Closed Bug 55738 Opened 24 years ago Closed 15 years ago

Cyrillic is not decoded in "Authentication request" popup

Categories

(Core :: Networking, defect, P3)

x86
All
defect

Tracking

()

RESOLVED DUPLICATE of bug 41489

People

(Reporter: ezh, Unassigned)

References

Details

(Keywords: arch, intl)

Attachments

(4 files)

1. Go to the URL.
2. Login dialog appears "User name for xxx" in xxx must be cyrillic works, but 
they are displayed without right encoding.

NN4.75 displays fine.

2000100708-M18
I don't see the difference using 4.7 and 6.0.
I don't know that is the corrupted text or not. Please specify the expected result.
Attached image NN 4.45
Attached image NN 4.75
And the mozilla I see what are you seeing.
Cc to IQA people, since I cannot reproduce this.
May be because my system does not have Cyrillic support installed.
I tried Windows2000 with Cyrillic support but could not reproduce the problem (I
mean I cannot see it correctly even with 4.7).
Eugene, could you specify the system you tested? Also any pre-condition (e.g.
charset menu)?
i am not able to display the login dialog in neither 4.x nor nscp6 correctly ( i
do have russian fonts on my machine)
Which default page encoding do you have? Try to set it to win1251 or KOI-8.

I'll try to test it tomorrow (I have dunk a little bit vodka tonight :) ).
Could it be related to regioanl settings? Maybe 4.x uses the regional setted 
font (system) to display this, but Mozilla do it different?

I tried ie5 and it also displayed this right to me (on russian).

My regional settings are Russia, russian.
That could be a reason, I don't know much about font.
Erik, do you have any idea?

BTW, I tried again with default to set to Russian (1251 and KOI8-R) but could 
not see the dialog correctly with 4.7.
I was able to the dialog with 4.x (but not with 6.0) after I set a default
region to Russia on my Win2000.
Reassign to network group.
I got a help from ftang to debug this. The string is coming from the server.
In HTTP Authentication code, currently the code assumes that realm is ASCII 
(line 2175).
We want to know if the HTTP spec specifies a way to indicate a charset. If so, 
we may apply a charset conversion there.

nsHTTPChannel.cpp
2152         //TODO localize it!
2153         nsAutoString message; message.AssignWithConversion("Enter username 
for "); 
2154         // later on change to only show realm and then host's info. 
2155         // find the realm 
2156         PRBool foundRealm = PR_FALSE;
2157         const char* realm = authLine.GetBuffer() + authType.Length();
2158         while (realm && nsCRT::IsAsciiSpace(*realm)) realm++;
2159 
2160         if (nsCRT::strncasecmp(realm, "realm", 5) == 0)
2161         {
2162             realm += 5;
2163             while (realm && (nsCRT::IsAsciiSpace(*realm) || *realm == '='))
2164                 realm++;
2165 
2166             const char * end = realm;
2167             if (*realm == '"') {
2168                 realm++; end++;
2169                 while (end && *end != '"') end++;
2170             } else {
2171                 while (end && !nsCRT::IsAsciiSpace(*end) && *end != ',') 
end++;
2172             }
2173 
2174             if (realm != end) {
2175                 message.AppendWithConversion(realm, end - realm);
2176                 foundRealm = PR_TRUE;
2177 
2178                 // Remember this realm; we will set it as an attribute of 
the new channel.
2179                 authRealm.Assign(realm, end - realm);
2180             }
2181         }
Assignee: nhotta → gagan
->darin
Assignee: gagan → darin
Target Milestone: --- → Future
Here's a sample HTTP 401 response from the site (generated from TestProtocols):

Response headers:
	server: Domino-Go-Webserver/4.6.2.2
	date: Sat, 14 Oct 2000 01:52:41 GMT
	connection: keep-alive
	accept-ranges: bytes
	content-type: text/html; charset=IBM-850
	content-length: 210
	last-modified: Thu, 25 Nov 1999 10:20:39 GMT
	expires: Sat, 14 Oct 2000 01:52:41 GMT
	pragma: no-cache
	cache-control: no-cache
	www-authenticate: Basic realm="Дистанционное обучение"

So, it looks like we can do a charset conversion as specified in
the content-type header.
Status: NEW → ASSIGNED
correcting summary
Summary: Cyrillic words are not disolayed correct → Cyrillic words are not displayed correct
This can be releated:

When saving a file to a directory with a cyrillic symbols in the save progress
window I see the same problem.

In save dialog (it is from windows) all is OK, but when appears the save
progress then the location is not cuyrillic.

Can it be related to using the wrong system font encoding?
Currently the API for displaying the dialog does not provide a means of
setting the charset.  Hence, we need to revise the API to include such info.
added keyword arch.
Keywords: arch
Moved to component Networking:HTTP
Component: Internationalization → Networking: HTTP
Keywords: intl
Still seeing this bug with moz 2001061620
Same on XP, changing to all OS.
OS: Windows 98 → All
Keywords: mozilla1.2
Summary: Cyrillic words are not displayed correct → Cyrillic words are not displayed correctly
See bug 41489
So, is it a dup of bug 41489?
-> default owner
Assignee: darin → nobody
Status: ASSIGNED → NEW
Component: Networking: HTTP → Networking
QA Contact: teruko → networking
Target Milestone: Future → ---
Summary: Cyrillic words are not displayed correctly → Cyrillic is not decoded in "Authentication request" popup
blocking2.0: --- → ?
Blocks: 61681
I don't see anything here that's not a dupe of bug 41489.  Marking as dupe.
Status: NEW → RESOLVED
Closed: 15 years ago
Resolution: --- → DUPLICATE
No longer blocks: 61681
blocking2.0: ? → ---
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: