532872 - Umlauts in URI lost after hitting Enter in location bar

Status: RESOLVED DUPLICATE of bug 461304

(Firefox :: General, defect)

Description

[bselu42]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.1; de; rv:1.9.1.5) Gecko/20091102 Firefox/3.5.5 (.NET CLR 3.5.30729)
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.1; de; rv:1.9.1.5) Gecko/20091102 Firefox/3.5.5 (.NET CLR 3.5.30729)

When an URI parameter contains umlauts, e.g. in an input field, an URL like

http://my.place.de/test.html?input=%C3%B6%C3%A4%C3%BC%C3%9F

is transferred to the browser, UTF-8 encoded. This is translated by Firefox which displays then

http://my.place.de/test.html?input=öäüß

Hitting the reload button is fine, the parameter is still transferred as %C3%B6%C3%A4%C3%BC%C3%9F (verified using Firebug network window). But when I klick into the location bar and hit Enter, the encoding seems to change. Now

http://my.place.de/test.html?input=%F6%E4%FC%DF

is transferred to the server. That looks ISO-8859-1 encoded and seems wrong to me. If Firefox resolves %C3%A4%C3%BC%C3%9F to öäüß, it also must re-resolve it back to %C3%A4%C3%BC%C3%9F when submitting.

Reproducible: Always

Steps to Reproduce:
1. Open http://www.google.de/search?hl=de&source=hp&q=%C3%B6%C3%A4%C3%BC%C3%9F&btnG=Google-Suche&meta=

Notice the location bar shows http://www.google.de/search?hl=de&source=hp&q=öäüß&btnG=Google-Suche&meta=

2. Klick into the location bar. 
3. Hit Enter.
Actual Results:  
The URL transferred to the server is:
http://www.google.de/search?hl=de&source=hp&q=%F6%E4%FC%DF&btnG=Google-Suche&meta=


Expected Results:  
The URL transferred should be:
http://www.google.de/search?hl=de&source=hp&q=%C3%B6%C3%A4%C3%BC%C3%9F&btnG=Google-Suche&meta=


Google can handle this change in the encoding. But not all websites can.

Comment 1

[David Dahl :ddahl]

I am not an expert in char encoding, but I can confirm this is happening as reported. I cannot confirm the improper encoding of 'öäüß'.

Comment 2

[David Dahl :ddahl]

The above does appear to be the proper encoding: 

encodeURIComponent("öäüß");

%C3%B6%C3%A4%C3%BC%C3%9F

Comment 3

[bselu42]

Does that mean the bug report is qualified?

Comment 4

[Stefan]

(In reply to bselu42 from comment #0)
> Hitting the reload button is fine, the parameter is still transferred as
> %C3%B6%C3%A4%C3%BC%C3%9F (verified using Firebug network window). But when I
> klick into the location bar and hit Enter, the encoding seems to change. Now
> 
> http://my.place.de/test.html?input=%F6%E4%FC%DF
> 
> is transferred to the server. That looks ISO-8859-1 encoded and seems wrong
> to me. If Firefox resolves %C3%A4%C3%BC%C3%9F to öäüß, it also must
> re-resolve it back to %C3%A4%C3%BC%C3%9F when submitting.

I can reproduce this error with current Fx 3.6.23 Linux x86_64 and when running with LANG=de_DE@euro (which is ISO-8859-15 aka latin9). Running under LANG=de_DE.utf8 works OK. 

The current Fx 10 (mozilla-central)/latin9 is also affected. In addtion to 3.6.23 the error occurs when you select+copy the urlbar content and paste it into a new tab.

Comment 5

[Stefan]

Testcase:

https://render01.fontshop.com/fonts/font_rend.php?idt=f&id=175770&rbe=fsifr&rt=%C3%9C%E2%82%AC&rs=128&w=1900&bg=ffffff&fg=000000&tp=0.0

STR:

1. Start Fx with LANG=de_DE@euro
2. Open Testcase-URL
3. Press RETURN in the urlbar.

Expected: 2. and 3. 'Ü€'

Actual: 2. 'Ü€' 3. a rectangle character