167237 - assuming UTF-8 instead of Latin1?

Status: RESOLVED DUPLICATE of bug 165908

(Core :: Layout, defect)

Description

[Jamie Zawinski]

Mozilla 1.1b, Red Hat Linux 7.2

The text of http://www.ocweekly.com/ink/03/01/cover-moxley.php contains a lot of
question marks, e.g., 

    "Today�s vicious attacks were clearly intended

That question mark is actually \222, the CP-1252 "right single quote" character.
Mozilla should have presented it as ' (apostrophe, the closest equivalent in my
Latin1 fonts.)

See also http://www.jwz.org/doc/charsets.html (which you also don't display
properly.)

I'm guessing that the root cause here is that when I do "Page info" on either of
those pages, it says "Encoding: UTF-8."  This is clearly not true, but the pages
do not specify a charset (as nearly nothing on the web does.)

The bugs:

  - The default charset, in absence of claims to the contrary via a Charset
    header or DTD reference, should be a Latin1 superset, as it had been in
    the past;

  - For flexibility and tolerance of the web as it actually exists, that
    superset should be CP-1252;

  - When displaying CP-1252 characters that are not part of Latin1, when using
    Latin1 fonts they should be converted to the closest glyphs available in 
    those fonts.

Preferences / Languages says (default settings):

   Languages for web pages:
     English/United States [en-us]
     English [en]

   Default Character Coding:
     Western (ISO-8859-1)

This is a regression: Mozilla 1.0 treated the default charset as Latin1, as
evidenced by bug bug 156200 -- I can't even see the (still un-fixed) error
described by that bug because now it's not even attempting any character
subsitution at all.

Comment 1

[Henri Sivonen (:hsivonen)]

Duplicate.

*** This bug has been marked as a duplicate of 165908 ***