47755 - TABLE ISSUE: width="100% vs width="100%"

Status: VERIFIED WONTFIX

(Core :: DOM: HTML Parser, defect, P3)

Description

[Alan S. Jones]

Mozilla build 200080408  M18 build on win95...

Some may say this bug is invalid, but I wanted to bring it up any way as I think
it is important we try and be flexable in quirks mode and when no doctype is given.


The last few days http://www.cnn.com has had this thing to get a daily e-mail
from Larry King.  Just what I have always wanted :)
However that section would not render properly.


We narrowed it down to with="100% vs width="100%", they left off the extra quote.
I know the quotes are important, but CNN's page does not have a doctype set so
we should try and be as flexable as possible.


In case CNN changes their page I am going attach a test case of the correct and
incorrect version of that small segment.

Comment 1

[Alan S. Jones]

Attachment: Test version width="100%

Comment 2

[Alan S. Jones]

Attachment: Correct Version

Comment 3

[Mats Palmgren (inactive)]

Adding testcase keyword, reassigning to Parser.

Comment 4

[Devoti Paolo]

see also bug 47744

Comment 5

[Hixie (not reading bugmail)]

highly recommend WONTFIX or INVALID.

Comment 6

[harishd]

Will debug this in FUTURE.

Comment 7

[rickg]

In fact, I'm marking this bug wontfix, because the context of the problem is 
indistinguishable from a legal case. 

[foo="this bar="width"] (vs)
[foo="this or that" bar="width"]

Comment 8

[Hixie (not reading bugmail)]

*** Bug 55581 has been marked as a duplicate of this bug. ***

Comment 9

[Patrick]

I don't get it: EVERY other browser is able to handle this and Mozilla wouldn't.
Please explain me!

Comment 10

[Hixie (not reading bugmail)]

Well, if you want to have a go at implementing it, I am sure we would welcome 
patches. If it's so easy, what are you waiting for?

Make sure that when you write this you can distinguish the author errors from 
the valid markup -- here's a simple test case:

1: The author missed out the double quotes after the word 'left':
  <div> Symbol: "<p align="left title="></p>">. </div>

2: The author made no error (this is a valid fragment of HTML):
  <div> Symbol: "<p align="left title="></p>">. </div>

Oh wait, they are the same. How silly of me.

Comment 11

[Patrick]

1) I never said it was easy.
2) In this project, I am a user-tester, not a developer.(It would take me much
too much time for me to find where to implement it first.)
3) I don't get your <quote>silly</quote> example.
4) We were talking about width="100%", I don't see it in your example.

Comment 12

[Hixie (not reading bugmail)]

> 1) I never said it was easy.

You said:

: I don't get it: EVERY other browser is able to handle this and Mozilla 
: wouldn't. Please explain me!

I assumed you were implying that it was easy. My bad, sorry.


> 2) In this project, I am a user-tester, not a developer. (It would take me 
> much too much time for me to find where to implement it first.)

Well, the developers have said that they cannot see a way to implement the logic
required to work around this uncommon case of bad markup. So unless _someone_
can spend the time to implement it, this will not get fixed.


> 3) I don't get your <quote>silly</quote> example.
> 4) We were talking about width="100%", I don't see it in your example.

|width| is just an attribute as far as our parser is concerned, just like
|title| or |align|. My example is quite simple: it demonstrates that there is
no way of telling when the missing quote (|"|) is an error or is intentional:

  <div> Symbol: "<p align="left  title="></p>">. </div>
  <div> Symbol: "<p align="left" title="></p>">. </div>

Both of these are completely valid examples. Imagine instead of |align| you
have your |width| attribute -- how do you know if the author intended to put
the |title=| text inside the |width| attribute or if it was supposed to be 
another attribute? (Remember that at the parser level you do not know anything
about what the |width| attribute actually is.)

Comment 13

[Patrick]

Sorry for being so long to reply to you but I'm like over busy these days.

Wether it has to be fixed or not is a delicate question to which I will not risk
to answer. I would only give the two possible scenarios I can imagine.

	- Mozilla has a strong market power and will soon impose itself on the browser
"market". All the webmasters will correct their pages to comply with the standards.
	- Mozilla will have some difficulties to impose itself and people won't like it
because the pages are not well displayed nor will the webmasters correct their
pages.

About the question if it is fixable or not, I would say "yes" because it was
done in preceeding generation browsers. This not true if the implementation is
radically different and keeps you from working this around. I think you might
ask the Netscape developers how they managed it.

Patrick

Comment 14

[Andreas M. "Clarence" Schneider]

Dupes: bug 40262, bug 44485.

Comment 15

[Andreas M. "Clarence" Schneider]

*** Bug 40262 has been marked as a duplicate of this bug. ***

Comment 16

[Blake Kaplan (:mrbkap) (inactive)]

*** Bug 285598 has been marked as a duplicate of this bug. ***