From: Frank da Cruz (fdc@columbia.edu)
Date: Wed Sep 27 2006 - 12:44:44 CST
> My solution: I edit my HTML sources on my PC, transfer them to
> the Unix server, and there I remove the coding signature (using
> the vi program). This works, though not very elegantly. Of
> course, the file transfer program should handle the coding
> signature (the same way it handles the different line feeds),
> but as long as I can't have that, I will work around this
> problem, at the point of its origin -- and nowhere else.
>
By the way, Kermit:
http://kermit.columbia.edu/ (The Kermit Project)
http://kermit.columbia.edu/k95.html (Kermit for Windows)
http://kermit.columbia.edu/ckermit.html (Kermit for Unix, etc)
when used as a file transfer agent, can do this. It converts from
various "legacy" character sets to UTF-8 (or vice-versa), and will
add a BOM if desired, or remove one if undesired. If the file to
be uploaded already is UTF-8, the same choices apply:
http://www.columbia.edu/kermit/ckermit70.html#x6.6
The pertinent command is SET FILE UCS BOM { ON, OFF }.
A few advances in this area were made in the next Kermit release,
8.0, most notably the character-set aware FTP client:
http://www.columbia.edu/kermit/ckermit80.html#x3
and automatic character-set detection for local files when
uploading:
http://www.columbia.edu/kermit/ckermit80.html#x4
Kermit can be used to transfer files over serial-port connections,
modem connections, Telnet/RSH/SSH connections (both clear-text and
secure), and FTP connections (ditto), and can also convert the
encoding of local files.
I'm not sure that the FTP client does anything with the BOM, but
if not this can easily be added since the code is already there
for Kermit protocol. If anybody is intrerested in trying this,
let me know; if any adjustments are necessary I'll make them.
The latest version (whose most notable new feature is support
for "large files") is here:
http://www.columbia.edu/kermit/ckdaily.html
- Frank
This archive was generated by hypermail 2.1.5 : Wed Sep 27 2006 - 12:46:26 CST