185753 - Special characters like ä, ö, ' etc. change to �. This does not happen all the time.

Status: RESOLVED INVALID

(Core :: Internationalization, defect)

Description

[Kasper Heikkilä]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.2.1) Gecko/20021130
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.2.1) Gecko/20021130

In the page at
http://firingsquad.gamers.com/hardware/all_in_wonder_9700_pro_review/page19.asp ,
special character " changes to �.


Reproducible: Sometimes

Steps to Reproduce:
1. Install Mozilla 1.2.1 / 1.1 to the Windows 2000 SP3, default installation
2. Use it some time and then special characters change to � character.
3.

Actual Results:  
Special characters change to � character.

Expected Results:  
Special characters should be ", ä, ö, etc.

Comment 1

[Karl Ove Hufthammer]

wfm

Comment 2

[Kasper Heikkilä]

It also works for me if I open new browser (not just navigator window). When I
use the same browser longer then special characters start to change �
characters. Example from the page source code before and after:

Even if you don’t wish to integrate your PC with your home theater, it’s hard
not to like the ALL-IN-WONDER 9700 PRO.

Even if you don�t wish to integrate your PC with your home theater, it�s hard
not to like the ALL-IN-WONDER 9700 PRO.

Same happens with multiple internet sites and pages.

Comment 3

[Yuying Long]

It works for me on 12-16 trunk build.

Reporter:
Could you please clear the Cache (Edit | Preferences | Advanced | Cache), and
try again?

Since the page that mentioned here:
http://cn.rd.yahoo.com/home/3link2/?http://cn.sports.yahoo.com/
is a w/o charset meta-tag page, it could be affected by the browser default
charset or Cache.

I think you can correct it by clear the Cache or manually select charset as
western charset, and next time, when you visit the same page again, then the
page will be displayed fine.

Comment 4

[Kasper Heikkilä]

I cleared the cache (memory and disk), all cookies, page history and location bar.  
After these changes special characters still change to �.

My default character coding is Western (ISO-8559-1) and in languages for web
pages I have [en-us] [en] and [fi].

Comment 5

[Yuying Long]

After you clear the Cache, can you try to exit the browser and relaunch again?
or manually select western charset to reset.

Change to i18n default engineer. Simon, feel free to reassign.

Btw, the url in comment #3 was wrong, the url that I meant was:
http://firingsquad.gamers.com/hardware/all_in_wonder_9700_pro_review/page19.asp

Comment 6

[Simon Montagu :smontagu]

Reassigning to ftang

Comment 7

[Kasper Heikkilä]

If I exit the browser and relaunch it then special characters are ok. But when I
use the browser for a couple of hours then this same thing happens again.

Comment 8

[Yuying Long]

Well, if you have never visited any page other than western charset, then I
believe the page will always display fine.  But if you happened visit a
non-western charset page, e.g. UTF-8 or whatever else, then Cache will remember,
this will help when you next time visit the same page.

The Cache will not affect to those pages w/ charset meta-tag.
Since the page that you mentioned:
http://firingsquad.gamers.com/hardware/all_in_wonder_9700_pro_review/page19.asp
is a w/o charset meta-tag page, so it will be affected by browser default
charset and Cache.

So you can either clear the Cache or manually select the correct charset or
maybe turn auto-detector ON, and visit that page again, then should be displayed
fine.

I think this is the expected behavior.  And this bug we can resolve as worksforme.

Comment 9

[Kasper Heikkilä]

I visited http://cn.sports.yahoo.com/
and after that
http://firingsquad.gamers.com/hardware/all_in_wonder_9700_pro_review/page19.asp
everything was still ok. If I find out in which page these special characters
start to change I will try if I can reproduce this behaviour. Clearing cache or
manually changing the correct charset does not help.

Comment 10

[Kasper Heikkilä]

This same "bug" is also in Mozilla 1.3a. Many users in Finland have this same
problem. With browsers like IE6, IE5, IE4, NS4.x this is not a problem. Only
thing that helps is to close and restart Mozilla. 

Comment 11

[Frank Tang]

this sound like a bad bug. Estimate debugging time 2-5 days.

Comment 12

[Kasper Heikkilä]

Attachment: Example from the page that shows special characters in the way they should be and at the same time with conversion to question marks.

Comment 13

[Frank Tang]

what a hack. I have not touch mozilla code for 2 years. I didn't read these bugs
for 2 years. And they are still there. Just close them as won't fix to clean up.

Comment 14

[Travis Chase]

Mass Reassign Please excuse the spam

Comment 15

[Travis Chase]

Mass Re-opening Bugs Frank Tang Closed on Wensday March 02 for no reason, all
the spam is his fault feel free to tar and feather him

Comment 16

[Travis Chase]

Reassigning Franks old bugs to Jungshik Shin for triage - Sorry for spam

Comment 17

[Jean-Marc Desperrier]

All the problematic pages referenced are windows-1252 encoded pages that don't
define a charset.

Finish users must frequently be accessing those pages using a link from a page
that is using a non iso-8859-1 character set, and the defined behaviour of
mozilla in that case is to reuse the character set of the previous page, leading
to problems with all characters that are not 7 bit ASCII, as described or shown
in the attachment images.

The finish users should use universal autodetect to alleviate the problem. I
believe the finish localization is already setting it on by default, and we have
a bug opened to do the same for all versions, but are still waiting for fixes
for the problems with universal autodetect.

When the page is opened from a bookmark, the charset used is the one memorized
inside the bookmark entry.

If it is proved Mozilla uses a non standard charset for reasons other than the
one listed above, then describe precisely when and open another bug to fix the
occurence.

Resolved as INVALID.