programming How N: was made relocatable

[tschak909]

I'm working on a video showing the process I took to make N: relocatable. Hopefully it will be of use to others.

 

Starting with, the needed mark-down of a listing, to find 3 byte addresses within the handler code to relocate.

[+David_P]

I seem to recall reading about someone building two versions, with origins at (say) $3000 and $4000, then running a program to identify all the difference, and using that dataset to build the install table.

[tschak909]

I will be showing a variation of this in the upcoming video, as well as doing it by hand.

 

[+David_P]

Doing it by hand is hardcore.

[tschak909]

Well, in this case, the ACTION! relocator has serious problems with some DOSes which don't keep the IOCB #1 open while loading binary load files (such as SpartaDOS), and the relocator itself is large, so ultimately N: uses a relocator that @flashjazzcat wrote as an example.

 

-Thom

[thorfdbg]

1 hour ago, tschak909 said:

Well, in this case, the ACTION! relocator has serious problems with some DOSes which don't keep the IOCB #1 open while loading binary load files (such as SpartaDOS), and the relocator itself is large, so ultimately N: uses a relocator that @flashjazzcat wrote as an example.

I wonder how this might work. If the DOS loader goes through the init vector ($2E2). If the IOCB is closed, the DOS would have to re-open it every time if the vector is called, and continue from the same offset in the file.