Date   : Wed, 14 Apr 1993 16:07:49 +0100 (BST)
From   : "I Stephenson" <ian@...>
Subject: Admin+Alpha software Release

First off - there have been a few problems with our mail over the last few  
days, but they seem to be fixed now.

OK now the main bussines:
I've placed an Alpha release of my shell and OS layer software on shiraz(os-uue  
- its uuencoded and arced[Gerben's progs for this are excellent]). I've  
mentioned this a few times over the last few weeks, but now it really exists,  
and I'd like to get some feedback/distribute the development load.

As stands the code is a bit of a mess, but it does work and provides stacking  
of executables, and redirection of io. There's a library of routines which load  
in at C000 (its compiled for 6502 co-processor, but I did get it running on the  
host several times during development). Theses are started up by *OS. This  
loads in the file "sh" and starts it running. From here you can execute other  
programs, and redirect their io (via > and <). This includes sub-shells, so  
shell scripts are supported (though currently shell variables and control flow  
aren't implemented).

The suplied programs are sh, ls (a shell script which calls *cat so can't be  
redirected), hello (prints hello world), cat (prints the contents of a SINGLE  
file), and echo (prints its arguments). These are just about enough to  
demonstrate that a fully working system would be usefull (which is the purpose  
of this release - don't think anyone is meant to actually use it!).

As discusses previously its pretty slow (as it pages to floppy), but RAM disc  
or hard disc support would make things a lot faster. Lots of work is needed (a  
complete rewrite of shell, and some sort of standard library interface), but  
I'd like to take this further. Hence this release - If anyones interested we  
can discuss what needs doing, and who can do it.


On a seperate but related note - is anyone using BCPL? I've got the source for  
a BCPL compiler, targeting to INTCODE if anyone is interested in trying it - it  
shoud be binary compatible with the Acorn system (probably the same code!).  
With a little work by someone ho understands this stuff we could have a PD BCPL  
compiler up and running. (I've also get the source to a simlarly interpreted  
pascal compiler if anyone wants a go at that...).

Ian