jeffcalc wrote:Curious , isn't it?
Bonjour Jean-Francois !
Once more you show me something I did not know yet. Again and again.
Yes, it certainly
is remarkable. My first test was on a CV with ROMs HFF, so all was fine. But it seems there is a quirk in the 41CX when looking for the external function shown as
XROM pp,qq when corresponding module is replaced by one with same ROM-ID but less functions. The query for the location of the module starts with page 5 (see RMAD10 @0810), after F the CX looks then at page 3:
- Code: Select all
0810 330 RDROM
F000 000 Data 0
0811 366 ?A#C X
0812 02B JNC +05 0817
0813 222 C=C+1 PT
0814 3E3 JNC -04 0810 RMAD10
0815 399 *
0816 0C6 NCGO 31E6
31E6 0D0 LC 3
31E7 15C PT= 6
31E8 330 RDROM
3000 019 Data 25
31E9 366 ?A#C X
31EA 360 CRTN
2F54 0C3 JNC +18 2F6C XROMNF
I did no CPU trace yet, so with just this firmware trace it looks to me as if cmd @31E9 compares apples with pears, ROM-ID and function number are the same. That explains why this bug does not show with XROM 29,26. Alas I can't see yet why it only occurs when 82143A is present.
If my idea of the bug is correct, it should show also with other modules. Proof (41CX only):
- I plugged my VORANOGE which has a FAT of 3 entries only,
- entered in prgm mode the function RN,
- used RAMED (function of ZEN ROM) to modify it to A4,99 so it ought result in XROM 18,25
- after exit from RAMED I see INSREC -- Heureka!
Steps 2 and 3 avoid the need of looking for modules with same ROM-ID and one of them wth at least 25 function, entering its function 25 in a program and exchange the modules.
Another example: Assign CTYPE from PANAME ROM to a key of your choice, switch machine off, unplug PANAME, plug ZEN ROM, switch on, hit that key of first step (in user mode of course).
To fix the bug would take more investigation. (All tests shown here done with my
Nut firmware interpreter.)
Note: due to heavy subject drift I started
a new thread.