Hi All,
Following are the problems i am facing.Resolution for these would be highly commendable:
a)I am trying to transfer files from my local C: directory to unix server using sftp.
The problem is coming when files with accent characters are picked for transfer.These files are not transferred.Error message of could not read/open file comes.
I am using putty to transfer the files.These files are dynamically generated through a C#.net application.
I also tried using psftp utility of putty.But that is also not able to read the file.
I had changed the character set to UTF - 8 and font to OEM mode but that did not help either.
b)I tried creating a dummy file on unix box through putty using touch command and renamed it with some accent characters.When i did ls to see the file the name of the file was changed to some other letters.Probably putty had encoded them.
My problem is if i try any tool that gets the file transferred then also on the unix box the name of the file is getting changed as it is getting encoded.Also the fiole is not opening.
Please help on this.How can i transfer the file so that the name and content of the file remains the same and easily readable.
My suggestion would be to gzip or bzip (or any other compressing method which is also supported by the server to decompress) the file and then transfer. Compressing the file essentially encodes the problematic characters and there should not be any problem in sftp-ing it to the server.
Compressing the file wouldn't fix the file name...
Theoretically, the filename is just a sequence of non-null and non-forward-slash characters. Whether or not the name looks pretty on the other end, it should still be representable; UNIX filenames are even more permissive than Windows ones, it shouldn't choke on anything a Windows filename has to offer. Unless perhaps they're extremely long?
Something may be attempting to translate the name to fit the remote system. I'd try disabling any translation if you can.
What architecture and OS version are you using?
An HP traditionally uses roman8 char set. You would have to use ISO... (cant remember...) what language are you using to have accents? (I have SF (swiss-french))
The file is getting dynamicaaly generated through a .net application.The name of the file is kept as per the entry made by the users from the interface.
These all files have french names and hence have the accent character.
I am on Windows XP OS as of now.
I tried using ssh and filezilla.Files are getting transferred and the same is shown on the tool interface but when i actually go to the server through putty the file names are different as they ae beogn encoded and als the file is not opening.
Imho this is a system design error. By all means hold the user's reference to the file within the file, but don't let them decide filenames on your system. Apart from your current issues with cross-platform character sets, you lay yourself open to code injection.
If you want to maintain cross-platform compatibility, stick to alphanumeric filenames within the 7-bit ASCII character set. When Windows and FTP is involved, the MSDOS 8.3 format (not the contracted Windows version of 8.3 which includes a tilde character) is preferred.
i.e. Filenames like XXXXXXXX.XXX (i.e 8-dot-3).
Btew. UTF-8 makes the stuation worse because it uses more than one character to express one special character.