
                          The Sandman's Diskmag Shell
                               [REVISED VERSION]

 Hmm. Well. Yes. What can I say apart from "Oops!". Here's a new version of the 
 diskmag  shell  that was given away on the second Kelstar Atari  Club  program 
 disk.  This program file is a direct replacement of the program file SHELL.PRG 
 that was in the self-extracting archive. To use it, follow these steps:

 [1] delete the original program file SHELL.PRG
 [2] copy  the file MAGSHELL.PRG off this issue's disk onto the disk where  you 
     had SHELL.PRG.

 Here's what's different about this version:

 [] printer routines removed as they were faulty. I can't test them myself as I 
    don't have a printer
 [] fixed for different TOS versions with the new STOSFIX routines for  greater 
    compatability.
 [] no  help  screen or screensaver - what you see is what  you  get.  All  doc 
    displayer functions are listed at the bottom of the viewer screen.
 [] there is currently no Degas .PC1/PC2 support. Maybe later.

 The program runs with exactly the same .INF file, so it's a direct swap of the 
 two program files with no mucking about. I'm going to keep this .INF file set-
 up  for  all  future revisions of this shell,  so it'll be a  case  of  simply 
 copying the latest version over.



                                THE MISSING LINK

 In  the original docs for the shell,  I whinged on about the Missing Link  not 
 likimg any other extensions being used. Just today (26/April/1997), Bob kindly 
 passed on a message from Tony Greenwood (gulp!) that this was rubbish -  well, 
 Tony, I totally agree with you, but I did have my reasons for thinking this.

 About 3 weeks ago I did a new backup of my compiler disk,  and discovered that 
 there  were  some dodgy sectors on it,  and they were the  kind  that  weren't 
 totally knackered but accessed okay after 2 or 3  tries.  Unfortunately,  when 
 you're compiling in STOS,  the compiler only attempts one access then  refuses 
 to  go  on if it finds an error.  I didn't realise there  were  dodgy  sectors 
 because  the drive didn't clunk like it normally does when it encounters  them 
 in GEM, so I wrongly assumed incompatability somewhere in the code. The reason 
 I assumed the Missing Link was to blame was because the program compiled  okay 
 if  I removed certain other extension commands.  Now that I'm working  with  a 
 100% error free compiler disk, everything works fine, as it should.

 So  there you go.  The problem wasn't the Missing Link,  so apologies  to  all 
 concerned.  The problem was a cross between a STOS compiler bug and a hardware 
 bug,  with  the blame lying firmly on my shoulders for not realising what  was 
 going on - I admit I should have checked all possibilities first. Sorry!



