Was there a conv buffer prob in 2.2.1a?

[Michael Cuff]

Ok, I see what's happening.
The string goes from utf-8 to codepage. The length of the codepage 
string is shoved into the cli struct. Then the string is converted to 
unicode. The resulting string is shorter than the length recorded since 
some the bytes represent multibyte chars.

I can work around that.
-Michael

Michael Cuff wrote:

> Seems the number of extra bytes matches the number of bytes in the share 
> name.
> In the example below there are 3 extra words (not 4 as stated) as there 
> are 3 characters in the share name: ku-ru-ma
> Longer share names produce more extra byte pairs......
> -Michael
> 
> Michael Cuff wrote:
> 
>> Urban, List,
>>
>> When using smbmount, I'm seeing too many characters in the share name 
>> [when viewing the network packet].
>>
>>
>> Setup:
>> Client - Linux 2.4.7 based, smbmount version 2.2.1a
>> Server - W2k server, and Japanese W2k Pro.
>>
>> Share: \\192.168.7.44\ and the Japanese word 'kuruma'
>> kuruma = UC - 30AF,30EB,30DE; 932 - 834e,838b,837d; UTF8 - 
>> e382af,e383ab,e3839e
>>
>> Smb.conf: coding system = utf8, client code page = 932, character set 
>> = iso8859-1
>>
>> The mount command receives the share name in UTF8 encoding.
>>
>> The result is tree connect failed: ERRDOS - 66
>> The packet shows:
>>
>> 00000070  00 00 5C 00 5C 00 31 00 39 00 32 00 2E 00 31 00 
>> ..\.\.1.9.2...1.
>> 00000080  36 00 38 00 2E 00 31 00 2E 00 31 00 32 00 37 00 
>> 6.8...1...1.2.7.
>> 00000090  5C 00 AF 30 EB 30 DE 30 00 00 19 5E 4B 44 A7 68 
>> \..0.0.0...^KD.h
>> 000000A0  3F 3F 3F 3F 3F 00                               ?????.
>>
>> The UTF8 to UC is good!!! But this stuff:00 19 5E 4B 44 A7 68 before 
>> the ????? puzzles me.
>>
>> When the mount is tried from another Linux machine the 4 extra words 
>> are different values.
>>
>>
>>
>> Thanks,
>> -Michael
>> Specifically, thank you Urban for pushing me in the right direction!
>>
>>
>>
>>
>