[Israel.pm] web-based script
yuval at windax.com
Tue Aug 31 04:34:39 PDT 2004
This can all be done from the shell.
And as much as I like to write Perl -- I'd rather use already-written tools.
> [in the browser]
> * goto url x.y.z
> * login (username+password) 
This can be done with wget and/or lwp-request and/or curl.
> <new page>
> * select a radio button
> * select another radio button
> * hit submit
Can you emulate this by hacking the URL?
Otherwise you can use WWW::Mechanize (which means you will write code, but
> <wait for a download of a HUGE (over 100 mg) file>
> [in the shell prompt]
> * run a script on the downloaded file
> <new file produced>
> * copy the new file to some directory
Of course, this is basic.
> * synchronize the directory with its mirror on another server
> * run another script that compiles something out of the file
> <after a few hours of processing a new file is created>
> * sftp the file to server A
> * ssh to server C
> * ssh from C to server B
> * sftp from server B to server A and get the new file
> * chmod the file on server B as it should be
Rsync runs over SSH, and can be used with SSH keys.
That means you can sync the directory, run scripts, chmod, etc.
I wrote a backup script for my former company that backs things up over
SSH using rsync and also chmods stuff.
> Can you please give me skeleton code for any of the tasks?
> I see here:
> session to handle (pose as a browser)
> doing stuff locally (that's easy... I already know this)
> synchronization in batch mode (how?!)
> transfer of the file by a chain of ssh and sftp 
Let me know if my way of thinking is correct :)
More information about the Perl