156824 - Charset can not be carried over after form submission

Status: VERIFIED FIXED

(Core :: Internationalization, defect)

Description

[Yuying Long]

Build: 07-10 1.0.1 branch build on all platforms

Steps:
Case one:
1. Create a new profile and launch browser.
2. Load a form page with HTML meta tag charset, I have a very simple example here:
http://babel/tests/browser/form/gb2312_formecho.html
- This page is with gb2312 meta charset, but there is no charset meta tag after
submission.
3. The charset is marked as gb2312 which is correct.
4. Input some Chinese characters.
- Note: If you reload page by doing Ctrl+R or View | Reload here, the input
character(s) won't go away.
5. Click submit.

Result:
Chinese charaters are displayed garbled in the result page, and charset is
marked as iso-8859-1.

6. Go back to the original form page, clear the Cache, reload the page, then
manually select gb2312.
7. Input some Chinese characters and submit it.

Result:
Still get display garbled characters with charset as iso-8859-1.

Case two:
1. Repeat step 1 to 4 in case one.
2. Manually select charset gb2312 instead of reload, then you will see the input
characters are gone.
3. Input some Chinese characters again, then click submit.

Result:
Now the Chinese characters are diplayed fine and the charset is marked as gb2312.

After fixed bug 144897, we don't have problem with carry over the charset info.
with sub-links, but still has problem in case one which is regression from N6.2x.

Comment 1

[Yuying Long]

reassign to Shanjian.

Comment 2

[Shanjian Li]

For some unknown reason, browser can always find a entry from bookmarks. Since
the bookmarks has a charset field, and most likely the entry is default to
iso-8859-1, it make the submitted page displayed as iso-8859-1. For other pages,
the charset pass over mechanism works fine. Need further investigation.

Comment 3

[Yuying Long]

I found this is a regression by bug 143579 or bug 144897.

And please ignore my case two result, because after restart browser, the result
will be same as case one.

There is another bug 162239 for real life example, and that bug is regression at
the same time. 

Nominate as nsbeta1.

Comment 4

[Yuying Long]

> I found this is a regression by bug 143579 or bug 144897.

I meant the regression is between 06-12 and 06-13.

Comment 5

[Shanjian Li]

this bug has been fixed.

Comment 6

[Yuying Long]

Yes, it has been fixed by bug 162239, mark as verified.