194478 - seeing squares containing FFFD between words

Status: RESOLVED INVALID

(Core :: Layout: Text and Fonts, defect)

Description

[Isam Bayazidi]

User-Agent:       Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.3b) Gecko/20030211
Build Identifier: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.3b) Gecko/20030211

I have compiled Mozilla with XFT support, I use Windows fonts in Mozilla in
Linux such as Arial, Courier New, and Tahoma to view pages. Occationly I run
into pages that show Boxes or squares between words, these squares contain the
UNICODE number FFFD

Reproducible: Always

Steps to Reproduce:
1. visit http://www.gnu.org/philosophy/philosophy.html#Speeches
2. Look at the Squares

Actual Results:  
You see squares containing FFFD

Expected Results:  
you should not see the squares

Comment 1

[Boris Zbarsky [:bzbarsky]]

xft, to blizzard.  This is a duplicate in any case....

Comment 2

[Christopher Blizzard (:blizzard)]

That means that there are characters on the page that couldn't be mapped into
unicode.  What locale are you using?

Boris, what do you mean DUPME?

Comment 3

[Boris Zbarsky [:bzbarsky]]

I mean "there is an older bug filed with exactly these symptoms and also only in
xft builds but I can't find the bug# right now".

Comment 4

[Isam Bayazidi]

Attachment: A screenshot of the faulty rendering

Comment 5

[Isam Bayazidi]

I am using en_US.UTF-8 .. I have to use UTF-8 locales because of Bug 188538 .. but I can  
see the squales even when I switch to the en_US locale .. it is not locale related.  

Comment 6

[Christopher Blizzard (:blizzard)]

That's generated when the unicode converters take the code off the page and turn
it into characters to render.  I suspect that for some reason the converters
can't convert some of the characters on the page.  Mozilla probably doesn't do
much with the .UTF-8 bits in the locale, maybe for local filename parsing, but
that's about it.

Comment 7

[Jungshik Shin]

Dupe or not. This is invalid. The page is apparently in ISO-8859-1, but the
reporter probably set the character coding (in View) to UTF-8. Latin letters
with diacritic marks in ISO-8859-1 are invalid UTF-8 sequences for which we use
U+FFFD (replacement character) to represent (as blizzard wrote). In the
following example, characters like é,â and è are replaced by U+FFFD. 

Audio recording in French of a speech,  l'éthique du système GNU/Linux et de la
communauté des logiciels libres, les tâches à accomplir et les risques à
envisager, given by Richard M. Stallman at CNIT à la Défense, Paris