80-Bus News

save it to disk instead. You’ve now got two files on the disk, the CCPZ and a CP/M system of matching size. Now use DDT or ZSID, or some such to merge the two files together putting CCPZ over the top of the original CCP and resave the the file. Finally use SYSGEN to put that file onto the system tracks. Bash RESET, and providing all is Ok, it should all fire up.”

This philosophy is fine provided you don’t change system size often. If you do, then all you need do is to keep a copy of the merged CCPZ and the CP/M system assembled for your favourite size on a disk, and when you change size just generate a CP/M to the system tracks using MOVCPM in the normal way. Ok, so you lose the features of CCPZ. When you’ve done, simply SYSGEN the CCPZ system image back onto the system tracks and you’re all set up again. (The version of CCPZ supplied with the Gemini Winnie has a relocator incorporated and whole thing is merged into MOVCPM. You can then use MOVCPM properly. Pity they don’t release this as a program, but I suspect that permutation problems would win in the end.)

Well this was all detailed in the installation documentation, and with the help of the dummy run, fairly easily understood. Finding the start of the CCP in an active CP/M of the chosen size proved a little difficult because the little program that should have been supplied to give you the address wasn’t there. So I thought about this for a while. My first thought was to use GEMDEBUG (I like ‘BUG’ over the others because I find it fairly easy to use even though it’s nothing like as powerful as ZSID) to find out where the CCP normally resided, but that was no use, as ‘BUG’ overlays the CCP and alters its size, and hence its jump location. At least I suppose this to be case with ‘BUG’, its’s certainly true of ZSID and DDT, I couldn’t be bothered to phone David to ask him. So using a debugger was out, I could calculate it, but as I only had a hazy idea of what I was about, this didn’t sound too clever, as it would introduce one more variable into the system to cause it to crash. Then I remembered STATUS, a little file I have rarely used because I thought I knew where everything was on my computer, apart from that, STATUS tests all the ports to see if something happens, as soon as it hits 0B3H, of course the video card resets, so you have to be very quick with a ^S if you hope to see the information prior to the video card reset. Anyway, I ran STATUS, and on the second try I caught it before it got to port 0B3H and, LO, there was my CCP start address.

Armed with the start address of the CCP, I had to modify the ORG statement of the source file. Easy? Not so. The source is 52K long and my version of PEN will only hold about 50K. I could have used WORDSTAR I suppose, but I don’t like it, it always wants to fight back at me. I always find it so much trouble to find out what the commands are, I find its layout of five sub-menus illogical and the commands I want to know about always seem to be scattered about in the five sub-menus and never in one. Further, there is no index to the sub-menus, so I have to work through them in order to find things. Enough slagging WORDSTAR, I know there are people out there who like it, and it is capable of jobs that PEN could never do (like writing books). Personally I find it all a bit of a pain, so enough personal bias from me. Anyway, horror of horrors, I resorted to ED!!! Now if there was something that should have been strangled at birth ....... With the aid of the CCPZ documentation, I patched the ORG statement and a couple of other things in the source, like “Do I want the prompt on USER 0 to be displayed as A> or A0>”, and things like that, and wrote the lot out to disk.

DRI MAC to the fore, I typed the assembly parameters exactly as in the demo (I have no idea what they meant) and let it get on with it. I ended up with a CCPZ.HEX file and a CCPZ.REL file. Ok, take the .REL file and LINK it, to make it appear as a true code file, but the linker wasn’t having any. Apart from that it was departure from the CCPZ instructions, but I thought I’d be clever and do it my way. Alright, don’t give up, get LOAD out of it’s dark drawer, blow the moths out of it, and see if that does the trick on the .HEX file. After all it’s supposed to take a .HEX file and convert it into a .COM file. Of course, LOAD