KUML

[Petric Frank]

Hello Heiko,

> I downloaded the file with Netscape running Windows choosing "Save link as"
> and selected the Linux machine (Samba) as the target directory. This way the
> archive was broken. Now I just run wget and everything was fine ... hmm.
> There seems to be some LF/CR conversion running - no other reason I can think
> of. But why ?!

That is a general problem. The web server must define the file like
tar.gz and tgz as binary files. So this was not done by this web server
administrator.
If netscape does not get such a identification it assumes the file type
'ascii'. And here it is dependent from the platform what happens. On
Win*** (and also OS/2) systems this CR/LF translation is invoked. On
*nix systems this translation will never be done (due not necessary).

Doing 'save link as' also works the way described above on win* systems.
There should be a 'save link as binary' menu item, as i proposed several
times to netscape.

You can test this assignment of the web server. If you explore the link
and get garbled text on the screen the file type was not set properly by
the web server (and you must use other tools like wget for download). If
you get a download box,  everythiong is ok.
Result: Every uploader should ensure that the file type will properly
set by the web server.

regards
  Petric