Good to hear that things are working (more or less) now.
Again, I'm not sure what the original question was, but if you created a folder with the old code page it may indeed still contain garbled characters (because the conversion never happened properly). If so, create the folder again, this time with correct settings.

FYI, for those interested :

As for MBCS and the limitations it causes, it is frustrating that Borland has never addressed this properly. For a unicode-internally application such as mine, it is sad to see Borland convert everything to MBCS, hand that to Windows (because VCL wraps around the MBCS APIs), and next Windows needs to converts to Unicode again.
For those interested I once put this to Team B in the newsgroups :

If you want your Borland app to display things properly for all languages on all systems you need to resort to third party Unicode components over the standard VCL components. What's stupid about this is that you need to completely replace all VCL objects and I personally haven't found the "moral strength" to start doing that in my apps :-)
E.g. : http://www.tntware.com/delphicontrols/unicode/
(This is NOT tested by me, but mentioned here just to complete the information on this topic)

Hope this helps.

Best Regards,
Peter Van Hove
CD and DVD Data recovery

----- Original Message ----- From: "Fastream Technologies" <[EMAIL PROTECTED]>
To: "ICS support mailing" <twsocket@elists.org>
Sent: Sunday, June 05, 2005 7:46 PM
Subject: Re: [twsocket] Folder names in THttpServer

Hello Peter,

Now I still see the folder name corrupted in DOS FTP but I can get into the
folder with the seen corrupted folder name. However filezilla works fine so
there should be no problem!

Thanks a LOT! :-))


----- Original Message ----- From: "Peter Van Hove" <[EMAIL PROTECTED]>
To: "ICS support mailing" <twsocket@elists.org>
Sent: Sunday, June 05, 2005 6:31 PM
Subject: Re: [twsocket] Folder names in THttpServer

I miss earlier mails of this thread (they were accidentally deleted) but
picking up on this one (I hope I'm not missing the point entirely)

Make sure you are not confusing Unicode with MBCS and SBCS.
Unicode is always 16 bits
MBCS (Multi Byte Character Set) as opposed to SBCS (Single Byte CS) can
contain, for certain characters, two (or more) bytes, as opposed to only one
byte for most other characters in that same character set.
I'm not sure if Turkish can have MB characters ?

If you're clear about that, also know that Borland VCL is limited in it's
capabilities to properly show non-latin characters on latin-set systems.
This is because VCL is completely MBCS inside, if it were Unicode we
wouldn't have all these problems.
So, suppose Win2K or XP, they are unicode based and convert MBCS via a
system defined code page to Unicode before displaying it anywhere.

If, for instance, your system's code page is set to "UK English" (as an
example) and you try to use foreign characters in a MBCS application (like
Borland builds them) then you will have problems seeing the correct
characters !
To work around that you have to tell XP that you want to use a different
code page (and next reboot your system)

To do this (XP) :

My Computer / Control panel / Date, Time, Language and regional settings /
Regional and Language Options / (third tab) Advanced
Language for non-unicode programs
See what language is specified there.  If not Turkish, try setting it to
Turkish and see if this fixes things (*)

(*) Without really knowing what the real issue is ;-))
Like said, I hope I'm not beside the point entirely

Best Regards,
Peter Van Hove
CD and DVD Data recovery


To unsubscribe or change your settings for TWSocket mailing list
please goto http://www.elists.org/mailman/listinfo/twsocket
Visit our website at http://www.overbyte.be

Reply via email to