822230 - Message won't show if in a folder with specific accented letters such as Úž(0xDA, 0x9E in windows-1252/windows-1250)

Status: NEW

(Thunderbird :: Message Reader UI, defect)

Description

[Martin Vendl]

Attachment: tb-error.png

User Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:17.0) Gecko/20100101 Firefox/17.0
Build ID: 20121128204232

Steps to reproduce:

1. Created folder with a name "Úžlabina" (case does matter)
2. Moved messages into it.


Actual results:

Thunderbird fails to display messages in this folder. Screen with yellow exclamation triangle appears with Try again button, but without any error message. Both in 3-pane view and standalone window.


Expected results:

It should display a message.

Comment 1

[Martin Vendl]

Interesting bonus facts:
- I wasn't able to find any other Czech accented letter combination other than "*Úž*" that causes that.
- Any additional accented letter will fix it (e.g. "Úžá" works fine).

Comment 2

[Martin Vendl]

Also it seems to work on IMAP, only reproducible on POP3.

Comment 3

[:aceman]

Yes, I can confirm this on POP3 (on Slovak Win XP, that should not have problems with these characters). In the filesystem the folder is created correctly (it is displayed fine in Explorer and also its .msf file).

There is no error in the Error console.

Comment 4

[:aceman]

However I was using en-US version of TB20.
Neither View source is working on the message:
The file view-source:mailbox:///D:/ApplicationData/thunderbird/Profiles/16c6hfdj.default/Mail/atlas.sk/ڞlabina?number=0 cannot be found. Please check the location and try again. The characters are shown wrongly, there is some internal encoding problem.

Comment 5

[:aceman]

Sorry, I mixed the comment.
This is the Tb error message I get in the View source window:
"The file view-source:mailbox:///D:/ApplicationData/thunderbird/Profiles/16c6hfdj.default/Mail/atlas.sk/ڞlabina?number=0 cannot be found. Please check the location and try again."

Comment 6

[WADA:World Anti-bad-Duping Agency]

(In reply to :aceman from comment #5)
> "The file
> view-source:mailbox:///D:/ApplicationData/thunderbird/Profiles/16c6hfdj.
> default/Mail/atlas.sk/ڞlabina?number=0 cannot be found. Please check the
> location and try again."

When the path string in this bug is Copy&Paste'ed to SeaMonkey's mail composition window, URLBar of SeaMonkey, Search field of www.google.com and so on, on Japanese MS Windows-XP(System charset=Shift_JIS, CP932, 7bits-ascii is similar to Windows-1252), following is pasted.
> D:/ApplicationData/thunderbird/Profiles/16c6hfdj.default/Mail/atlas.sk/%DA%9Elabina?number=0

Tb puts 8bit binary of 0xDA and 0x9E in the error message text.
Tb puts 8bit binary of 0xDA and 0x9E of non-unicode as-is(without escaping, not in UTF-8 which is used in ordinal URL of Browser/Firefox) in internal URL.
>                (English)                    (Czech)
>        windows-1252  iso-8859-1     windows-1250  iso-8859-2
> 0xDA : Ú == U+00DA   Ú == U+00DA    Ú == U+00DA   Ú == U+00DA
> 0x9E : ž == U+017E   Not defined    ž == U+017E   Not defined
>
> http://en.wikipedia.org/wiki/Windows-1252
> http://en.wikipedia.org/wiki/ISO/IEC_8859-1
> http://en.wikipedia.org/wiki/Windows-1250
> http://en.wikipedia.org/wiki/ISO/IEC_8859-2

(In reply to Martin Vendl from comment #1)
> Interesting bonus facts:
 - I wasn't able to find any other Czech accented letter combination other
> than "*Úž*" that causes that.
> - Any additional accented letter will fix it (e.g. "Úžá" works fine).

á is 0xE1.
>                (English)                    (Czech)
>        windows-1252  iso-8859-1     windows-1250  iso-8859-2
> 0xDA : Ú == U+00DA   Ú == U+00DA    Ú == U+00DA   Ú == U+00DA
> 0x9E : ž == U+017E   Not defined    ž == U+017E   Not defined
> 0xE1 : á == U+00E1   á == U+00E1    á == U+00E1   á == U+00E1

Phenomenon like following?
  Úžá => 0xDA9EE1 == corresponding UTF-8 binary exists.
  Úža => 0xDA9E61 == corresponding UTF-8 binary doesn't exist.
  If check as UTF-8 passes, Tb's process successfully reaches file.

(In reply to Martin Vendl from comment #2)
> Also it seems to work on IMAP, only reproducible on POP3.

When IMAP, Mbox name is passed in "Modified UTF-7" from server, and Tb uses the string of "Modified UTF-7" as local file name for .msf and offline-store file.
> http://en.wikipedia.org/wiki/UTF-7
So, this bug won't occur if IMAP.