28 Jun
2015
28 Jun
'15
3:56 p.m.
Hi list,
I'm implementing my own CP/M-80 work-alike using the FAT16 file system.
So far it's working nicely but there's a curious problem with Microsoft
BASIC. MBASIC51.COM will load files fine, but MBASIC52.COM won't,
behaving like the files are 0 bytes long. It calls the open-file routine
and that's it.
I'm just curious about the differences between version 5.1 and 5.2. I'm
assuming I'm not returning correct information in the FCB, but I'm
unable to figure out what that might be.
Here's a video of it: https://www.youtube.com/watch?v=XT4dBda2CKw
Cheers,
Alexis K.
28 Jun
28 Jun
4:19 p.m.
On 06/28/2015 03:56 PM, Alexis Kotlowy wrote:
I'm implementing my own CP/M-80 work-alike using
the FAT16 file system.
So far it's working nicely but there's a curious problem with Microsoft
BASIC. MBASIC51.COM will load files fine, but MBASIC52.COM won't,
behaving like the files are 0 bytes long. It calls the open-file routine
and that's it.
I'm just curious about the differences between version 5.1 and 5.2. I'm
assuming I'm not returning correct information in the FCB, but I'm
unable to figure out what that might be.
How about somewhere that both binaries can be obtained? Then
individuals can try it with their own favorite CP/M emulator.
--Chuck
4:36 p.m.
On 29/06/2015 8:49 AM, Chuck Guzis wrote:
How about somewhere that both binaries can be
obtained? Then
individuals can try it with their own favorite CP/M emulator.
Yes, that would help. They're at
http://kaput.homeunix.org/~thrashbarg/mbasic.zip
These are off one of the disk images that comes with the SIMH Altair
emulator.
Alexis.
5:55 p.m.
Don't know what to say--I ran both MBASIC versions with the same files.
Both load and save just fine under 22NICE 1.44 (running under FreeDOS
on DOSEMU running on Ubuntu on an AMD64 CPU).
--Chuck
6:12 p.m.
On 29/06/2015 10:25 AM, Chuck Guzis wrote:
Don't know what to say--I ran both MBASIC versions
with the same files.
Both load and save just fine under 22NICE 1.44 (running under FreeDOS
on DOSEMU running on Ubuntu on an AMD64 CPU).
--Chuck
Yes, they work fine under existing systems. I'm trying to figure out why
it doesn't work on the system I'm implementing. I'm emulating the
current record data in the FCB so even if a program modifies the extent
or current record without a random access call, it still loads the
correct sector from the file. So far, it works with WordStar 3. I'm
still implementing write support.
I've compared my emulated FCB to the one generated by a CP/M system and
from what I can tell, the CR/EX/S2 data is the same.
I've had a look at
the CP/M source, I don't see anything unusual there.
Really, the only
thing that's different is the allocation data.
I'm following the CP/M-86 convention for converting between FAT file
system data and CP/M FCB's. It's documented here:
http://www.seasip.info/Cpm/dosplus_fat.html
Knowing me, it's something silly or trivial, but until I find it... :)
Cheers,
Alexis.
6:59 p.m.
On 06/28/2015 06:12 PM, Alexis Kotlowy wrote:
Yes, they work fine under existing systems. I'm trying to figure out why
it doesn't work on the system I'm implementing. I'm emulating the
current record data in the FCB so even if a program modifies the extent
or current record without a random access call, it still loads the
correct sector from the file. So far, it works with WordStar 3. I'm
still implementing write support.
I've compared my emulated FCB to the one generated by a CP/M system and
from what I can tell, the CR/EX/S2 data is the same. I've had a look at
the CP/M source, I don't see anything unusual there. Really, the only
thing that's different is the allocation data.
I'm following the CP/M-86 convention for converting between FAT file
system data and CP/M FCB's. It's documented here:
http://www.seasip.info/Cpm/dosplus_fat.html
Knowing me, it's something silly or trivial, but until I find it... :)
So what does your FCB look like when CLOSE is invoked by the program?
It was almost 30 years ago, but I recall that 22NICE translates FCB
actions in ordinary handle calls--with a heckuva lot of code translating
FCB actions into handle-type file actions. So probably of no particular
help. WordStar is largely CP/M 1.4 compatible, IIRC.
--Chucj
9:26 p.m.
On 06/28/2015 06:12 PM, Alexis Kotlowy wrote:
Knowing me, it's something silly or trivial, but until I find it... :)
Well it was something trivial, in version 5.2 they decided to use a
random access read to rewind the file to the start. I didn't have random
access implemented, but now I have it works. What I don't understand is
why it didn't pop up as an unimplemented function call (causing the
confusion).
Crisis averted.
Cheers,
Alexis K.
3246
days inactive
3247
days old
6 comments
2 participants
participants (2)
-
cclist＠sydex.com -
thrashbarg＠kaput.homeunix.org