| T.R | Title | User | Personal Name
 | Date | Lines | 
|---|
| 222.1 |  | PARVAX::PFAU | Hacker for hire | Wed Mar 26 1986 19:30 | 5 | 
|  |     This reminds me of an old TECO macro that printed out the alphabet. The
    trick to the macro was that no character in the macro itself was
    repeated....
    
    tom_p
 | 
| 222.2 | How About "@"? | VAXUUM::DYER | Brewer - Patriot | Thu Mar 27 1986 00:22 | 5 | 
|  | 	    Have you tried this?
 		$ foo @SYS$SYDEVICE:[DECNET]DASTARDLY_FILE
			<_Jym_>
 | 
| 222.3 | How About ":"? | VAXUUM::DYER | Brewer - Patriot | Thu Mar 27 1986 00:30 | 5 | 
|  | 	    Have you tried this?
		$ foo :any nasty command you want
			<_Jym_>
 | 
| 222.4 | @, no, : maybe | SIERRA::OSMAN | and silos to fill before I feep, and silos to fill before I feep | Fri Mar 28 1986 14:47 | 19 | 
|  |     I experimented, without success with the
    
    		$ foo @something
    
    strategy.
    
    I appears that the "foo" symbol was originally defined as
    
    		$ foo == "$filespec"
    
    and that DCL doesn't execute the "@" on such a thing.  Instead,
    it merely runs the filespec and feeds the line "@something" to it.
    
    The ":" idea sounds promising, I'll let you know when I get results.
    
    /Eric
    
    p.s.	Has anyone figured out yet what system I'm hacking ??
    
 | 
| 222.5 |  | JON::MORONEY |  | Fri Mar 28 1986 15:24 | 4 | 
|  |     Yup.  You shouldn't pick on poor Mr. * with your hacking, with the
    service he's providing..
    
    -Mike
 | 
| 222.6 | . . . a.k.a. ____::__E_E_ | VAXUUM::DYER | Brewer - Patriot | Fri Mar 28 1986 16:22 | 2 | 
|  | 	    Yeah, leave E_ (____::|_____::) _e___e_____ alone!
			<_Jym_>
 | 
| 222.7 | I help, not hinder!  (: doesn't work) | SIERRA::OSMAN | and silos to fill before I feep, and silos to fill before I feep | Fri Mar 28 1986 16:44 | 18 | 
|  |     Actually, I've been helping him.  When I first found loopholes in
    his system, I told him about it by coercing his software to send
    him a message telling him what the hole was.
    
    Anyway, it looks like ":" isn't a loophole either.  I'm not sure
    why not.  I suspect the filter avoids executing
    
    	$ foo . . .
    
    if the first character after "foo" isn't alphabetic.
    
    It's too bad the filter has to check for specific things.
    
    As a meta question, is there a more elegant way to protect
    one's software than the current method of specifically checking
    for certain known-to-be-nasty ascii characters ?
    
    /Eric
 | 
| 222.8 | I'd Run It Through A Keyword Parser | VAXUUM::DYER | Brewer - Patriot | Mon Mar 31 1986 09:03 | 0 | 
| 222.9 | Coupla thoughts. | KLOV02::BROWN |  | Wed Apr 02 1986 11:58 | 21 | 
|  |     
    The best way to trap 'nasty' commands is to only execute specified
    commands within the command procedure. For example never do anything
    like :-
    $'Command
    Always do something like :-
    $If "''Command'".eqs."FOO" Then $Foo
    (Make sure FOO is defined within the command procedure as a local
    	symbol, and that the user has no chance to re-define it. To
    	play REALLY safe you can re-define the command FOO every time
    	just before invoking it)
    To pass parameters to a command use F$PARSE on the file names. For
    example (parameter P1 is a file name) :-
    $P1 = F$Parse(P1)
    $If P1.eqs."" Then $Goto ....
    $Command 'P1
                
    You can always give the user a menu of numbered commands and force
    	him to use a numbered command instead.
    
    
 | 
| 222.10 |  | KOALA::ROBINS | Scott A. Robins" | Wed Apr 02 1986 14:52 | 5 | 
|  |     re .9:
    If you want a 'secure' .com file, never use the "''command'"
    construction.  See the note earlier about 'in-use' procedures.
    
    Scott
 | 
| 222.11 | Secure command file is an oxymoron | ERIS::CALLAS | Jon Callas | Thu Apr 03 1986 13:41 | 4 | 
|  |     If you want a secure command file, convert it to an executable image
    (using TPU, your favorite compiler, and the linker) and install it.
    
    	Jon 
 | 
| 222.12 | cashe me if you can | PHENIX::SMITH | William P.N. Smith, CSM Components Eng. | Tue Jun 24 1986 15:15 | 7 | 
|  |     Umm, maybe I'm just being dense here, but I missed something.  What
    is the point of having a system remotely execute something you send
    it?  What can you do remotely that you can't do easier locally?
    Examples would be greatly appreciated...
    
    WPNS
    
 | 
| 222.13 | Is There A [GAMES] Directory? | VAXUUM::DYER | Banish Bigotry | Wed Jun 25 1986 00:18 | 2 | 
|  | 	    Look for goodies that aren't on your system, for starters.
			<_Jym_>
 | 
| 222.14 | check remote jobs is another reason | ROXIE::OSMAN | and silos to fill before I feep, and silos to fill before I feep | Wed Jun 25 1986 13:49 | 4 | 
|  | Another thing you may want is to look at the batch or print queues
on the remote system, to see how your PRINT/REMOTE job is humming.
/Eric
 | 
| 222.15 | ... | DSSDEV::ALDEN | Ken Alden | Wed Jun 25 1986 15:30 | 6 | 
|  |     Or compile a long bliss job when your local node is 730 and the
    remote node is an 8800.:-)
    
    
    
    
 | 
| 222.16 | :-) or :-( ? | SWIFT::PITT | Tony Pitt, UK CS, Basingstoke, England | Fri Jun 27 1986 10:46 | 3 | 
|  |     How about shutting down the network?
    
    T
 | 
| 222.17 | to access third-party software | REX::MINOW | Martin Minow, DECtalk Engineering | Mon Jun 30 1986 15:50 | 7 | 
|  | We have some third-party software that is only licensed for use on
a specific cpu.  We have people who want to use this software who
are located on a different cpu.  The easiest solution seems to be
to copy the files back and forth as needed.
Martin.
 |