 7

[Wladimir Palant]

Attachment: mochitest compatible testcase

Comment 8

[Robert Sayre]

Checking in test_bug238409.html;
/cvsroot/mozilla/testing/mochitest/tests/test_bug238409.html,v  <--  test_bug238409.html
initial revision: 1.1
done