88987 - should not support PLAINTEXT

Status: RESOLVED WONTFIX

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

Description

[Tim Altman]

Using Mozilla 0.9.2 on Windows 98 SE.

The XMP, LISTING, and PLAINTEXT tags are evaluated, though they are obsolete in HTML 4.01, even when using the Strict (not quirks) doctype switch (specifying 4.01 Transitional with the URL goes Strict, right?).  Additionally, Mozilla does not evaluate character entities (< and >) inside XMP and doesn't seem to have a problem when not using these.  I'll add a test case shortly.

Comment 1

[Tim Altman]

Attachment: Test case of Mozilla 0.9.2 behavior regarding obsolete XMP, LISTING, and PLAINTEXT

Comment 2

[Tim Altman]

Sorry, forgot to add this:

http://bugzilla.mozilla.org/show_bug.cgi?id=69792 may depend on this, as it talks about rendering XMP.

Comment 3

[Tim Altman]

Attachment: More simplified test case of problem (valid HTML 3.2, 'cept noted char. entities bug I'm demonstrating).

Comment 4

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

Just because they're obsolete doesn't mean we can't recognize them.  We'd like
to keep quirks mode / strtict mode differences to a minimum.  The goal of strict
mode is to be something that we *could* be applying to all pages if there hadn't
been bugs in previous browsers, and we wouldn't want to break old pages in such
a mode.  The goal of quirks mode is to make only the changes from strict mode
that are needed to display pages correctly.

We don't truly support DTDs for HTML.  That would require using an SGML parser,
which would break the web.  So marking WONTFIX, although one could also argue
that it's INVALID.

Comment 5

[Tim Altman]

If you are going to leave these in, then the issues with them need to be addressed, i.e. the problems noted in the test cases with character entities in XMP.  Additionally, all text in a web page following PLAINTEXT should be shown as text, not rendered as HTML.  This is not the current behavior in 0.9.2.

Opera 5.12 seems to have the desired rendering of the second test case.

Reopening bug.

Comment 6

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

Entities should not be expanded in XMP, LISTING, or PLAINTEXT.  So the only bug
here seems to be that we don't make PLAINTEXT take up the entire page.  However,
PLAINTEXT was obsolete even in HTML 2.0.  Quoting from
http://info.internet.isi.edu/in-notes/rfc/files/rfc1866.txt :

5.5.2.1. Example and Listing: XMP, LISTING

   The <XMP> and <LISTING> elements are similar to the <PRE> element,
   but they have a different syntax. Their content is declared as CDATA,
   which means that no markup except the end-tag open delimiter-in-
   context is recognized (see 9.6 "Delimiter Recognition" of [SGML]).

      NOTE - In a previous draft of the HTML specification, the syntax
      of <XMP> and <LISTING> elements allowed closing tags to be treated
      as data characters, as long as the tag name was not <XMP> or
      <LISTING>, respectively.

   Since CDATA declared content has a number of unfortunate interactions
   with processing techniques and tends to be used and implemented
   inconsistently, HTML documents should not contain <XMP> nor <LISTING>
   elements -- the <PRE> tag is more expressive and more consistently
   supported.

   The <LISTING> element should be rendered so that at least 132
   characters fit on a line. The <XMP> element should be rendered so
   that at least 80 characters fit on a line but is otherwise identical
   to the <LISTING> element.

      NOTE - In a previous draft, HTML included a <PLAINTEXT> element
      that is similar to the <LISTING> element, except that there is no
      closing tag: all characters after the <PLAINTEXT> start-tag are
      data.

Comment 7

[Tim Altman]

My apologies in presuming entities should be expanded--rereading my correspondence with Opera, the use PRE when given LISTING or XMP.

I'm not sure what you're saying about PLAINTEXT.  You say it's obsolete, but you don't say anything about how it should be handled.  It's obviously being handled incorrectly according to the RFC, as you pointed out.

Comment 8

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

Changing from "Obsolete XMP, LISTING, and PLAINTEXT recognized" to "should not
support PLAINTEXT" and moving to Parser.

This requires removing support from the parser and from html.css.  Do Nav 4.x
and MSIE do it "right"?  If so, we're probably better off just dropping support
for it since it was dropped before HTML 2.0.

Comment 9

[Tim Altman]

Netscape 4.77 and IE 4.77 support PLAINTEXT correctly, yes.  They both show the rest of the tags in the page as plaintext.

Comment 10

[harishd]

This bug has been marked "future" because the original netscape engineer working 
on this is over-burdened. If you feel this is an error, that you or another
known resource will be working on this bug,or if it blocks your work in some way 
-- please attach your concern to the bug for reconsideration -----

Comment 11

[Boris Zbarsky [:bzbarsky]]

Attachment: Patch to remove <plaintext> support

Comment 12

[Boris Zbarsky [:bzbarsky]]

I could probably hack things so that plaintext would actually "work" (that is,
just ignore its end tag).  Or we can remove this (per that patch).  Either is
fine with me, I guess....

Comment 13

[Boris Zbarsky [:bzbarsky]]

Attachment: Also fix fragment sink

Comment 14

[Boris Zbarsky [:bzbarsky]]

Attachment: Merged to tip

Comment 15

[Johnny Stenback (:jst)]

Comment on attachment 159120 [details] [diff] [review]
Merged to tip

r=jst

Comment 16

[Boris Zbarsky [:bzbarsky]]

Fixed for 1.8a5.

Comment 17

[Martijn Wargers (dead)]

So plaintext is not supported anymore?
Shouldn't these two rules also be removed, then?
http://lxr.mozilla.org/seamonkey/source/content/html/content/src/nsHTMLSpanElement.cpp#126
http://lxr.mozilla.org/seamonkey/source/content/html/content/src/nsHTMLSpanElement.cpp#138

Comment 18

[Boris Zbarsky [:bzbarsky]]

Attachment: Indeed

Comment 19

[Boris Zbarsky [:bzbarsky]]

Attachment: And attaching the right file...

Comment 20

[Johnny Stenback (:jst)]

Comment on attachment 162485 [details] [diff] [review]
And attaching the right file...

r+sr=jst

Comment 21

[Boris Zbarsky [:bzbarsky]]

Checked that patch in.

Comment 22

[Blake Kaplan (:mrbkap) (inactive)]

Reopening. Plaintext support was just checked back in for bug 287990...

Comment 23

[Boris Zbarsky [:bzbarsky]]

OK, this is apparently wontfix; we now have correct support for <plaintext>.