93269 - Charset meta tag is ignored on view page source window when auto-detector is on

Status: VERIFIED FIXED

(Core :: Internationalization, defect)

Description

[ji]

*****Observed with linux 07/26 branch build*****

When auto-detector is on, the charset meta tag is ignored on view page source
window.

Steps to reproduce:
1. Turn on auto-detector on Japanese.
2. Go to http://home.netscape.com/zh/cn
3. View page source on this page.
   Simplified Chinese characters are not displayed correctly. 
4. Turn-off auto-detector, view this page source again, this time it's displayed
correctly.

Comment 1

[Matthias Versen [:Matti]]

*** Bug 93267 has been marked as a duplicate of this bug. ***

Comment 2

[Yuying Long]

Actually, I had a bug 92632 talking about same problem:
Auto-detect set to different language, the Page Source will shows in-correctly.
 However, when I filed that bug, I thought it was because the fix of bug 90288.
 Then I checked 2 builds (07-25 and 07-26/27), and auto-detect set to different
condition (07-25 OFF, and 07-27 with wrong language setting), so I thought it
was my mistake, then I closed it cause I think that we can not handle such kind
of case.

Xianglan, if Auto-detect to Chinese, then the Page Source will display fine, right?

I'm wondering do we consider this is a problem?

it's on all platforms.

Comment 3

[ji]

My point is charset meta tag should have higher priority than auto-detector on 
both navigator window and view page source window. 

Comment 4

[ji]

And stored charset for a page w/o charset meta tag should also have higher 
priority than auto-detector for view page source window. For example, I turn on 
auto-detector on Japanes, when I visit a Chinese page w/o meta tag, I correct 
the view by selecting charset menu, then I view page source on this window, I 
should be able to view the page source window in the Chinese charset I set by 
selecting the menu, not in the charset sniffed by auto-detector.

Comment 5

[Shanjian Li]

This is a problem. I am investigating now. 

Comment 6

[Shanjian Li]

Attachment: proposed patch

Comment 7

[Shanjian Li]

Harish, this problem is related with your change in parser. Meta charset might 
be the only exception that need to be handled correctly in view source. I think 
it should be ok to do this special handling inside meta charset observer. Please 
review my code and make sure it is consistent with your change in parser. If 
everything looks OK, please give a r=, thanks!

Comment 8

[Shanjian Li]

Attachment: revised patch after talked with harish

Comment 9

[harishd]

+  while(IS_SPACE_CHARS(httpEquivValue[lastIdx]) && lastIdx > 0) 
+    lastIdx--;


This will still fail when out of bounds. Should be

+  while(lastIdx > 0 && IS_SPACE_CHARS(httpEquivValue[lastIdx])) {
+    lastIdx--;
+  }

Comment 10

[harishd]

Actually, it should be 

+  while(lastIdx >= 0 && IS_SPACE_CHARS(httpEquivValue[lastIdx])) {
+    lastIdx--;
+  }

Comment 11

[Shanjian Li]

Attachment: one more update to my patch

Comment 12

[Shanjian Li]

Harish, I rechecked my patch and I found out that it is not really necessary to 
check the last quote character. So I update my patch one more time. Could you 
take a look? thanks!

Comment 13

[harishd]

r=harishd for patch id=45255.

Comment 14

[Shanjian Li]

blizzard, can you sr this patch?

Comment 15

[Christopher Blizzard (:blizzard)]

sr=blizzard

Comment 16

[Shanjian Li]

fix checked in. 

Comment 17

[Yuying Long]

Fixed verified on 09-04 trunk build on Win2k-CN and Linux RH6.2-Ja.