#: 8520 S10/OS9/6809 (CoCo)
    01-Dec-90  17:59:35
Sb: #CoCo 3 Emulator?
Fm: John M Semler 74020,736
To: all
Would it be legal to create a Coco 3 hardware emulator that runs on Macintosh
hardware?   I am asking this because I would like to run OS9 Level II on my
MacIIcx to keep up with you guys!
Of course if I create such a thing It will be uploaded for free.
It would be interesting to see how much faster this emulated CoCo 3 would run
compared with the real thing <smile>.
John Semler
There is 1 Reply.

#: 8530 S10/OS9/6809 (CoCo)
    02-Dec-90  05:38:55
Sb: #8520-#CoCo 3 Emulator?
Fm: Kevin Darling (UG Pres) 76703,4227
To: John M Semler 74020,736 (X)
John - interesting thought!
One of the guys here has been working on a similar thing: it runs L-II/6809
software under OS9/68K.  It's written well enough that it could probably be
made to run under other systems also (altho I suspect that it would help a lot
if the other system was also a multitasking one).
But even tho this emulator is highly tuned, it's still tens of times slower (at
least) than a coco-3, even when run on an 030.  It's still quite useful, but it
points out that 6809 code is very optimized and powerful... emulating the 09
seems to be harder than emulating other, simpler processors.
There are 2 Replies.

#: 8534 S10/OS9/6809 (CoCo)
    02-Dec-90  11:06:06
Sb: #8530-#CoCo 3 Emulator?
Fm: John M Semler 74020,736
To: Kevin Darling (UG Pres) 76703,4227 (X)
> "... emulator ... tens of times slower ..."
I would of thought that I could at least break even with a 6809 emulator
running on a 16MHz 68030.  Case in point, here is an excerpt from page 30 of
the manual for SoftPC EGA/AT for 68030 Mac's (It runs MSDOG version 3.30):
"The absolute speed of SoftPC EGA/AT depends primarily on the speed of your
Macintosh.  On the slowest SoftPC-compatible Mac, the original Mac II, the
Norton Computing Index for SoftPC EGA/AT is about two and a half times the
speed of the original PC.
On the fastest Macs, the Computing Index is about the same as that of the
original AT, or about seven times the Index of the original PC."
Who is this other guy that implemented a LII emulator?
Is he willing to share his work with others for free?
Besides, I already got a cute Icon made up for the CoCo III emulator.  I just
need to add a CODE resource to it to complete my project <grin>.
There are 4 Replies.

#: 8543 S10/OS9/6809 (CoCo)
    02-Dec-90  15:39:30
Sb: #8534-CoCo 3 Emulator?
Fm: Kevin Darling (UG Pres) 76703,4227
To: John M Semler 74020,736 (X)
John -
Bob Santy wrote the emulator over the last year or so...  I believe that he'll
be selling it.  It includes source for the customizable I/O section, I think
(the places where I$ and SS.calls must be translated).
I have no idea what would be involved in a Mac port of it, altho we've talked
about his porting it to say, OS-9000. (He wrote it in C, but has def'd it so
that his extensive 68K asm quick routines can be compiled instead).
I've used it at strange times:  like when I didn't have a wordcount util under
68K.  So I just used my coco "wc" util under the emulator instead. Sure, it's
slower than a 68K util, but when you want to use something in a hurry without
writing a new one, you can't beat it <grin>.  He says the 6809 C compiler even
runs under it.
As for the SoftPC comparisons.... well, remember how we all would talk about
how a 2mhz 6809 is pretty fast at many things?  We weren't kidding! And
remember that the 6809 has addressing modes even a 680x0 cpus doesn't have! 
Plus you have two separate 8-bit regs (A and B) which sometimes are
concatenated into the one 16-bit reg (D). Not to mention that it sets condition
codes on almost all instructions, which really slows down emulations.
So faking a 6809 is _much_ harder than emulating early Intel cpus. A 680x0
emulating an AT is kinda like making an adult fake babytalk: no problem. A
680x0 emulating a 6809 is like making the same adult fake teenage slang:
similar language, but sophisticated in a different way, and harder to fake ;-).
PS: <big grin> on the CODE resource comment! hehe - kev

#: 8565 S10/OS9/6809 (CoCo)
    03-Dec-90  06:59:56
Sb: #8534-CoCo 3 Emulator?
Fm: Bob Santy 76417,714
To: John M Semler 74020,736 (X)
John:
I've tried my darndest to get the emulator to go faster (in pure 6809 CPU
terms) and I haven't figured out a way to do it and still have a working
program.  The emulator I have is a 6809 object code "interpreter" with an
OS9-6809 "emulator".  I use the terms individually to describe the two basic
parts of the utility because they are different conceptually.  The 6809 address
space sections (TEXT and DATA) is kind of "emulated" as well.
Here's a list of things to think about:
 o The 6809 is an 8 bit machine with no restrictions on the
   address of an object.  The 68000 is a 16 bit machine with byte
   and word object differentiation.  At any point in the
   interpreter, a memory reference to a word object can take
   place.  Therefore, all accesses to the emulated address space
   are byte wide, slowing it down.
 o As Kevin pointed out, the D register is a nasty.  I keep the A
   and B registers in global 68000 registers, but must glue them
   together every time D is referenced.
 o All 6809 registers are kept in global 68000 registers.  X and
   Y in data registers and PC, S and U in address registers.
   But, as mentioned above, stack pushes and pulls force the
   disassembly and re-assembly of the words.  The address
   registers have the added slow-down of having to be normalized
   to the 6809 64K space.
 o Perhaps the nastiest slow-down is the darned condition codes.
   The 6809 and 68000 condition codes are tantalizingly similar
   and annoyingly dis-similar.  Plus, the 6809 CC register must
   be faithfully emulated so the program dosen't fail on a
   branch.  I'm still trying to figure out how this can be
   streamlined.  For now, every instruction that can alter the CC
   register has to do it correctly according to the 6809 rules.
   Check out the differences in rotates and the lack of a 68000
   X-bit conditional branch.
 (continued)

#: 8566 S10/OS9/6809 (CoCo)
    03-Dec-90  07:00:45
Sb: #8534-#CoCo 3 Emulator?
Fm: Bob Santy 76417,714
To: John M Semler 74020,736 (X)
 (continued)
I should note that the "emulator" dosen't appear to be slow if a program does a
lot of I/O.  The operating system interface (especially the I/O stuff) runs
pure 68000 code and isn't too bad.  A port to the MAC should be fairly
straightforward assuming that the MAC can support the OS9 level II system calls
needed. The current version runs on OSK and most of the calls are one to one. 
I don't know the MAC, so I can't say how easy it would be. Start by comparing
your MAC's C runtime library with OS9 level II's system calls.
Any ideas on the above would be appreciated.
Bob Santy
There is 1 Reply.

#: 8583 S10/OS9/6809 (CoCo)
    04-Dec-90  03:41:53
Sb: #8566-#CoCo 3 Emulator?
Fm: John M Semler 74020,736
To: Bob Santy 76417,714 (X)
Bob -
Maybe I do have an idea...
Why couldn't we build an OS9 6809 object program loader that also creates an
equivalent 680X0 phantom executable object.  This peusdo interpreter would not
need to perform any 6809 opcode fetches since it "knows" where it is at and
what it is doing at all times.
What I am proposing is a fast 6809 to 680x0 object code translater/loader that
creates a highly optimized "peusdo" interpreter for each 6809 object program it
loads.  Each 6809 opcode would be compiled into a series of 680x0 instructions
obviating the need for an opcode fetch.
This model would break down for OS9 programs with self modify code. The OS9
Debug command would not be able to set breakpoints since this phantom is
created only at program load time.  Also programs that run amuck and destroys
its own 6809 object code space will not crash this phantom.
What do you think?
There is 1 Reply.

#: 8585 S10/OS9/6809 (CoCo)
    04-Dec-90  07:36:51
Sb: #8583-#CoCo 3 Emulator?
Fm: Bob Santy 76417,714
To: John M Semler 74020,736 (X)
John:
Such a "translator" is actually in the works here.  The problem to be solved is
not as simple as it appears.  In order to correctly translate 6809 object to
68000, the 6809 code must be interpreted and converted only when the actual
program function is verified. How many disassemblers (as good as they may be)
are actually guaranteed to produce 100% correct source?  This is my basic
problem.  I need to guarantee that the resultant 68000 code is faithful to the
original.  By the way, I can only generate 68000 code in USER mode to allow the
program to be used on a 68000.  I don't have a 68020 to test on and will leave
any optimizations available on the '20 (and '30) for later.
With the above in mind, it would be possible to generate a 68000 object that
was faithful to the original to the extent that the original was interpreted. 
One would "exercise" the 6809 object to take it through all possible paths. 
Some method would be available (a special trap) to inform the user of an
unexercised path in the 68000 executable.  Then the program would be extended
by running the interpreter again.  That's the current theory anyway.
Even with that kind of conversion working, there's still the problem of
byte/word boundries.  What if the 6809 program had a series of byte addressable
objects that were ALSO referenced by word operations as well?  This would be a
tough problem to solve although not impossible.
Bob Santy
There is 1 Reply.

#: 8605 S10/OS9/6809 (CoCo)
    04-Dec-90  21:59:41
Sb: #8585-#CoCo 3 Emulator?
Fm: John M Semler 74020,736
To: Bob Santy 76417,714 (X)
> "...the 6809 code must be interpreted and converted only when the actual
program function is verified..."
I disagree.  The translator I have in mind can translate 6809 object code into
680x0 code with only 4 bytes of lookahead in a single pass!  I don't care if it
is code or data as I am only concerned about constructing macro cells for each
program memory location that specifies how the state of the 6809 microprocessor
changes if the PC pointer was pointing to it (macro cells are all the same
size).  Kinda like a strip down version of the MC6809 for each memory location.
The "Fast" mode of 6809 emulation assumes that the OS9/6809 program is well
behaved (software does not generate code and execute it).  The "Slowpoke" mode
of emulation would cause the pseudo interpreter to incrementally update itself
for each memory write for complete 6809 emulation.  I believe most OS9 programs
are well behaved which makes the "Fast" mode of emulation a practical first
choice.
I will illustrate my point with a short code fragment (Assume that the cell
size is 16 bytes in this example).
 Virtual 6809 address space   |  680X0 address space
                              |                        Relative
 Location   Value  Opcode     |  6809 Macro Cell       Cell address
  $E503      86    LDA  #$43  |  "LDA #$43"            $E5030
  $E504      43               |  "COMA"                $E5040
  $E505      8B    ADDA #$53  |  "ADDA #$53"           $E5050
  $E506      53               |  "COMB"                $E5060
    .               .                   .
    .               .                   .
    .               .                   .
Note that I have generated macro cells for "COMA", and "COMB".  This ensures
that the emulator will correctly execute even if the programmer got sneaky and
jumped into the middle of a "valid" 6809 program instruction.
John Semler
There are 4 Replies.

#: 8645 S10/OS9/6809 (CoCo)
    08-Dec-90  15:33:02
Sb: #8605-CoCo 3 Emulator?
Fm: Bob Santy 76417,714
To: John M Semler 74020,736 (X)
John:
I decided to check out your suggestion of using a cell approach
to translation.  The following message shows the 68000 code that
I believe is necessary for your example instruction sequence
mapped out into cells.  I'm afraid that maintenance of the 6809
CC register prohibits the use of 16 byte cells (they are too
small to accomodate the code necessary in COMA, COMB and ADDA
#$53).  These cells are full with a 32 byte size.
Since I have only taken the example instructions and coded them
into cells and they are relatively simple ones, I have a sinking
feeling that 32 bytes is not even enough.  They DO however
illustrate my original statement that the 6809 condition codes
must be emulated for automated translation to actually work.
I have not determined how the data space references would be
handled using this cell approach.  I presume that the data
references would be handled by using the first byte of each cell
to hold the 6809 data and the code in the cells would assemble
and disassemble 16 bit references to this data space.  That would
certainly add to the complexity in such cells.
Another troublesome issue is how this approach would handle stack
frame references.  I think that psh and pul instructions would
probably use the 68000 stack, but then what about stack offset
references?
 (continued)

#: 8646 S10/OS9/6809 (CoCo)
    08-Dec-90  15:33:51
Sb: #8605-CoCo 3 Emulator?
Fm: Bob Santy 76417,714
To: John M Semler 74020,736 (X)
 (continued)
Finally, I feel it necessary to voice a concern that I have about
the practicality of a cell approach.  16 byte cells (which most
probably cannot work at all because they are too small) would
require 1 Mb of cells for the maximum 6809 space of 64k.  The 32
byte cells that I illustrate (marginal at best) require 2 Mb.
I think that means many OS9 68000 users would not be able to use
it.
The example in the following message assumes the following:
1.   That D0 is used globally for the emulated A register.
2.   That D1 is used globally for the emulated B register.
3.   That data space for the running program is available with
     its base address in A6.  This is OS9 68000 convention.  The
     global 6809 CC register is shown as the first byte of this
     space.
4.   That several 68000 data registers are available for scratch.
     I have used D4 and D5.
Bob Santy

#: 8647 S10/OS9/6809 (CoCo)
    08-Dec-90  15:34:56
Sb: #8605-#CoCo 3 Emulator?
Fm: Bob Santy 76417,714
To: John M Semler 74020,736 (X)
*
* code for LDA #$43
*
          cell_0
0000 103c 0043       move.b   #$43,d0        this is easy
0004 022e 00f10000   andi.b   #$f1,cc09(a6)  clear N,V and Z
000a 6034            bra.s    cell_2         skip next cell
000c 4e71            nop                     this cell is padded <grin>
000e 4e71            nop
0010 4e71            nop
0012 4e71            nop
0014 4e71            nop
0016 4e71            nop
0018 4e71            nop
001a 4e71            nop
001c 4e71            nop
001e 4e71            nop
*
* code for COMA
*
          cell_1
0020 1a2e 0000       move.b   cc09(a6),d5    get 6809 CC
0024 1805            move.b   d5,d4          save EFHI bits
0026 0204 00f0       andi.b   #$f0,d4
*
* Put 6809 CC in 68000 CCR
*
002a 44c5            move     d5,ccr         prepare for "COMA"
*
* 68000 not.b will set Z, N and V properly
*
002c 4600            not.b    d0             emulator A
002e 42c5            move     ccr,d5         get CCR
*
* But 6809 COMA sets C!
*
0030 0005 0001       ori.b    #1,d5          set C
0034 0205 000f       andi.b   #$0f,d5        preserve NZVC
0038 8a04            or.b     d4,d5          restore EFHI
003a 1d45 0000       move.b   d5,cc09(a6)    save 6809 CC
003e 4e71            nop                     fill cell
 (continued)
There is 1 Reply.

#: 8653 S10/OS9/6809 (CoCo)
    08-Dec-90  16:41:59
Sb: #8647-#CoCo 3 Emulator?
Fm: James Jones 76257,562
To: Bob Santy 76417,714 (X)
Wouldn't it be possible, when translating 6809 code, to do some lookahead to
see whether an instruction is followed by other instructions that both (1)
don't depend on the current value of the condition code and (2) set the
condition code themselves, so that when translating a given instruction
preceding it, you know from context that it really doesn't have to dot the i's
and cross the t's on condition code setting?
You'll almost certainly want to do something special for the usual
implementation of the BASIC09 syscall procedure, which pushes an SWI2 on the
stack. :-) Programs that generate code on the fly, even if they do it in a
re-entrant way, are nasty for this sort of code translation!
There is 1 Reply.

#: 8654 S10/OS9/6809 (CoCo)
    08-Dec-90  17:50:36
Sb: #8653-CoCo 3 Emulator?
Fm: Bob Santy 76417,714
To: James Jones 76257,562 (X)
James:
Well, yes.  However, I hardly think that the cell size can be
reduced to 16 bytes by such a technique.  Even if it could, a 16
byte cell would call for a 1 Mb space for the cells in a 64k
translated address space.  Just not practical.
My original suggestion of doing the translation by actually
running the target program emulation would seem to offer the most
opportunity for the tightest 68000 code.  The resultant code
would not be cell oriented and consequently no restrictions on
the 6809 to 68000 relationship.
The exercising approach would be better able to handle the
BASIC09 and other such nasties as well.  I suppose that this kind
of a translator would be able to figure out that the PC  got into
data space and complain.  I don't think that a 68020 would allow
the address space to change like that.  That would depend on the
memory management arrangement for the particular machine.  We
could probably get a 68000 to fake it.  In any event, I think a
complaint and refusal would be the way I'd go.  BASIC09 would be
un-translatable!
However, BASIC09 does run using the working emulator/interpreter.
Bob Santy

#: 8648 S10/OS9/6809 (CoCo)
    08-Dec-90  15:36:02
Sb: #8605-#CoCo 3 Emulator?
Fm: Bob Santy 76417,714
To: John M Semler 74020,736 (X)
 (continued)
*
* code for ADDA #$53
*
          cell_2
0040 183c 0053       move.b   #$53,d4        get byte
0044 1a2e 0000       move.b   regs(a6),d5    get CC
0048 1405            move.b   d5,d2          copy and
004a 0202 00f0       andi.b   #$f0,d2        preserve EFHI
004e 44c5            move     d5,ccr         make CC current
0050 d004            add.b    d4,d0          add to A
0052 42c5            move     ccr,d5         get CC
0054 0205 000f       andi.b   #$0f,d5        mask NZVC
0058 8a02            or.b     d2,d5          fix CC
005a 1d45 0000       move.b   d5,regs(a6)    in ram
005e 6020            bra.s    cell_4         skip next cell
*
* code for COMB
*
          cell_3
0060 1a2e 0000       move.b   cc09(a6),d5    get 6809 CC
0064 1805            move.b   d5,d4          save EFHI bits
0066 0204 00f0       andi.b   #$f0,d4        
*
* Put 6809 CC in 68000 CCR
*
006a 44c5            move     d5,ccr         prepare for "COMB"
*
* 68000 not.b will set Z, N and V properly
*
006c 4601            not.b    d1             emulator B
006e 42c5            move     ccr,d5         get CCR
*
* But 6809 COMB sets C!
*
0070 0005 0001       ori.b    #1,d5          set C
0074 0205 000f       andi.b   #$0f,d5        preserve NZVC
0078 8a04            or.b     d4,d5          restore EFHI
007a 1d45 0000       move.b   d5,cc09(a6)    save 6809 CC
007e 4e71            nop                     fill cell
          cell_4                             
There are 3 Replies.

#: 8664 S10/OS9/6809 (CoCo)
    10-Dec-90  03:12:01
Sb: #8648-CoCo 3 Emulator?
Fm: John M Semler 74020,736
To: Bob Santy 76417,714 (X)
Bob,
I recoded the example over again and I was able to make them fit in 16 bytes. I
made use of the following facts:
o On entry to or exit from any cell, the 68000 NZVC bits are exactly the same
as the emulated 6809 NZVC bits.
o Code that will overflow a cell is made into a template.  I can pass up to two
short words of data to a template without destroying the integrity of the NZVC
bits. I don't know how many templates I would have to make but I know that it
is less than 1500 (<grin>).
o The emulated 6809 program counter is implicit in the 68000 program counter.
The code for "adda #$53" is complex because of the H bit.  Did you take this
into account in your code?  Did I calculate it right?  The code for "coma" and
"comb" is trivial.  The code for "lda #$43" is slightly more complex because I
have to preserve the 68000 C bit.
I also included two more examples.  One is "BSR $100" and it illustrates the
absolute limits in calling a template.  It also demonstrates how the 6809 stack
frame references are emulated. The other "TFR CC,B" shows how I remap the 68000
NZVC bits into the 6809 CC register before swapping CC and B.
I concede the point that this idea is impractical on all but the largest
system.  I have an 8Mb 68030 system so this will not deter me from trying out
the idea.
The register assignment I am thinking about for my model is as follows:
    D0 emulates A register
    D1 emulates B register
    D2 emulates CC register
    D3 emulates DP register
    A0 emulates S register
    A1 emulates U register
    A2 emulates X register
    A3 emulates Y register
The following data structure emulates the 6809 address space:
        vsect
 m6809  dz.b 65536
        ends
This is the first time I have coded in 68000 assembly language so it is
definite possibility that the code fragments have some bugs.  The code fragment
follows...
   (continued)

#: 8665 S10/OS9/6809 (CoCo)
    10-Dec-90  03:14:41
Sb: #8648-CoCo 3 Emulator?
Fm: John M Semler 74020,736
To: Bob Santy 76417,714 (X)
 psect test_a,0,0,0,0,0
 nam test_a
         vsect
 *
 * 6809 address space (data space emulated here)
 *
 m6809   dz.b    65536
         ends
 *
 * Code for "LDA #$43"
 *
 cell_0
         bcs.s   cell_0a      % preserve C bit
         moveq   #$43,d0      % NZVC bits are ok
         bra.s   cell_2       % skip to next cell
 cell_0a
         moveq   #$43,d0      % NZV bits are ok
         ori     #$01,ccr     % set C bit
         bra.s   cell_2       % skip to next cell
         nop
 *
 * Code for "COMA"
 *
 cell_1
         eori.b  #$ff,d0      % NZV bits are ok
         ori     #$01,ccr     % set C bit
         bra.s   cell_2       % skip to next cell
         nop
         nop
         nop
 *
 * Code for "ADDA #$53"
 *
 cell_2
         move.b  #$53,d7      % set up parameter
         jsr     adda_imed    % call template
         bra.s   cell_4
         nop
         nop
 *
 * Code for "COMB"
 *
 cell_3
         eori.b  #$ff,d1      % NZV bits are ok
         ori     #$01,ccr     % set C bit
         bra.s   cell_4       % skip to next cell
         nop
         nop
         nop
 *
 * Code for "EXG CC,B"
 *
 cell_4
         jsr     exg_cc_b     % call template
         bra.s   cell_5       % skip to next cell
         nop
         nop
         nop
         nop
 *
 * Code for "BSR $100"
 *
 cell_5
         movea.w #$0007,a5    % PC offset
         movea.w #$0100,a4    % bsr address
         jsr     bsr_rel      % call template
         bra.s   cell_7       % goto next cell
         (continued in next message)

#: 8666 S10/OS9/6809 (CoCo)
    10-Dec-90  03:17:33
Sb: #8648-CoCo 3 Emulator?
Fm: John M Semler 74020,736
To: Bob Santy 76417,714 (X)
          (continued from previous message)
 *
 * Template for "xBSR $<xxxx>"
 *
 bsr_rel
         move    ccr,d7       % save NZVC
         subq    #2,a0        % SP predecrement
         clr.l   d6           % clear b32-b16
         move.w  a0,d6        % save SP in b15-b0
         move.w  a5,d5        % free up an addr reg
         move.l  a6,a5        % data area ptr
         adda.l  d6,a5        % add in SP offset
         move.w  d5,(a5)      % save PC on 6809 stack
         clr.l   d4           % clear b31-b16
         move.w  a4,d4        % get bsr address
         asl.l   #4,d4        % calculate cell offset
         lea     cell_0,a4    % get first location
         adda.l  d4,a4        % add cell offset
         move    d7,ccr       % restore NZVC
         jsr     (a4)         % call subroutine
         addq    #2,a0        % pop 6809 SP
         rts
 *
 * Template for "ADDA #$<xx>"
 *
 adda_imed
         andi.b  #$df,d2      % clear H bit in CC
         move.b  d0,d6        % make a copy of A
         andi.b  #$0f,d6      % mask off lsn
         move.b  d7,d5        % get operand
         andi.b  #$0f,d5      % mask off lsn
         add     d5,d6        % add the two halves
         andi.b  #$10,d6      % discard b0..b3
         asl     #1,d6        % H bit in b5
         or.b    d6,d2        % or H bit into CC
         add.b   d7,d0        % NZVC bits ok
         rts
 *
 * Template for "EXG CC,B"
 *
 exg_cc_b
         andi    #$f,ccr      % mask off 68000 NZVC
         move    ccr,d7       % fetch 68000 NZVC
         andi.b  #$f0,d2      % discard invalid 6809 NZVC
         or.b    d7,d2        % save NZVC to 6809 CC
         move.b  d1,d7        % make a copy of B
         exg     d1,d2        % CC <-> B
         andi.b  #$0f,d7      % mask off new NZVC
         move    d7,ccr       % set 68000 CC
         rts
 ends

#: 8876 S10/OS9/6809 (CoCo)
    25-Dec-90  13:01:39
Sb: #8534-#CoCo 3 Emulator?
Fm: MOTD Editor..Bill Brady 70126,267
To: John M Semler 74020,736 (X)
But the book on SoftPC/EGA/AT fibs. On my 32Mhz '030 Mac II Norton pegs it at
only the speed of a NEC V20. (1.5x an 8088 XT @ 4.77MHz.) & Norton don't
measure screen speed. SoftPC is a snail. Try running PC Paintbrush! zzzzz
There is 1 Reply.

#: 8925 S10/OS9/6809 (CoCo)
    28-Dec-90  06:14:41
Sb: #8876-CoCo 3 Emulator?
Fm: John M Semler 74020,736
To: MOTD Editor..Bill Brady 70126,267
Bill,
Those are the kind of results I would expect for a 16Mhz '030.  What kind of
accelerator are you using?
Accelerators that consist solely of a 68030 and a 32Mhz crystal only speeds up
the execution unit, bus cycle time is still 16Mhz.  Only marginal improvement.
(Maybe I am wrong?)
The 32Mhz Dove MaraThon 030 board speeds up the Mac II by a third according to
the August 1990 MacWorld review on accelerators.
John

#: 8537 S10/OS9/6809 (CoCo)
    02-Dec-90  12:51:26
Sb: #8530-CoCo 3 Emulator?
Fm: Mark S 76004,373
To: Kevin Darling (UG Pres) 76703,4227 (X)
Kev, Keep in mind that the Coco-3 was optimized in hardware for level II It's
more then just a 6809 emulation to achive equivilent results.

#: 8525 S10/OS9/6809 (CoCo)
    01-Dec-90  20:48:53
Sb: #VRN
Fm: Ted Miller 76545,457
To:  76625,2273 (X)
Hello Bruce;
     Just a message of appreciation for your VRN upload. Its cleared up a long
standing and puzzling problem I was having with my system. Also looking forward
to your new ACIA driver. Hopefully it will clear up a few rs232 problems I've
been having.
Ted Miller
There is 1 Reply.

#: 8546 S10/OS9/6809 (CoCo)
    02-Dec-90  16:38:07
Sb: #8525-#VRN
Fm: Bruce Isted (UG VP) 76625,2273
To: Ted Miller 76545,457 (X)
Ted,
  Glad you like VRN.  If you don't mind, what was the problem that it cleared
up?  Just wondering...
  Bruce
There is 1 Reply.

#: 8556 S10/OS9/6809 (CoCo)
    02-Dec-90  20:23:50
Sb: #8546-#VRN
Fm: Ted Miller 76545,457
To: Bruce Isted (UG VP) 76625,2273 (X)
Hi Bruce;
Its kind of a long story. The ramdisk software I have installed in my
system,written by Ken Drexler, would not work with the cron utility. Cron would
error out saying that it didn't see a ramdisk installed. Cron wasn't the
problem as it would work with a different ramdisk(i.e Kevin Darlings 'Rammer').
No matter what combination of boot disks I tried cron wouldn't work.I now
realize that my nil driver (being the common denominater on all boot disks)must
have been the problem for cron works fine with your VRN and nil descriptor
installed in conjunction with Ken Drexler's ramdisk.
The other problem was with a C Compiler user interface I use when playing
around with C. All temporary files and libs are used with the ramdisk. With
VDGInt, FTDD,and AGIvirq installed my source would not compile when using the
interface program. I would get a ramdisk sector error. Without the above
modules the interface works fine with the ramdisk. However with your
VRN,ftdd,vi installed I have experienced no problems with the interface program
and VDGInt in the boot disk.
I don't have any  idea why I should have had these problems.All I know is that
the problems disappeared when I installed your VRN modules. Once again thanks.
Ted Miller
There is 1 Reply.

#: 8580 S10/OS9/6809 (CoCo)
    04-Dec-90  00:01:37
Sb: #8556-#VRN
Fm: Bruce Isted (UG VP) 76625,2273
To: Ted Miller 76545,457 (X)
Ted,
  Very strange... I don't know why VRN should apparently work better than
NILDRV, because  never found anything really wrong with NILDRV.  It was more a
case of VRN acted like a null driver anyway, so why not make a NIL descriptor
for VRN?  However, both FTDD and AGIVIRQDr had a few things wrong with them.
But I can't see that they could cause any problems if they weren't in use, as
when you're using the C compiler.  Or do you play Leisure Suit Larry while
waiting for C compiles?  <grin>
  Perhaps its just because VRN is a littel smaller than FTDD, but does the job
of FTDD plus AGIVIRQDr plus NILDRV, and saved you a bit of system space?  I
don't know, but I'm glad it helped you out.
  Bruce
There is 1 Reply.

#: 8623 S10/OS9/6809 (CoCo)
    06-Dec-90  22:20:26
Sb: #8580-#VRN
Fm: Ted Miller 76545,457
To: Bruce Isted (UG VP) 76625,2273 (X)
Hi Bruce;
       The nil driver software I was using was copied from the 'Complete
Rainbow guide to Os9' originally designed for a level 1 system. Maybe something
in that caused a problem although it looked so simple.When I tested it and it
worked it didn't dawn on me that it was the problem. I do have a large bootfile
(>35k) so maybe that was it.
    Also I've applied your cc3io patch for the regular hires mouse. The most
noticable difference? Before I couldn't run my terminal without losing
characters when the gshell screen was displayed on my host Coco. Now there is
no problem. Right now I'm limited to 4800 baud as I don't have the irq hack
intsalled for my terminal port. Perhaps your aciapak replacement will allow
9600 baud without the hack? Any way great work.
Ted Miller
There is 1 Reply.

#: 8652 S10/OS9/6809 (CoCo)
    08-Dec-90  16:40:30
Sb: #8623-VRN
Fm: Bruce Isted (UG VP) 76625,2273
To: Ted Miller 76545,457 (X)
Ted,
  Glad you like the stuff, and thanks.  Kevins Darling's trick to improve IRQ
response was pretty slick, I think!  I'd actually done similar things in other
drivers, but it never occurred to me that it'd work in CC3IO too.
  Bruce

#: 8570 S10/OS9/6809 (CoCo)
    03-Dec-90  10:33:23
Sb: #serial printer on RS232
Fm: Mike Guzzi 76576,2715
To: all
I have heard of people using an ACIAPAK port for a serial printer. I have the
COMM-4 and now own a CGP-220 printer and would like to use it along with my
epson thats hooked onto /p.
i know it would work.. the cgp-220 has a serial port but my concern is the busy
line. if i do something like "list file >/t4" and my cgp-220 is on /t4 will
data flow stop while the printer is busy? what pin on the RS232 should the BUSY
like be hooked to?
Mike
There is 1 Reply.

#: 8610 S10/OS9/6809 (CoCo)
    05-Dec-90  13:59:26
Sb: #8570-#serial printer on RS232
Fm: Lee Veal 74726,1752
To: Mike Guzzi 76576,2715 (X)
Mike, I'm using a Comm-4 to send serial data to my printer, but I'm not using a
Tandy-style serial interface (4-pin DIN).  The serial device that I'm sending
the data to uses a standard RS-232 interface.
In your situation, I think I'd tie the BUSY line from the printer to the DTR
line in the COMM-4 interface.
Tying it to DSR might also be an option, but I'd try DTR first.
Lee
P.S. If you use Multi-Vue (or the Hi-Res Joystick adaptor), then you might
experience some lost data when data is being sent to the printer (via the
Comm-4 port) and the mouse pointer is being moved about.
LV...
There is 1 Reply.

#: 8670 S10/OS9/6809 (CoCo)
    10-Dec-90  20:13:24
Sb: #8610-serial printer on RS232
Fm: Mike Guzzi 76576,2715
To: Lee Veal 74726,1752 (X)
hmmm ill have to try that and see which works.
Mike

#: 8604 S10/OS9/6809 (CoCo)
    04-Dec-90  21:22:52
Sb: #Disto CCHdisk (SCSI)
Fm: james pottage 71750,2012
To: Kevin Darling
Kevin, this is Jim Pottage, I talked to you a while ago about a problem with my
ST125N (SCSI) hard drive. I mentioned that it would not access the drive in the
first attempt. Therefore, it would not run my startup file off the hard drive,
or boot to the hard drive. Well, after a lot of work with Ken Scales (he did
all the modifications and most of the work, I just tested the driver) Ken has a
version of CCHdisk that allows the drive to work on the first access. I believe
he has inserted a restore to track zero to accomplish this bbut for more
accurate information you could talk to Ken. Further, the driver seems to work
with the format command. I just thought I would let you know that the problem
seems to have been solved. Further, if any one else has this problem they might
try getting a hold of Ken.
                                              Jim Pottage
There is 1 Reply.

#: 8607 S10/OS9/6809 (CoCo)
    04-Dec-90  23:02:23
Sb: #8604-Disto CCHdisk (SCSI)
Fm: Kevin Darling (UG Pres) 76703,4227
To: james pottage 71750,2012 (X)
Thanks for the info, Jim!
Oddly, I had just run into a similar thing myself on an embedded SCSI drive,
and ended up having the cc3go module do a coupla chd/chx's to the hard disk in
order for it to finally "take".  I believe that some tandy/coco-HD people found
they had to do the same thing.
I'll talk to Ken. thx! - kev

#: 8613 S10/OS9/6809 (CoCo)
    05-Dec-90  19:52:52
Sb: #How to install 512K?
Fm: MICHAEL ROSEN 73340,2756
To: KEVIN DARLING
KEVIN,
     THE ONLY PROBLEM NOW IS I HAVE (2) COCO 2'S AND NOW A SET OF COCO 3'S I
HAVE AN UNPOPULATED 512K BOARD (TANDY) AND I'M WONDERING WHAT CHIPS TO USE TO
MAKE IT WORK.. HAVE YOU GOT ANY IDEAS ON HOW TO INSTALL 512K IN A COCO 3? ALSO
WHAT CHIPS DO I NEED? I GUESS I CAN HAVE 1 COMPUTER TO HACK ON <GRIN>.. ANYHOW
THANKS FOR TRYING TO HELP ME..I'M NEW AT A COCO 3. I'VE HAD JUST ABOUT ALL THE
ANTIQUE COMPUTERS RADIO SHACK HAD. (2) COCO 1'S,A 4P,A MODEL 3,I STILL HAVE (2)
MC-10'S,(2) COCO 2'S, AND NOW (2) COCO 3'S.. I GUESS IT'S AFFIES GRAVEYARD..
  MICHAEL 73340,2756
There is 1 Reply.

#: 8616 S10/OS9/6809 (CoCo)
    06-Dec-90  01:35:38
Sb: #8613-#How to install 512K?
Fm: Kevin Darling (UG Pres) 76703,4227
To: MICHAEL ROSEN 73340,2756 (X)
Michael - you'll need 16 of the 256K x 1-bit dynamic RAM chips... they usually
have numbers like 41256 or anything xx256 on them.
You remove the original 4 coco ram chips, and also two capacitors nearby...
unfortunately my memory fails me as to their numbers on the coco3 board.
Guys???
There is 1 Reply.

#: 8617 S10/OS9/6809 (CoCo)
    06-Dec-90  05:58:23
Sb: #8616-#How to install 512K?
Fm: Dan Robins 73007,2473
To: Kevin Darling (UG Pres) 76703,4227 (X)
Kev,
  I -THINK- the old ones are the 4154's, maybe (then maybe not...I'm going off
gray cell memory on this one).
   Dan
There is 1 Reply.

#: 8618 S10/OS9/6809 (CoCo)
    06-Dec-90  16:53:25
Sb: #8617-#How to install 512K?
Fm: Kevin Darling (UG Pres) 76703,4227
To: Dan Robins 73007,2473 (X)
Thx Dan.... but I meant the capacitor numbers to clip out for the 512 upgrade.
<grin>  Thanks!
There is 1 Reply.

#: 8621 S10/OS9/6809 (CoCo)
    06-Dec-90  20:51:01
Sb: #8618-How to install 512K?
Fm: Dan Robins 73007,2473
To: Kevin Darling (UG Pres) 76703,4227 (X)
Kev,
  Oh wellst. Sorry. Guess that means I don't get any points, eh?
  Dan

#: 8620 S10/OS9/6809 (CoCo)
    06-Dec-90  19:56:44
Sb: #HD problems
Fm: LUTE MULLENIX 70721,2230
To:  76703,4227 (X)
Kevin:
I'm having some trouble putting togeather this HD system, I keep getting an
ERROR 246 (Device not ready). Is it possible that a bad cable will do this? I
have sent the drive and the Disto adapter back and I get the same error. I even
pulled my 3-1 board out of the controller and put the adapter in there with the
same results. What do you think?
If it might be the cable, where can I get about a three foot SCSI cable? The
one I have is one I cobbled up out of some stuff I had around here.
     >Lute<
There are 2 Replies.

#: 8625 S10/OS9/6809 (CoCo)
    06-Dec-90  23:52:14
Sb: #8620-#HD problems
Fm: Kevin Darling (UG Pres) 76703,4227
To: LUTE MULLENIX 70721,2230 (X)
Lute - could be the cable, or perhaps the power supply (?).
Try this first tho: make a boot disk with H0 renamed to HD or something. Boot,
then try an "iniz /hd" and see if that works.
Or, if you're already doing this... ummm... what kind of hard disk is it? Did
it format okay using that disto rsdos program?
There is 1 Reply.

#: 8637 S10/OS9/6809 (CoCo)
    08-Dec-90  11:08:55
Sb: #8625-HD problems
Fm: LUTE MULLENIX 70721,2230
To: Kevin Darling (UG Pres) 76703,4227 (X)
Kevin:
I found the problem, it was an addressing thing. However there is something
else you may be able to help me with.
Quite often when attempting to run Basic programs or those with basic
subroutines I get an error #43. And now that the HD is going some of the stuff
that used to work is giving it to me.
When working up a boot file for the HD I smoked the one on my previous master
disk (the one the basic stuff worked from) and don't remember what was in it.
and the backup I had doesn't seem to work.
Any ideas? Ad9 dosn't want to work and tis the season!
     >Lute<

#: 8630 S10/OS9/6809 (CoCo)
    07-Dec-90  09:13:09
Sb: #8620-#HD problems
Fm: Pete Lyall 76703,4230
To: LUTE MULLENIX 70721,2230 (X)
Cable is a prime candidate. If I recall, SCSI is simply 50 pin end-end. You
should be able to build another cheaply, or pick one up from your local
computer place for $20 or under.
Pete
There is 1 Reply.

#: 8638 S10/OS9/6809 (CoCo)
    08-Dec-90  11:09:47
Sb: #8630-#HD problems
Fm: LUTE MULLENIX 70721,2230
To: Pete Lyall 76703,4230 (X)
Pete:
Found the problem, it was an addressing thing. Thanks for the input though.
About the local computer shops. We have Radio Shack and Ultra Inc. Though I
rate the local RS as excellent they carry no SCSI stuff, and Ultra, well they
sell Macs but they weren't sure what SCSI was, and were sure they had never
seen a 50 pin cable. Oh well.
     >Lute<
There is 1 Reply.

#: 8878 S10/OS9/6809 (CoCo)
    25-Dec-90  13:27:18
Sb: #8638-HD problems
Fm: MOTD Editor..Bill Brady 70126,267
To: LUTE MULLENIX 70721,2230 (X)
Come to think of it, I've never seen a 50-pin *cable* myself! <grin>

#: 8622 S10/OS9/6809 (CoCo)
    06-Dec-90  20:57:33
Sb: #8575- StermHelp
Fm: VERN STOCKMAN 70415,1057
To: Kevin Darling (UG Pres) 76703,4227 (X)
list msgs/thanks

#: 8624 S10/OS9/6809 (CoCo)
    06-Dec-90  23:08:30
Sb: COCO lv 1.2 BBS
Fm: LAVERN SCHOONOVER 73700,3217
To: all
Does anyone know of a COCO lv.1.2 BBS program that I could buy I need it in
order to get my hard drive to work. It only is accessable under lv 1.2
                                -=*< L.V. >*=

#: 8633 S10/OS9/6809 (CoCo)
    07-Dec-90  18:38:59
Sb: #PCoid SCSI and Burke**2?
Fm: James Jones 76257,562
To: All
I got a JDR Microdevices (or whatever it is) catalog in the mail today.  In it
I found a SCSI controller card for PClones (floppy and hard $80, hard only $50,
to one significant digit I think :-).
What I'm wondering is this: can one connect this beast to a Burke**2 interface
and, with appropriate drivers, connect SCSI stuff to one's CoCo?  The floppy
hardware is claimed to handle all the various flavors of floppy sizes up to
1.44 Mbytes.  That would be a nice setup, I think, if it could be pulled off.
Anyone know whether it's possible?
There is 1 Reply.

#: 8641 S10/OS9/6809 (CoCo)
    08-Dec-90  13:25:17
Sb: #8633-#PCoid SCSI and Burke**2?
Fm: Zack Sessions 76407,1524
To: James Jones 76257,562 (X)
The CoCo-XT by Burke&Burke does not support SCSI controllers.
There is 1 Reply.

#: 8644 S10/OS9/6809 (CoCo)
    08-Dec-90  13:39:11
Sb: #8641-#PCoid SCSI and Burke**2?
Fm: James Jones 76257,562
To: Zack Sessions 76407,1524 (X)
Shucks.  I don't know enough about the hardware involved to know whether it was
possible.  Ah, well...I wonder how hard it would be to do?
There is 1 Reply.

#: 8879 S10/OS9/6809 (CoCo)
    25-Dec-90  13:34:43
Sb: #8644-PCoid SCSI and Burke**2?
Fm: MOTD Editor..Bill Brady 70126,267
To: James Jones 76257,562 (X)
James, the SCSI adapters for clones have a driver on-board which is loaded by
BIOS at boot time. (as a BIOS extension). You could disassemble the code on a
PC & write a driver for OS-9. (you'd prolly have to disable the ROM later) The
floppy side would be easier. I bought one of those boards for $30. The one with
the floppy & SCSI was about $6 more.

#: 8696 S10/OS9/6809 (CoCo)
    12-Dec-90  19:19:12
Sb: #File structure
Fm: Denise Tomlinson 71021,3274
To: All
Can anyone tell me how a file is put on a disk for a Color computer under os9?
I have a arced file from the music library that consists of 36 granules. This
file is too big to use "dosor9" or others to transform a Dos format to a os9
format. I thought if I split it into 2 files and used "dosor9" to 2 different
disks and then copied them to a genuine os9 format, then I could modify the gat
tables and file ending bytes to combine into 1 file. I have done this to files
in Dos format using a disk editor. But I have never done it under os9. I have
the "zapper". Lets just say for simplicity that a disk I want to trace a file
has only a "root " directory to make things simple. Thanks, Denise PS: I only
access compuserve via Dos system because I don't have a modem pak for os9
operation.
There are 2 Replies.

#: 8698 S10/OS9/6809 (CoCo)
    12-Dec-90  20:11:30
Sb: #8696-File structure
Fm: James Jones 76257,562
To: Denise Tomlinson 71021,3274 (X)
If you split the file on the DECB side into some number of parts, then the
simplest way to get them back together on the OS-9 side would be to do
something like
merge file1 file2 file3 >bigfile
(change as appropriate for the number of parts and the names you actually do
give them).  Doing that does imply the existence of two copies of the data, at
least for a while.  Is the file big enough that two copies won't fit on one
disk?  If that's the case, and you only have one disk, then life becomes more
difficult.
On the other hand, though, I thought someone had come up with a fix for one of
the DECB->OS-9 file copy programs that had to do with files longer than 32K. 
You might look around to see whether there's a program here that does the job
and avoids the file size problem.

#: 8720 S10/OS9/6809 (CoCo)
    13-Dec-90  17:17:53
Sb: #8696-File structure
Fm: Zack Sessions 76407,1524
To: Denise Tomlinson 71021,3274 (X)
If getting the file split and copied over worked, you can them combine the two
with a merge command:
OS9: merge file1 file2 >combined_file
Zack

#: 8706 S10/OS9/6809 (CoCo)
    13-Dec-90  07:21:34
Sb: #SS.WTRK
Fm: William Phelps 75100,265
To: ALL
How does the SS.WTRK system call know what drive to use? If a path cannot be
opened on an unformatted drive, then how can a path number be generated?
William
There are 2 Replies.

#: 8711 S10/OS9/6809 (CoCo)
    13-Dec-90  09:17:59
Sb: #8706-#SS.WTRK
Fm: Pete Lyall 76703,4230
To: William Phelps 75100,265 (X)
William -
Good question... What about the combination of doing an open on /DD@, followed
by an SS.Freeze (inhibits update of path/disk parameters the next time LSN0 is
read)?
Pete
There is 1 Reply.

#: 8732 S10/OS9/6809 (CoCo)
    14-Dec-90  06:44:45
Sb: #8711-#SS.WTRK
Fm: William Phelps 75100,265
To: Pete Lyall 76703,4230 (X)
SS.FREEZE?
What page of the manual is that on?
William
There is 1 Reply.

#: 8738 S10/OS9/6809 (CoCo)
    14-Dec-90  09:12:42
Sb: #8732-#SS.WTRK
Fm: Pete Lyall 76703,4230
To: William Phelps 75100,265 (X)
William -
You may be a victim of the RS OS9 manuals. The $40 I spent eons ago to get the
'real' MW manuals was one of the best $40 I ever spent.
Page 11-70 of the OS9 System Programmer's Manual says:
  SS.FRZ
  Input:   A - Path Number
           B - SS.FRZ function code ($0A)
  Output:  None
Function:  Inhibits the reading of the identification sector (LSN0) to memory
DD.xxx variables (that define the disk formats) so non-standard disks may be
read.
Pete
P.S. May already be defined in your os9defs or OS9.H files.
There is 1 Reply.

#: 8746 S10/OS9/6809 (CoCo)
    15-Dec-90  04:31:15
Sb: #8738-#SS.WTRK
Fm: William Phelps 75100,265
To: Pete Lyall 76703,4230 (X)
When does SS.FRZ have to be called if reading a disk?
William
There is 1 Reply.

#: 8749 S10/OS9/6809 (CoCo)
    15-Dec-90  08:15:42
Sb: #8746-SS.WTRK
Fm: Pete Lyall 76703,4230
To: William Phelps 75100,265 (X)
William -
I believe SS.FRZ is a single shot affair. It prevents the next would-be read of
LSN0. In that case, the time to use it is before any implicit or explicit read
of LSN0. I'd do it just after opening the path.
Pete

#: 8715 S10/OS9/6809 (CoCo)
    13-Dec-90  11:16:20
Sb: #8706-#SS.WTRK
Fm: Kevin Darling (UG Pres) 76703,4227
To: William Phelps 75100,265 (X)
Pete has it.  Try a simple basic09 program which opens "/d0@"... you'll see
that the drive light doesn't come on.  Thus you have an open path. This is
exactly what format does.
There is 1 Reply.

#: 8733 S10/OS9/6809 (CoCo)
    14-Dec-90  06:47:15
Sb: #8715-#SS.WTRK
Fm: William Phelps 75100,265
To: Kevin Darling (UG Pres) 76703,4227 (X)
I tried that, but the system tried to read the disk.
William
There is 1 Reply.

#: 8739 S10/OS9/6809 (CoCo)
    14-Dec-90  09:13:32
Sb: #8733-#SS.WTRK
Fm: Pete Lyall 76703,4230
To: William Phelps 75100,265 (X)
William -
You tried open /dd@ ? Are you sure it wasn't open /dd ?
Pete
There is 1 Reply.

#: 8747 S10/OS9/6809 (CoCo)
    15-Dec-90  04:32:00
Sb: #8739-#SS.WTRK
Fm: William Phelps 75100,265
To: Pete Lyall 76703,4230 (X)
Well, I finally got it to work with:
  OPEN #path,"/d0@":UPDATE But now I have two more questions.  Is it possible
to read sector 0 if the standard OS-9 info is not there?  If the pathname is
input externally, then how can one tell the difference between a name for a
floppy and a name for another type of disk. That last question assumes that the
descriptors and drivers do not have to be the standard ones.
William
There are 2 Replies.

#: 8750 S10/OS9/6809 (CoCo)
    15-Dec-90  08:18:29
Sb: #8747-#SS.WTRK
Fm: Pete Lyall 76703,4230
To: William Phelps 75100,265 (X)
William -
Not sure I understand the question. If you use SS.FRZ, I believe it's your
responsibility to know what the disk's parameters are before you attempt to
read it, and set them accordingly in the path yourself (using SS.OPT).
The business about the name being external slipped right by me... care to
rephrase & reask?
Pete
There is 1 Reply.

#: 8773 S10/OS9/6809 (CoCo)
    16-Dec-90  04:39:45
Sb: #8750-#SS.WTRK
Fm: William Phelps 75100,265
To: Pete Lyall 76703,4230 (X)
If the path to be opened is not in the program but rather taken from the
command line or an input from the keyboard, then how can it be determined that
a device is definitely a floppy drive?  Reading IT.TYP is not good enough
because some devices like ram disks look like flopies.
Does Microware have any manuals that explain everything in the defs files? 
Most are obvious, but some ...
William
There is 1 Reply.

#: 8785 S10/OS9/6809 (CoCo)
    16-Dec-90  12:07:24
Sb: #8773-#SS.WTRK
Fm: Pete Lyall 76703,4230
To: William Phelps 75100,265 (X)
William -
Basically, it the descriptor _says_ it's a floppy, it's a floppy. That's as
sure as you can be. What other assurances do you want/need? Of course if it HAD
to be a floppy for some obscure reason, you could write some low level nasty
code to watch the index pulse LED from the disk bus, but that's incredibly
ugly. I guess I fail to see the need? Care to explain why it's so important?
Half of the beauty of os9/unix is that all devices and files look pretty much
the same. Looks like you're trying to head the other way.
Pete
There is 1 Reply.

#: 8797 S10/OS9/6809 (CoCo)
    17-Dec-90  05:11:42
Sb: #8785-#SS.WTRK
Fm: William Phelps 75100,265
To: Pete Lyall 76703,4230 (X)
Actually, I wanted to add some extra disk functions without writing a whole new
driver.  The new functions might cause trouble if used on anything other than a
floppy.
William
There is 1 Reply.

#: 8803 S10/OS9/6809 (CoCo)
    17-Dec-90  12:25:32
Sb: #8797-#SS.WTRK
Fm: Pete Lyall 76703,4230
To: William Phelps 75100,265 (X)
Well, in that case you should just observe the floppy/hard bit in the path
(device) descriptor.
Pete
There is 1 Reply.

#: 8830 S10/OS9/6809 (CoCo)
    19-Dec-90  16:27:08
Sb: #8803-SS.WTRK
Fm: William Phelps 75100,265
To: Pete Lyall 76703,4230 (X)
I guess I will narrow the possibilities down to floppy or hard disks using the
port address.  Then I will use the TYP byte.
William

#: 8762 S10/OS9/6809 (CoCo)
    15-Dec-90  19:50:46
Sb: #8747-#SS.WTRK
Fm: Kevin Darling (UG Pres) 76703,4227
To: William Phelps 75100,265 (X)
William - I think few drivers support SS.Frz, so the general answer is: it's
pretty tough to read LSN 0 of a non-os9 disk.  If you're using any of the
Santy-like floppy drivers... the disto cc3disk, bob's cc3disk, or I think
isted's cc3disk... then I believe you can use SS.Opts to set the non-OS9 disk
bit ($40) in the Typ byte of the path descriptor.
That causes those drivers to ignore LSN0, and use the defaults from the device
descriptor (actually, the defaults in the path descriptor, which were copied
from the dev desc).  - kev
There is 1 Reply.

#: 8774 S10/OS9/6809 (CoCo)
    16-Dec-90  04:40:10
Sb: #8762-SS.WTRK
Fm: William Phelps 75100,265
To: Kevin Darling (UG Pres) 76703,4227 (X)
If the TYP byte were changed in the Device descriptor, would that lockout error
#249 regardless the of cause.
William

#: 8714 S10/OS9/6809 (CoCo)
    13-Dec-90  10:52:05
Sb: #LISP for CoCo3
Fm: David Betz 76704,47
To: all
I just received a letter from someone in Belgium asking about a more recent
version of XLISP that runs on the CoCo3.  He's got version 1.1 and is looking
for 1.7 or later.  Unfortunately, I didn't do the CoCo3 port of 1.1 and don't
know who did.  I would doubt that version 1.7 would fit in the 64K address
space allowed by OS-9 L2 (or is it 64K of code and 64K of data?).  So, does
anyone know of a version of LISP or Scheme that will run on a CoCo3 under OS-9
L2?
Thanks in advance, David Betz
There is 1 Reply.

#: 8722 S10/OS9/6809 (CoCo)
    13-Dec-90  18:11:20
Sb: #8714-#LISP for CoCo3
Fm: James Jones 76257,562
To: David Betz 76704,47 (X)
I did the 1.1 port; I also ported 1.2.  I don't recall seeing 1.3, and I think
1.4 got too big (no separate I & D space on the 6809).  I tried SIOD, and it
runs fine on OSK but pukes immediately on my CoCo.  I really need to go ahead
and upload XLisp 1.2 here if I haven't already.
There is 1 Reply.

#: 8734 S10/OS9/6809 (CoCo)
    14-Dec-90  08:53:52
Sb: #8722-#LISP for CoCo3
Fm: David Betz 76704,47
To: James Jones 76257,562 (X)
Thanks.  I figured that 1.7 wasn't going to fit.  To tell the truth, I can't
remember what the differences between 1.1 and 1.2 were.  Also, there never was
a version 1.3 released.  For some reason, I went right from 1.2 to 1.4.  I'm in
the process of porting XScheme to OSK, but I doubt it would fit into 64K on the
CoCo3.  Do you know of any commercial implementations of LISP on the CoCo?
There is 1 Reply.

#: 8743 S10/OS9/6809 (CoCo)
    14-Dec-90  23:08:41
Sb: #8734-#LISP for CoCo3
Fm: James Jones 76257,562
To: David Betz 76704,47 (X)
I don't know about the CoCo, but there definitely is or was a Lisp for
OS-9/6809 sold in Japan, called "Lisp09."
There is 1 Reply.

#: 8779 S10/OS9/6809 (CoCo)
    16-Dec-90  08:07:47
Sb: #8743-#LISP for CoCo3
Fm: Dan Robins 73007,2473
To: James Jones 76257,562 (X)
James,
  I recall seeing a LISP for OS9 in the User's group library, only I think they
called it XLISP in this case. It may not be there now...but worth asking Gwit
to see if it's archived.
   Dan
There is 1 Reply.

#: 8783 S10/OS9/6809 (CoCo)
    16-Dec-90  10:52:55
Sb: #8779-#LISP for CoCo3
Fm: James Jones 76257,562
To: Dan Robins 73007,2473 (X)
That's David Betz's Lisp interpreter (with object-oriented extensions).  I
guess I should go ahead and upload the 1.2 version in the "Languages" DL.
There is 1 Reply.

#: 8798 S10/OS9/6809 (CoCo)
    17-Dec-90  06:05:10
Sb: #8783-#LISP for CoCo3
Fm: Dan Robins 73007,2473
To: James Jones 76257,562 (X)
James,
   Not a bad idea (hint,hint,grin)...since it is topical at the moment, and
others might want to play with it as well. Whatchathink?
   Dan
There is 1 Reply.

#: 8802 S10/OS9/6809 (CoCo)
    17-Dec-90  07:52:47
Sb: #8798-#LISP for CoCo3
Fm: James Jones 76257,562
To: Dan Robins 73007,2473 (X)
OK.  It's kinda hectic this week, but I'll see if I can't do it.
There is 1 Reply.

#: 8883 S10/OS9/6809 (CoCo)
    25-Dec-90  14:06:19
Sb: #8802-#LISP for CoCo3
Fm: MOTD Editor..Bill Brady 70126,267
To: James Jones 76257,562 (X)
>>>> Rumor control! There is no 64k limit in Level 2. Just a 64 k segment space
that the application must manage itself. (unfortunately).
There is 1 Reply.

#: 8904 S10/OS9/6809 (CoCo)
    26-Dec-90  15:10:20
Sb: #8883-#LISP for CoCo3
Fm: David Betz 76704,47
To: MOTD Editor..Bill Brady 70126,267 (X)
Does that mean that an individual application can access more than one 64K
segment at a time?  How does it do that?  Are there mapping calls?
There is 1 Reply.

#: 8908 S10/OS9/6809 (CoCo)
    27-Dec-90  05:06:12
Sb: #8904-#LISP for CoCo3
Fm: MOTD Editor..Bill Brady 70126,267
To: David Betz 76704,47 (X)
What it means is that an application can use more that 64k. You can a
non-mapped load, which puts modules in memory, but don't link them. You can
them call them into your space when you need them. WizPro does this by keeping
16k free in its workspace, then RUNning & KILLing procedures as needed. Last
time I tallied the total program size was about 128k.
There is 1 Reply.

#: 8911 S10/OS9/6809 (CoCo)
    27-Dec-90  11:24:11
Sb: #8908-LISP for CoCo3
Fm: David Betz 76704,47
To: MOTD Editor..Bill Brady 70126,267
Can you use this trick to get access to more than 64K of data?  XLisp and
XScheme mostly need more data space, not more code space.

#: 8752 S10/OS9/6809 (CoCo)
    15-Dec-90  14:48:30
Sb: #Help
Fm: The Rev. Wayne C. Paul 76477,142
To: Mike Haaland 72300,1433 (X)
Dear Mr. Haaland: On pg 6 of the doc's for EdPoint 1.2 you mention the
following: Datadialog and icon display rotuines from Toby Farley BASIC09 MVDemo
button routine from Kevin Darling. Are these files available on CIS? I am still
a beginning programmer, but t is starting to come together. I have started
working on an OS9 version of MAx-10. As soon as I get any working procedures, I
will upload them to CIS. Thank you Brother Jeremy, CSJW CIS- The Rev. Wayne C.
Paul 76477,142
There are 2 Replies.

#: 8807 S10/OS9/6809 (CoCo)
    17-Dec-90  20:39:11
Sb: #8752-#Help
Fm: Mike Haaland 72300,1433
To: The Rev. Wayne C. Paul 76477,142 (X)
The files you are looking for are in Lib 10.  The one with the neat Icon
display and DataDialog routines is called EDCON.AR and was written by Toby
Farley.  The second Basic09 source file is called MVDEMO.AR and was written by
Kevin Darling.  Both are in Lib 10.
Good Luck on the Max-10 clone.  Hope it's a BIG success,
Mike
There is 1 Reply.

#: 8821 S10/OS9/6809 (CoCo)
    19-Dec-90  01:00:07
Sb: #8807-Help
Fm: The Rev. Wayne C. Paul 76477,142
To: Mike Haaland 72300,1433 (X)
Thank you Mike, I will post anything that I come up with.

#: 8808 S10/OS9/6809 (CoCo)
    17-Dec-90  20:49:59
Sb: #8752-#Help
Fm: Mike Haaland 72300,1433
To: The Rev. Wayne C. Paul 76477,142 (X)
OOps!  The second file is named MVTEST.AR  (not MVDEMO.AR)  Sorry about that.
Mike
There is 1 Reply.

#: 8822 S10/OS9/6809 (CoCo)
    19-Dec-90  01:01:05
Sb: #8808-Help
Fm: The Rev. Wayne C. Paul 76477,142
To: Mike Haaland 72300,1433 (X)
Got it. Thanks again. WCP+

#: 8778 S10/OS9/6809 (CoCo)
    16-Dec-90  07:52:36
Sb: 900 Number Revisited
Fm: James Jones 76257,562
To: All
Well...I called the 900 number again last night--guess I just couldn't resist.
Evidently "Ask Mr. Computer" does update the information, because this time
they mentioned a slightly different number (OS-9 ahead by 125), and gave a
total vote (227 votes altogether, which one can readily deduce from that
"RSDOS" got 51 votes and OS-9 176--i.e. 77% for OS-9, 23% for DECB, which
finally gives some idea of what's happening).  I still am pretty curious about
just who set up this poll.

#: 8792 S10/OS9/6809 (CoCo)
    16-Dec-90  19:17:32
Sb: #Starting out (part 1)
Fm: Carlyle Hudkins 72040,2754
To: all
     I have a 128k CoCo 3, 1 FD-502 drive, which I have had for about 3 years. 
I recently bought OS9 and am having trouble adjusting to this new environment. 
I am conversant in CoCo BASIC, IBM BASIC, and IBM DOS.  OS9 seems somewhat like
IBM DOS, with enough changes to make me think I know what I'm doing when in
reality I don't.
     My most recent problem is this:
     I went through the tedious disk-swapping using CONFIG, to create a
customized System Master, as described in the "Getting Started" section of the
manual, and created a disk with options PIPE, D040D, TERMWIN, W, W1, and FULL
command set.  Now I need to know what I will need to do in order to change the
options, without going through the immensely painful CONFIG experience again. 
For instance, I want to change the usable Windows so that I might use graphics
windows from BASIC09.  I noticed that OS9gen can redo the Boot file, but there
is a warning that this can fragment the Boot file, rendering the disk unusable.
Other questions:
     Is the disk I made with CONFIG actually double-sided?  I tried to format
it using "format '40' 2," but through trial and error I discovered that the
option "2" was not usable with the drive, as OS9 saw it at that time as a
35-trk, SS drive.  Therefore, I did a regular '40' format and proceeded to
CONFIG, telling CONFIG that I wanted D040D, in hopes that I will at least be
able to format a DS from that disk.  I don't yet know of a way to test a disk
to see if it is DS, except from regular BASIC, and since I made the disk 40-trk
BASIC wouldn't be able to use it anyway.  It does not seem likely, in
hindsight, that the disk I made is actually double sided, as only one side was
formatted.  I hope I will not have to use CONFIG again in order to make a real,
DS, "Customized System Master"!
     Can I run a pure ML program, such as my terminal (Ultimaterm v4.0) or word
processor (Simply Better v1.1) from a window so that I do not have to leave OS9
(which involves hitting Reset twice, necessitating removing the disk)?  The WP
uses a regular BASIC boot program to
There are 2 Replies.

#: 8795 S10/OS9/6809 (CoCo)
    17-Dec-90  02:04:02
Sb: #8792-Starting out (part 1)
Fm: James Jones 76257,562
To: Carlyle Hudkins 72040,2754 (X)
You cannot run a program under OS-9 that was not written to run under OS-9, at
least not unless you go to a considerable amount of trouble modifying it, as
Chris Burke figured out how to do to the Color BASIC interpreter to get RSB.
Download the "dmode" utility from here, and use it to modify th device
descriptor for /d0 and /dd in memory to say the dt(rive is double-sided, then
try formatting a disk.

#: 8801 S10/OS9/6809 (CoCo)
    17-Dec-90  07:48:27
Sb: #8792-#Starting out (part 1)
Fm: William Phelps 75100,265
To: Carlyle Hudkins 72040,2754 (X)
CONFIG does not change the format of the disk it writes on; so, the disk you
made is still a plain SS one.  Reboot using that disk; then xmode the window
and printer descriptors to your liking.  Format a new disk, and use COBBLER ow
it.  Copy "startup" over to the new disk.  Make a "CMDS" directory, and copy
shell and grfdrv into it(you can also copy any other commands you want).  Now
you will have a 40trk DS bootdisk.
William
There is 1 Reply.

#: 8804 S10/OS9/6809 (CoCo)
    17-Dec-90  17:03:38
Sb: #8801-#Starting out (part 1)
Fm: Carlyle Hudkins 72040,2754
To: William Phelps 75100,265 (X)
Thanks, William, but as a complete beginner how should I xmode the windows (I
don't have a printer)?  The manual says a lot about pauses, pages, null, and
the like, but I'm not sure exactly what I should use it on in order to get the
right results.
        If I understand this, the disk I now have should allow me to format a
new disk DS.  Then, I must create a bootfile (using Cobbler, which will put the
modules I booted with onto the new disk).  Then I copy "startup," use makdir to
put CMDS on the new disk, copy "shell" & "grfdrv" (if I understand, shell is
the actuall program I talk to OS9 through, and grfdrv is necessary if I want to
enable a graphics window).  I will try these, and see what happens!
                Carlyle H
There are 3 Replies.

#: 8805 S10/OS9/6809 (CoCo)
    17-Dec-90  19:51:52
Sb: #8804-#Starting out (part 1)
Fm: DAVID DE FEO 71630,721
To: Carlyle Hudkins 72040,2754 (X)
Carlyle, to be able to do a double sided format, the drive descriptor in the
boot file (d0 and ddd0) have to be the type for double sided drives. Your
bootfile has to contain the following: d0_40d.dd and ddd0_40d.dd. You might
want to use Config again to substitute these drivers in for the old ones, but
if you feel adventerous, you can use OS9gen.  It will probably be faster.  You
need to "build" a file called bootfile containing all the modules that you want
in your system(you can find all the modules in the Modules directory of the
config disk).  Some modules that you have to include are: init, ioman, rbf.mn,
cc3disk, d0_40d.dd, ddd0_40d.dd, scf.mn, cc3io, term_win.dt, clock.60hz, cc3go,
w.dt, w1.dt, vdgint.io, grfint.io. Once the file is made "chd /d0/Modules" and
"chx /d0/cmds" for single drive: os9gen /d0 -s </d0/bootlist. Remember that the
bootlist must be in /d0.        Have fun, Dave.
There is 1 Reply.

#: 8806 S10/OS9/6809 (CoCo)
    17-Dec-90  20:01:05
Sb: #8805-Starting out (part 1)
Fm: DAVID DE FEO 71630,721
To: DAVID DE FEO 71630,721 (X)
Oops, some clarifications....you have to put this boot on a 35 track disk and
then make a startup file and put shell and grfdrive in CMDS along with whatever
other commands you want.  Then you can format a disk for 40 ds. The disk you
format (format /d0) will have a double sided format.  Then, load "cobbler" into
memory. Then, put your newly formated disk into your drive and do "cobbler
/d0".  Now you have a bootable 40 trk doublesided disk(don't forget to put
startup on it as well as put shell and grfdrive in the Cmds dir).             
Dave

#: 8816 S10/OS9/6809 (CoCo)
    18-Dec-90  00:38:58
Sb: #8804-Starting out (part 1)
Fm: Jim Peasley 72726,1153
To: Carlyle Hudkins 72040,2754 (X)
Carlyle;
  There's lots of startup help in the libs here - some of the ones that come to
mind right away are Bruce Isted's WINVDG.AR in lib 10, Kevin Darling's
BESTOF.TXT also in 10, Mark Griffith's tutorial on making boot disks painlessly
(can't remember the name, but a search on 76070,41 should get it for you.) and
many others.
  If you have specific needs, ask and someone will remember where that 'just
what you need' file resides.
  Oh, yeah, you asked about xmode'ing windows - that's what Bruce's WINVDG does
- it sets all the descriptors that you will EVER need, from w1 - w14 and v1 -
v7.  There's also a WMODE.AR that'll let you modify descriptors in real time
(can't remember the author tho).
 ...Jim

#: 8831 S10/OS9/6809 (CoCo)
    19-Dec-90  16:27:36
Sb: #8804-#Starting out (part 1)
Fm: William Phelps 75100,265
To: Carlyle Hudkins 72040,2754 (X)
If you don't know what you want to change, then don't change it until you need
to change it.  I was guessing about the printer; do you use a full blown word
processor just for programs?
William
There is 1 Reply.

#: 8840 S10/OS9/6809 (CoCo)
    21-Dec-90  12:18:31
Sb: #8831-Starting out (part 1)
Fm: Carlyle Hudkins 72040,2754
To: William Phelps 75100,265 (X)
I use the WP to for my poems and stories, and to compose mail.  Haven't written
any progs with it yet.
                        Carlyle H

#: 8793 S10/OS9/6809 (CoCo)
    16-Dec-90  19:21:20
Sb: Getting started (part 2)
Fm: Carlyle Hudkins 72040,2754
To: all
Continued from previous message:
    The WP uses a BASIC boot program to set up defaults and load the ML--can
this be accomodated by a window?
    Thanks for your time and trouble--
                Carlyle H (I wrote the last message offline with Simply
Better--there may be a few glitches due to different character compatibilities
with CIS; I also unknowingly exceeded the msg size requirements, and I got hung
up while uploading!  What a day!)

#: 8794 S10/OS9/6809 (CoCo)
    16-Dec-90  20:59:31
Sb: #disto SCII boot disorder
Fm: KENHEIST 71750,551
To: all
General info to any in possible need.
 I have used a Disto SCII for about 9 months and have had a lot of trouble with
dist format and a correct read and writes to dist. Called CRC ,even sent back
the controller and 4n1 card.They checked it out, even replaced the drive
card;but .... to no avail. The problem contenued. It was most pronounced with
the MPI and extra cards in place, allmost gone with just the disto in the
computers side slot.
 Any way, even got Tony D. into the problem but still nothing.
 Finally this evening after months of runnning in the sleep mod tried something
that I had not thought of in this connection. Reworked the boot
list.......BINGO!!!!
  I'v had boot disorders be for but never this way. The problem only arises
when the IRd is used. Moved RBF.mn below the disk drive discriptors and driver.
This left OS9p1 OS9p2 OS9p3 INIT and IOMAN at the top of my boot.
   Sometimes you just have to let go..... \exit
There is 1 Reply.

#: 8799 S10/OS9/6809 (CoCo)
    17-Dec-90  06:09:42
Sb: #8794-disto SCII boot disorder
Fm: Dan Robins 73007,2473
To: KENHEIST 71750,551 (X)
Ken,
   It's known as the BLOB....the Boot List Order Bug. It's not a definable bug,
with a known fix, but you did what many others have done...play with it until
it works. I think Kevin Darling placed a file in LIB 10 that discusses the BLOB
in better detail.
   Dan

#: 8812 S10/OS9/6809 (CoCo)
    18-Dec-90  00:01:17
Sb: #Eliminator repair
Fm: JOERG SATTLER 74016,631
To:  76625,2273 (X)
Hi Bruce. A couple days ago Frank Hog Labs. informed me that my Eliminator
board was send 12/12/90) Would you please be kind enough to advise me via this
forum what the result of your investigation is, and also could you tell me what
if any other then the WD1002-05 might be an alternative controller to be used
with the Eliminator. It seems my controller crapped out at the same time that
the Eiminator decided to quit. It might be a case one causing the other to go
haywire, but if so I will have to replace the controller, and Frank tells me he
does not have access to the WD1002-05 any longer. So what next to do.     
Thanks Joerg
There is 1 Reply.

#: 8884 S10/OS9/6809 (CoCo)
    25-Dec-90  14:17:53
Sb: #8812-#Eliminator repair
Fm: MOTD Editor..Bill Brady 70126,267
To: JOERG SATTLER 74016,631 (X)
Joerg, call Frank back and get the number for EMERALD, I think. They will
re-work your WD board for you.
There is 1 Reply.

#: 8914 S10/OS9/6809 (CoCo)
    27-Dec-90  16:29:02
Sb: #8884-Eliminator repair
Fm: JOERG SATTLER 74016,631
To: MOTD Editor..Bill Brady 70126,267
Thanks for the name of "EMERALD".    And a happy new year to you to.   Joerg

#: 8872 S10/OS9/6809 (CoCo)
    24-Dec-90  18:02:14
Sb: Eliminator repair
Fm: JOERG SATTLER 74016,631
To: Bruce Isted (UG VP) 76625,2273 (X)
Thanks for the reply !!!!! And a merry X mass and and a Happy New Year.

#: 8839 S10/OS9/6809 (CoCo)
    21-Dec-90  11:57:22
Sb: Inside OS9
Fm: REX GOODE 73777,3663
To: Kevin Darling
Kevin,
 Where can I get a copy of your book, "Inside OS9"?
   Rex

#: 8841 S10/OS9/6809 (CoCo)
    21-Dec-90  12:20:00
Sb: #Downloading stuff
Fm: Carlyle Hudkins 72040,2754
To: all
   A few people have advised me of files to get to help with things I need to
do.  Text files are no problem--I can read them with my WP.  However, I know
from experimenting that I cannot save a file to an OS9 disk (especially a
40-trk one) from outside OS9 (i.e. my terminal program).  How, then, am I to
use a utility, such as the recommended DMODE.AR, once I've downloaded it?
 Would running my term (Ultimaterm v4.0, pure ML) from a window help, and if so
how do I manage this?
   Is there a text file in a Lib that would explain all this to me?  If so,
point me at it and I'll grab it and chew on it for a while.
                   thanks,
                          Carlyle H /post to:all;sb:Downloading Stuff
There is 1 Reply.

#: 8842 S10/OS9/6809 (CoCo)
    21-Dec-90  13:10:52
Sb: #8841-Downloading stuff
Fm: Pete Lyall 76703,4230
To: Carlyle Hudkins 72040,2754 (X)
Carlyle -
There's a program and documentation in DL10 called DOSOR9.BAS and DOSOR9.DOC.
Grab these, and save them to an RS-DOS diskette. These will allow you to make
an RSDOS diskette readable to OS9. That's one way you can move files over
there.
Another way is to use an OS9 terminal program. You'll NEED a serial port (not
the one in the back of the machine, but a real one like an RS-232 pak, or one
of the 3rd party versions), but it's worth it. Sterm for os9 supports ASCII,
XModem, and B+ protocols, and is the terminal program of choice for os9
Compuserve folks. Sterm and goodies are in DL7.
Pete

#: 8843 S10/OS9/6809 (CoCo)
    21-Dec-90  15:05:08
Sb: #Color of GFX Cursor?
Fm: Zack Sessions 76407,1524
To: ALL
I really don't have time to disassemble WindInt to figure this out,
so if someone familiar with it or with this particular question, I would
appreciate some input.
Basically, what determines which palette slot(s) govern the color of
an autofollow mouse cursor?
What I know is that, the color of the mouse can't stay the same palette
all the time. What if it is moved to an area which is completely filled
by that palette? It would disappear! So WindInt (or somebody) can change
the palette of any part of an autofollow mouse cursor (and does!) as it
chooses. Just how does "it chooses"?
TIA,
Zack Sessions
There is 1 Reply.

#: 8844 S10/OS9/6809 (CoCo)
    21-Dec-90  15:24:06
Sb: #8843-#Color of GFX Cursor?
Fm: Kevin Darling (UG Pres) 76703,4227
To: Zack Sessions 76407,1524 (X)
Zack - the cursor is simply an XOR of the bits on the screen.  Thus the color
depends upon the background, and the palette slot the data XORs out to. 
Nothing fancier than that.
In other words, if your cursor is over a background of palette 0 in a 4-color
screen, then the bits become palette 3.
Of course, if both palettes 0 and 3 are say, white... then you get no visible
cursor <grin>.  Which usually is no concern except in gfx paint programs. Which
is why old Max9 had two-color cursors, so that your chances increased for being
able to see the thing on most pictures.  (the two colors were done by creating
put buffers with mixed bits in them... I guess that wouldn't work ... ummm yes
it should work for autofollow too, i think).
There are 2 Replies.

#: 8846 S10/OS9/6809 (CoCo)
    22-Dec-90  18:39:45
Sb: #8844-#Color of GFX Cursor?
Fm: Joseph Cheek 76264,142
To: Kevin Darling (UG Pres) 76703,4227 (X)
I believe you can create 4- and 16-color mouse pointers.  just use the image
you need and make sure you set the STY type correctly.  should i explain it
better?
There is 1 Reply.

#: 8849 S10/OS9/6809 (CoCo)
    22-Dec-90  20:14:06
Sb: #8846-Color of GFX Cursor?
Fm: Kevin Darling (UG Pres) 76703,4227
To: Joseph Cheek 76264,142 (X)
Joe - I think you're right.  THe mouse cursor is really just a PUTBLK each time
in XOR logic mode... so a different STY type and data should work as you said.

#: 8847 S10/OS9/6809 (CoCo)
    22-Dec-90  18:58:43
Sb: #8844-#Color of GFX Cursor?
Fm: John Ranck 73540,246
To: Kevin Darling (UG Pres) 76703,4227 (X)
        Kevin,
        What if you want ot change the cursor to an underscore. Thats what it
is in school on the Vax and I always like my CoCo to be more like the Vax and
the Vax to be more like my CoCo. I've always wandered if there was a way to
change it. Could you change like you can change mouse pointers or something. Is
it mapped in w/ the fonts as a character or is it in CC3io itself? Is there a
way to change the pattern of the cursor? Is there anything about this in Inside
OS9? I didn't see it so thats why I'm wondering. Its a good book though. Lots
info.
        L8R & Thanx
        Mike
There is 1 Reply.

#: 8850 S10/OS9/6809 (CoCo)
    22-Dec-90  20:30:34
Sb: #8847-Color of GFX Cursor?
Fm: Kevin Darling (UG Pres) 76703,4227
To: John Ranck 73540,246 (X)
Well, I was about to say "sure, piece of cake"... until I looked at that part
of Grfdrv... and it's not just a matter of changing something in place, as I'd
hoped.
Ummm. It's a not a char in the font, no.  The text cursor routine reverses the
attributes on a rom-text screen, and does an XOR of the whole char area on a
gfx-screen... using the normal charput routines instead of a special routine...
making it harder to just patch in place.  (note: they partly did it this way
because the cursor also had to handle proportional text, too - tough one).

#: 8929 S10/OS9/6809 (CoCo)
    28-Dec-90  12:45:34
Sb: #bonk etc.
Fm: Joseph Cheek 76264,142
To: all
Now that you all have had a while to look at it, does anyone have any comments
for my bonk.pak upload?
 while i'm at it, does anyone have a good random-number generator in C?  I
wrote one, but it's SLOW.  anyone?  anyone?  thanks . .  .
There is 1 Reply.

#: 9000 S10/OS9/6809 (CoCo)
    02-Jan-91  18:27:42
Sb: #8929-bonk etc.
Fm: Kevin Darling (UG Pres) 76703,4227
To: Joseph Cheek 76264,142
Question - what is the View command that's needed?  Is it available here?
  thx! - kev

#: 8936 S10/OS9/6809 (CoCo)
    29-Dec-90  11:25:46
Sb: #Format
Fm: REX GOODE 73777,3663
To: All
What are the differences, if any, between an OS9 FORMAT and DECB's DSKINI0?
  Rex
There is 1 Reply.

#: 8944 S10/OS9/6809 (CoCo)
    29-Dec-90  20:09:34
Sb: #8936-#Format
Fm: Kevin Darling (UG Pres) 76703,4227
To: REX GOODE 73777,3663 (X)
Rex - if there are any differences, they can't be much.  A lot of us will
format under one OS, and then write the needed info to make the disk work under
the other OS.
The biggest diff might be related to writing... L-II/coco no longer uses
precompensation because they say that it's not needed with modern equipment and
disks.  Doesn't seem to matter much tho.
There is 1 Reply.

#: 9004 S10/OS9/6809 (CoCo)
    03-Jan-91  12:12:15
Sb: #8944-Format
Fm: REX GOODE 73777,3663
To: Kevin Darling (UG Pres) 76703,4227 (X)
Kevin,
 Thanks for the tip. Suddenly I have been getting errors when formatting with
OS9, yet formatting the same disk under DECB seemed OK. I'll try using the DECB
disks.
  Rex

#: 8937 S10/OS9/6809 (CoCo)
    29-Dec-90  14:54:47
Sb: Osterm
Fm: NEAL STEWARD 72716,1416
To: ALL
Can anyone tell me how the ASCII upload works with OSTerm v.2.0.8?  I input a
prompt (that is sent by a local BBS) and the program just idles without
attempting to send a file.

#: 8951 S10/OS9/6809 (CoCo)
    30-Dec-90  09:34:01
Sb: patches
Fm: Hugo Bueno 71211,3662
To: All
I missed the Rainbow article by Bruce Isted concerning clock patches and such.
Would anyone care to give me a synopsis of article?
Hugo

#: 8953 S10/OS9/6809 (CoCo)
    30-Dec-90  09:41:21
Sb: #bonk etc.
Fm: Hugo Bueno 71211,3662
To:  76264,142 (X)
I like Bonk a lot. The only thing that takes getting used to is mouse movement.
Also, I'm never sure how the ball will bounce off the ship. It seems random.
I also noticed that the ball is surrounded by a square. Is there any way to XOR
or whatever, to just have the ball without an outline?
Hugo
There is 1 Reply.

#: 8972 S10/OS9/6809 (CoCo)
    31-Dec-90  12:31:21
Sb: #8953-#bonk etc.
Fm: Joseph Cheek 76264,142
To: Hugo Bueno 71211,3662 (X)
the ball bounces at a 45 deg angle if it hit the ship on the edges, or at a 23
deg angle if it hit in the middle.  also, 1/4 of the time if bounces back the
same direction.
  re: ball with outline.  I'll see what i can do.
There is 1 Reply.

#: 8993 S10/OS9/6809 (CoCo)
    01-Jan-91  21:49:18
Sb: #8972-#bonk etc.
Fm: Hugo Bueno 71211,3662
To: Joseph Cheek 76264,142 (X)
As far as ball bouncing, it would be better to be able to put "english" on the
ball.
Also, I noticed when I quit the game, the VIEW utility is not unlinked from
memory. You should probably make sure that all graphics buffers are also
KILLed.
Hugo
There is 1 Reply.

#: 8999 S10/OS9/6809 (CoCo)
    02-Jan-91  16:13:27
Sb: #8993-#bonk etc.
Fm: Joseph Cheek 76264,142
To: Hugo Bueno 71211,3662 (X)
english?  please repeat . . .
There is 1 Reply.

#: 9002 S10/OS9/6809 (CoCo)
    02-Jan-91  18:47:45
Sb: #8999-bonk etc.
Fm: Hugo Bueno 71211,3662
To: Joseph Cheek 76264,142
Well, "english" a term used in billiards, means putting a spin on the ball. In
video games, it means if the racquet is travelling to the left when the ball is
hit, then the ball will also tend to travel to the left. Many versions of
BREAKOUT had this feature.
Hugo

#: 8967 S10/OS9/6809 (CoCo)
    31-Dec-90  09:25:13
Sb: Wanted to buy
Fm: Paul Rinear 73757,1413
To: All
     Am desperately seeking a copy of OS-9 Level One, used or
   otherwise; preferably with docs. Missed the tent sale at
   Radio Shack and now there are no copies to be found. If you
   have one for sale, please send E-mail to:
                                Paul Rinear
                                73757,1413

#: 8970 S10/OS9/6809 (CoCo)
    31-Dec-90  10:49:18
Sb: Osterm208
Fm: Chris Bergerson 72227,127
To: Neal Steward 72716,1416
Neal,
     I met Vaughn at a party a few weeks ago, and asked him about
the osterm ASCII upload problem.  I had tried many times to get
it to work.  Well, I won't bother trying anymore... he said that
it never did work, and that he has no intentions of fixing it!
     There is a rumor that the source might be available for the
upload sections of the program... I'm investigating, and if true,
I will probably fix it.  In the meantime, to accomplish the same
goal, I'd recommend flipping to another window, and outputting
your text file to /t2, using a simple utility to send a
character, pause, send a character, etc.
     BTW, good to see you here!

#: 8982 S10/OS9/6809 (CoCo)
    01-Jan-91  02:54:50
Sb: #MultiVue
Fm: NAM PUI 73347,3324
To: all
I   got MultiVue since it hit the market. However, after the novelty worn  off
after  the first month, I never bother with it again. Since then I have  added
hard  drives  to all my COCO IIIs. I have been getting along fine without  the
use of MultiVue. Just days ago I was going through the drives to clean out the
garbage  in  my  data files. I thought it would be great to be able to  use  a
point  and click enviroment to view and delete the unwanted files. So out come
Multivue  again.
I have been pulling my hair off since. Here is my system hardware set up. 512K 
COCO  III  with  J&M JFD-CP floppy controller  and  Owlware  hard  drive
interface/Omti  5200  controller  on  a  Y-Cable.  Seagate  St225  and  3-720k
floppy(2-5.25" and 1 3.5").
Software set up is as follow.
   Module Directory at 03:28:41 REL         Boot        OS9p1       OS9p2      
Init RBF         CC3Disk     D0          D1          D2 CCOmti      dd         
h0          IOMan       SIO SCF         CC3IO       WindInt     Term        W
W1          W2          W3          W4          W5 W6          W7         
PRINTER     P           CPPRINT P1          Clock       CC3Go       GrfDrv     
Shell Cat         Dir         Display     Del         gotoxy ds          MDir
CC3Disk is the version patched for msdos ccomti.dr, dd, h0 are the Owlware
driver and descriptors Term starts up in 80 column. P1, CCPRINT are for the
parallel port on the J&M.
All  the /d0s in the autoex file is patched to /dds. Autoex renamed to Mv  for
manaul start up.
Now,  here  is the problem. Everytime I move from the root directory to a  sub
directory  on  the  hard  drive,  the window in mv  just  hangs  in  the  hour
mode(wait).  It seem to have effect the time share as well. The other  windows
slowed down as well. The whole things work well in the floppys.
Hope someone can help.
Thanks in advance.
Nam
There are 2 Replies.

#: 8984 S10/OS9/6809 (CoCo)
    01-Jan-91  10:27:59
Sb: #8982-#MultiVue
Fm: Zack Sessions 76407,1524
To: NAM PUI 73347,3324 (X)
Try downloading the GShell+ patches and apply them. They fixed some bugs which
could cause what you're seeing.
There is 1 Reply.

#: 9008 S10/OS9/6809 (CoCo)
    03-Jan-91  21:37:32
Sb: #8984-MultiVue
Fm: NAM PUI 73347,3324
To: Zack Sessions 76407,1524 (X)
Thanks. I am heading for DL10 right now.
Nam

#: 8985 S10/OS9/6809 (CoCo)
    01-Jan-91  11:02:36
Sb: #8982-#MultiVue
Fm: Kevin Darling (UG Pres) 76703,4227
To: NAM PUI 73347,3324 (X)
Hmm. I think this can happen if you have lots of files and some 3-letter
extensions.  At the least, apply the MVFIX.SCR patch to gshell (from Lib 10)..
that should help.
If you do a "bro /key:gshell" there you'll find more patches.  Some people
began to use it after the patches ;-).
There is 1 Reply.

#: 9007 S10/OS9/6809 (CoCo)
    03-Jan-91  21:36:06
Sb: #8985-MultiVue
Fm: NAM PUI 73347,3324
To: Kevin Darling (UG Pres) 76703,4227 (X)
Thanks. I will give it a shot.
Nam

#: 8991 S10/OS9/6809 (CoCo)
    01-Jan-91  18:49:04
Sb: disto&hardrv
Fm: KENHEIST 71750,551
To:  76703,4224
kevin I'm using a disto scii w/ 4n1 and just got a Seagate 157n. I keep getting
a 247 error when I bootup with ddd0 and do a chx /h0/cmds it takes two or more
tries to get it to catch and of course if I try to do a ddh0 I get bootfail
every time. Whats up? Used the hmode to change the cyls and hds and inittbl.
i.e.
  cyls=615 hds=6 inittbl=026706 installed h0_4in1scsi.dd and cchdisk_scsi.dr.
Did I miss something?

#: 9003 S10/OS9/6809 (CoCo)
    02-Jan-91  23:56:38
Sb: CoCo 3 discontinued
Fm: James Jones 76257,562
To: All
Well...after seeing an earlier message posted here about the change in the
status of the CoCo 3 (from "active" to "discontinued"), I went over to a local
RS store and asked the fellow there to look it up.  At that time
(mid-December), his system claimed it was "active."
Yesterday, I got back from spending the holidays with family, and found among
the messages on the answering machine one from the same fellow, saying that the
CoCo 3 had indeed changed status to "discontinued," and he had one left at the
store--so I'd better come on over if I wanted it.
So...perhaps this store got the information later than others, but I guess it's
true.

#: 9031 S10/OS9/6809 (CoCo)
    05-Jan-91  15:58:50
Sb: #VDG windows
Fm: Denise Tomlinson 71021,3274
To: all
Is there a way to have 2 vdg windows with shells and 2 programs running in
them? I know about /w1 /w2 and soforth, and how to use the shell command and
the <clear> key. The two programs I want to run at the same time use the
standard 32/16 /term screens. I have to be operating out of that type of screen
to boot the program. I try booting out of /w1 but it gives me a "no vdg window"
error. Thanks, Denise
There is 1 Reply.

#: 9039 S10/OS9/6809 (CoCo)
    06-Jan-91  02:24:40
Sb: #9031-VDG windows
Fm: Kevin Darling (UG Pres) 76703,4227
To: Denise Tomlinson 71021,3274 (X)
Yes, assuming you have enough system map space (for the text vdg windows), you
can pick a descriptor and for example:
 xmode /w6 type=1
 shell <>>>/w6&
The "type=1" sets it as a vdg type window default.  Other people sometimes make
up a set of renamed descriptors (/v0, /v1, etc) just for this purpose (there
may even be a set in Lib 10 here).
But the xmode will work fine.  Be sure to xmode back to type=80 afterwards...
just in case later on when you exit your program, an open to /w doesn't
accidentally find and use /w6 (or whatever) as a vdg window.  I'm rambling, but
I think you get the gist. <grin>  kev

#: 9036 S10/OS9/6809 (CoCo)
    05-Jan-91  23:57:55
Sb: #pmap!
Fm: Everett Chimbidis 76370,1366
To:  76703,4227 (X)
Where do I find pmap?
There is 1 Reply.

#: 9037 S10/OS9/6809 (CoCo)
    06-Jan-91  01:06:23
Sb: #9036-pmap!
Fm: Dan Robins 73007,2473
To: Everett Chimbidis 76370,1366 (X)
Everett,
   It (PMAP) is located in Kevin Darling's UTIL3.BIN upload, which is located
in LIB 10. You can:   BRO UTIL3.*      and it should pop up.
   Dan

#: 9038 S10/OS9/6809 (CoCo)
    06-Jan-91  01:10:31
Sb: #FD501 power transformer
Fm: Dan Charrois 70721,1506
To: all
Just today the primary for the transformer in my disk drive died.  The markings
on the transformer are pretty well obscure so I was wondering if anyone could
clue me in on what voltage it puts out.  I know it is tapped in 4 places, but
that's about all I can come up with when it doesn't work!
  Hopefully I can pick one up for cheaper than a whole power supply...
  Thanks for your replies
                                   Dan
There is 1 Reply.

#: 9086 S10/OS9/6809 (CoCo)
    11-Jan-91  00:56:39
Sb: #9038-FD501 power transformer
Fm: Wayne Day 76703,376
To: Dan Charrois 70721,1506
A disk drive power supply should be putting out +12v dc and +5v dc regulated,
so... somewhere in the area of 13v and 6v respectively?
Wayne

#: 9048 S10/OS9/6809 (CoCo)
    06-Jan-91  19:36:14
Sb: Disto Hard Drive
Fm: james pottage 71750,2012
To:  71750,551 (X)
The problem you are encountering with the disto 4 in 1 board and the seagate
157n is the same problem that I encountered with my St125n drive. Ken Scales
has written a patch for the disto drivers that cure this problem and also allow
formatting of the hard drive under OS9. You can contact Ken by leaving him a
message on compuserve, although it would be quicker if you have access to Delfi
and left him a message there.

#: 9052 S10/OS9/6809 (CoCo)
    07-Jan-91  00:31:32
Sb: Eliminator repair
Fm: JOERG SATTLER 74016,631
To:  76625,2273
Hi there Bruce! I hope that there has been some progress toward the resolution
of the problem with my Eliminator board in the meantime. I would just love to
get back up and runnning if at all possible. please let me know as soon as
possible.
               Thanks for your effort in resolving this little problem.
oerg a
                                              Joerg Sattler 74016,631

#: 9075 S10/OS9/6809 (CoCo)
    10-Jan-91  01:11:31
Sb: #data base
Fm: Everett Chimbidis 76370,1366
To: all
What the Best data base for the coco in os9?? AND where can i get it?
There are 3 Replies.

#: 9076 S10/OS9/6809 (CoCo)
    10-Jan-91  07:15:49
Sb: #9075-#data base
Fm: Ed Gresick 76576,3312
To: Everett Chimbidis 76370,1366 (X)
Everett,
We sell SCULPTOR by MPD for the CoCo.  We think its the best. Frank Hogg sells 
IMS by Clearbrook - he thinks that is the best. There may be some others, but
I'm not familiar with them.
Ed Gresick - DELMAR CO 302-378-2555
There is 1 Reply.

#: 9079 S10/OS9/6809 (CoCo)
    10-Jan-91  08:03:07
Sb: #9076-#data base
Fm: Steve Wegert 76703,4255
To: Ed Gresick 76576,3312 (X)
Ed,
I'm curious (really!) ... what's the going price for a CoCo version of Sculptor
these days? I've lost track.
Steve
There is 1 Reply.

#: 9087 S10/OS9/6809 (CoCo)
    11-Jan-91  03:18:23
Sb: #9079-data base
Fm: Ed Gresick 76576,3312
To: Steve Wegert 76703,4255 (X)
Steve,
Current price for CoCo Version (version 1.16) of SCULPTOR is $260.00.
Ed Gresick - DELMAR CO

#: 9077 S10/OS9/6809 (CoCo)
    10-Jan-91  07:26:27
Sb: #9075-data base
Fm: JIM HICKLE 76672,602
To: Everett Chimbidis 76370,1366 (X)
Never used a commercial DB on the coco; If you wish to roll your own, check out
Al Stevens' "C Database Development" (MIS Press).  It has a relational
database, b-tree indexing and other good stuff.
     -jim

#: 9078 S10/OS9/6809 (CoCo)
    10-Jan-91  08:02:06
Sb: #9075-data base
Fm: Steve Wegert 76703,4255
To: Everett Chimbidis 76370,1366 (X)
Everett,
I've been monkeying around with IMS from Clearbrook, and find it facinating.
I've had it for years ... and finally got around to setting up a few databases
to help keep me organized (That's a full time job in itself!).
Steve

#: 9089 S10/OS9/6809 (CoCo)
    11-Jan-91  12:01:23
Sb: #mroff & Kreider docs
Fm: Paul Rinear 73757,1413
To: Pete Lyall
        I downloaded MROFF and was able to print the mroff docs
  using mroff. It worked nicely. When it came time to print the
  Kreider C-lib docs, mroff gave out scads of unrecognized
  command errors and no results. The command in question seems
  to be .de which is not mentioned in the docs. What am I doing
  wrong ?
                            Paul R.
There is 1 Reply.

#: 9090 S10/OS9/6809 (CoCo)
    11-Jan-91  15:32:10
Sb: #9089-#mroff & Kreider docs
Fm: Pete Lyall 76703,4230
To: Paul Rinear 73757,1413 (X)
The MROFF docs were prepared using a 'flavor' of Mroff created by Mark
Griffith. You need Mark's copy, which _should_ be in DL6 or DL9.
Pete
There is 1 Reply.

#: 9092 S10/OS9/6809 (CoCo)
    11-Jan-91  16:03:56
Sb: #9090-#mroff & Kreider docs
Fm: Paul Rinear 73757,1413
To: Pete Lyall 76703,4230 (X)
      I can't seem to find it in either library using keywords of
  "mroff" or "print". Will look some more.
There is 1 Reply.

#: 9093 S10/OS9/6809 (CoCo)
    11-Jan-91  16:32:24
Sb: #9092-#mroff & Kreider docs
Fm: Paul Rinear 73757,1413
To: Paul Rinear 73757,1413 (X)
      Still can't find it. The description of "clibdo.ar" in lib 3
  says Mark Griffith's mroff is included. Upon de-arcing, only the
  C-source for mroff is there. I could probably compile it but I
  am missing one DEF file; scfstat.h .
                                         Paul
There is 1 Reply.

#: 9097 S10/OS9/6809 (CoCo)
    11-Jan-91  21:38:35
Sb: #9093-mroff & Kreider docs
Fm: Pete Lyall 76703,4230
To: Paul Rinear 73757,1413
Paul -
Drop a note to Steve Wegert.... I believe he's in daily digital contact with
Mark's system, and could probably have mark mail the file.
Pete

#: 9105 S10/OS9/6809 (CoCo)
    12-Jan-91  21:02:11
Sb: #9076-#data base
Fm: Greg Law 72130,23
To: Ed Gresick 76576,3312 (X)
Ed,
        In the case of Sculptor versus IMS, the issue of 'best' is extremely
particular to the sections you are speaking of. Overall, I think Sculptor is
the better database engine and it has a lot more functions than IMS. On the
other hand, I think IMS was implemented better. Where Sculptor generates
tokenized files in the current data directory, IMS creates executable code in
the current execution directory. All you need is a simple 4-byte patch to Shell
(Type/Lang=$1B, run-time module=imsi) and you can run IMS database applications
directly. I.E., I run 'menu' for a database I developed and Shell turns that
into 'imsi menu' and gets me up and running. Sculptor requires a little more
manual labor. IMS also came with a screen-form generator (screen painter) and
the original version of Sculptor didn't. I don't know if one has been developed
for Sculptor yet.
        In my humble opinion, I'd like the language features of Sculptor merged
into the rest of IMS. I think Sculptor is more powerful, but IMS is easier to
set up and get going.
        -- Greg
There is 1 Reply.

#: 9138 S10/OS9/6809 (CoCo)
    14-Jan-91  12:41:33
Sb: #9105-#data base
Fm: Ed Gresick 76576,3312
To: Greg Law 72130,23 (X)
 Greg,
 Whoops!  You threw me a curve ball!!!  Forgot all about that trick.  It's
 very useful and should be used more often.  Maybe you should write an
 article about this.  Also, this might induce some vendors (MPD) to change
 the header of their compiled programs to permit this kind of execution.
 (Jeez, making me look things up!)  I told John Dickey about it.  John was
 heavily into IMS and is converting over to SCULPTOR.  He hadn't heard about
 it but tried it after we figured out where and how to patch 'shell'.  Worked
 just fine.
 We were curious whether we could do the same thing with the SCULPTOR programs
 'sage' and 'sagerep'.  The SCULPTOR compiled programs ___.g and ___.q are
 simply data files as far as OS9 concerned so we had to fake it.  We made a
 dummy data module named with the program we want to run.  Patched 'shell'
 and it worked.  The actual sequence is -
    shell looked for the program in memory - data module was there with
       that name.
    shell found a module name associated with the type byte in this case
'sage'.
    shell called 'sage' and passed the program name to it.
    sage looked for that program in the CWD with a .g extension and then
       executed it.
 Interesting experiment - value ????
 I agree with most of what you said (except I think the SCULPTOR 'menu'
 program is better and easier to use but ...).  One point I feel should also
 be made (sorry MPD); CLEARBROOK's IMS is easier to learn.  As I recall, their
 tutorial is very good.  The documentation provided by MPD for CoCo SCULPTOR
 leaves much to be desired;  a terse reference manual with a very sparse
 'tutorial' section.  The reference manual for version 2.0 of SCULPTOR has
 been much improved and I think is pretty good.  MPD is supposed to be working
 on a tutorial but it's not available yet.
There are 2 Replies.

#: 9139 S10/OS9/6809 (CoCo)
    14-Jan-91  12:42:50
Sb: #9138-#data base
Fm: Ed Gresick 76576,3312
To: Ed Gresick 76576,3312 (X)
 --- Continued ---
 The latest version of SCULPTOR for the CoCo is 1.16.  MPD has no plans to
 update it.  The official reason is that the size of the code for version 2.0
 will not fit into the 64K work-space allowed by the 6809.  IMO, I believe the
 real reason is that they don't feel the market is large enough to justify the
 work necessary for the 2.0 port and that CoCo users will not pay what they
 (MPD) think is a fair price.  Also IMO, when GIMIX and others (not Tandy)
 stopped making 6809 machines, MPD felt the 8-bit market was dead.  Incidently,
 the CoCo version is basically the same as the 6809 version with the addition
 of a vdu descriptor for the CoCo.
 SCULPTOR version 2.0 does include a screen painter, improved utilities and
 more functions and commands.  It also provides a 'windowing' environment
 within 'sage' (up to 8 screens per program).  A pre-proccessor has been added
 which allows all kinds of neat things including conditional compilation, the
 ability to include library files, manifest constants and some additional
 commands.
 I do have a copy of CLEARBROOK's IMS but it goes back a few years.  I had
 considered it for my purposes.  I selected MPD's SCULPTOR because of its
 portability, acceptance by the business community and my perception of its
 power.  Software written in SCULPTOR and compiled for the CoCo will run on
 almost any OS and machine including mainframes using the appropriate runtime
 module.  As far as I know, no other DBM system runs on such a wide variety of
 platforms without recompiling.  And its performance is respectable - holding
 its own or superior to other systems (Oracle, Informix, SyBase, etc.).  Also,
 it is easier to program in than the others.
 Ed Gresick - DELMAR CO
There is 1 Reply.

#: 9141 S10/OS9/6809 (CoCo)
    14-Jan-91  20:41:17
Sb: #9139-#data base
Fm: Dan Robins 73007,2473
To: Ed Gresick 76576,3312 (X)
Ed,
  RE: "no other DBM system runs on such a wide variety of platforms.....
  Although it's not on OS9...I'm impressed with Progress! Are you familiar with
it?
  Dan
There is 1 Reply.

#: 9161 S10/OS9/6809 (CoCo)
    16-Jan-91  04:52:53
Sb: #9141-#data base
Fm: Ed Gresick 76576,3312
To: Dan Robins 73007,2473 (X)
Dan,
I've heard of it - good reports.  No, I've never used it.  None of my customers
use it.
Ed Gresick - DELMAR CO
There is 1 Reply.

#: 9162 S10/OS9/6809 (CoCo)
    16-Jan-91  06:32:30
Sb: #9161-#data base
Fm: Dan Robins 73007,2473
To: Ed Gresick 76576,3312 (X)
Ed,
  It runs on many platforms (unfortunatly, not OS9 or OSK), but having both
Sculptor and IMS, in comparison, it's so much simplier and much more powerful!
  Of course, that's my opinion...but then you know what people say about
opinions, dontcha? <grin>
   Dan
There is 1 Reply.

#: 9171 S10/OS9/6809 (CoCo)
    17-Jan-91  02:40:15
Sb: #9162-#data base
Fm: Ed Gresick 76576,3312
To: Dan Robins 73007,2473 (X)
Dan,
How does it compare with either Informix or Oracle?
Ed
There is 1 Reply.

#: 9172 S10/OS9/6809 (CoCo)
    17-Jan-91  05:19:13
Sb: #9171-data base
Fm: Dan Robins 73007,2473
To: Ed Gresick 76576,3312 (X)
Ed,
  Can't speak on Informix, as I've seen it demonstrated only once, and then, it
wasn't much of a demo.
  However, in comparison to Oracle, both the setting up of the database (and
linking in to other produced relational database files) and it's ability to
make both viewable & printable reports could roughly be done in about half the
time.
  To be honest, when I saw Progress being demoed and then with assistance (a
Progress programmer looking over my shoulder) being able to create a simple
database, I was in awe at the speed that I could do it. My exposure to the
product was not by any salesman but a guy who works for a company that develops
under the system.
  It's an awesome system...if you can take a gander at it...it's worth the
time!
   Dan

#: 9183 S10/OS9/6809 (CoCo)
    18-Jan-91  19:27:38
Sb: #9138-data base
Fm: Greg Law 72130,23
To: Ed Gresick 76576,3312 (X)
Ed,
        I can't argue with that. IMS is easier to learn than Sculptor, if for
no other reason than the included tutorials. I always wondered why MPD decided
to go the data file route for its database applications, especially since the
modular organization of OS-9 makes it very easy to create new executable
modules with a known run-time module. It soon dawned on after seeing the MS-DOS
version of Sculptor. I had forgotten that MPD is offering Sculptor for so many
different platforms, including OS-9, Unix, MS-DOS, and others.
        Even so, I think MPD can do a much better job with the manuals. Of
course, the reference manual should be included. Unfortunately, a reference
manual doesn't tell you diddly-squat about how to actually use the program or
to develop applications. It also comes with so many executable files that
you're CMDS directory will be swimming with files that you'll quickly forget
where they came from. I wonder if it would have been better to split the CMDS
directory on the distribution disks to the master executable in CMDS and the
maintenance/reorganization stuff in the CMDS/MAINTENANCE or CMDS/SCULPTOR
directory.
        -- Greg

#: 9107 S10/OS9/6809 (CoCo)
    12-Jan-91  21:47:07
Sb: #9097-mroff & Kreider docs
Fm: Paul Rinear 73757,1413
To: Pete Lyall 76703,4230 (X)
     Thanks, will do.

#: 9110 S10/OS9/6809 (CoCo)
    12-Jan-91  22:34:17
Sb: Renegade Windows
Fm: Paul Rinear 73757,1413
To: Kevin Darling
       Something I've been afraid to ask about for a long time and
    have just been tolerating.
       I have a Coco3 running OS9 level II, 1 Meg upgrade, and
    three double-sided floppies, RS-232 Pak, and CM-8 Monitor.
    After the 1 Meg is activated, I load View, Viewgif, and
    sometimes Max09 (Is that one of yours?). Both these view
    programs have modes that permit rapid switching of images
    to simulate more than 16 colors or 4 gray shades.
       Say I'm watching a picture that contains 2 vef files
    alternating in this manner, and then I break the program
    to stop it. Once in awhile one of the images will remain
    in memory somewhere (you can see it flash by when hitting
    the clear key to change windows, but you can't stop on it).
    When this happens, an attempt to view another 2-file image
    with either view or viewgif will result in a 'window already
    defined' error.
       Using 'procs' you can see there is no process running
    keeping this window active. Similarly, 'deinizing' the
    windows doesn't matter.
       More detail: I boot with a vdg TERM, and define W1 type 2,
    W2 type 8, and W3 type 7. Running a viewer program from one
    of the graphics windows will sometimes turn W1 into the
    renegade window described above.
       The only way I've been able to view multi-window pictures
    again after this happens is to reboot. Got any ideas on
    what is happening here ?
                                          Paul

#: 9115 S10/OS9/6809 (CoCo)
    13-Jan-91  15:56:30
Sb: #9052-#Eliminator repair
Fm: Bruce Isted (UG VP) 76625,2273
To: JOERG SATTLER 74016,631 (X)
Joerge,
  I meant to get on CIS and tell you a while ago, but I just haven't had the
time... anyway, your Eliminator has been repaired, and you should receive it
any day now.  The WD 1002-05 interface was pretty much popped, as was the Tx
data line on one of the serial ports.  Do you have an extra tough static
problem there?  Thats the only thing I can think of that'd cause the damage I
saw.  Anyway, not to worry, just don't let it happen again, OK?  <grin>
  Bruce
There is 1 Reply.

#: 9117 S10/OS9/6809 (CoCo)
    13-Jan-91  17:27:05
Sb: #9115-Eliminator repair
Fm: JOERG SATTLER 74016,631
To: Bruce Isted (UG VP) 76625,2273
There is some occasional static but I dont use T1 for outside (modem)
connections, only for remote terminal service. Could you gine me the name of
the HD controller repair again ?. Seen to have lost the original notes I made
at the time.    Thank you.                  Joerg

#: 9142 S10/OS9/6809 (CoCo)
    14-Jan-91  21:37:35
Sb: #Gshell
Fm: NAM PUI 73347,3324
To: Kevin Darling
Thanks   Kevin,   and Zack Sessions. After reading the messages from  both you 
I searched the library and downloaded all patches related to  Gshell. Applied
most of it to get gshell+ 1.24a. It worked great. No more hangups. With the new
Grfdrv. It is fast.
Now I am getting greedy. Is there a patch to get a 80*24 window (or 40*24)
after  you  click on a file and click on list to lis the file  to  screen?
Also,  Is there a cc3io patch to patch the cc3io patched for the  Dynastar
keytable.  I  use  the  <ctrl>-<:> to get the cursor keys  to  work  under
Dunastar  a  lot.  I  also  liked the ability to use  the  button  on  the
joystick/mouse to move the screen.
One more thing, does the serial mouse driver work with MV?
Thanks again, guys, for all the helps.
Nam
There is 1 Reply.

#: 9160 S10/OS9/6809 (CoCo)
    15-Jan-91  23:20:56
Sb: #9142-#Gshell
Fm: Kevin Darling (UG Pres) 76703,4227
To: NAM PUI 73347,3324
Nam - glad to help, and even gladder that you managed to apply all the patches
<grin>!  A bunch, eh?
Don't think there's a DS key patch (I could be wrong).  No patch yet for a
bigger list window either.
The serial mouse driver should work, I'd think... the mod should be to CC3IO,
which means Windint and Grfdrv are left alone as you've patched them. Stay
"greedy", btw... keeps us on our toes <hehe>. best - kev
There is 1 Reply.

#: 9192 S10/OS9/6809 (CoCo)
    19-Jan-91  07:29:05
Sb: #9160-Gshell
Fm: Mike Passer 72750,420
To: Kevin Darling (UG Pres) 76703,4227
~ Kevin,
    If when you mentioned "DS key patch" (I didn't get the whole thread) you
meant a patch for Dynastar keys, there is indeed one in the DL's from Ken
Scales.  It works quite well, and Ken is a heck of a guy!
Mike

#: 9170 S10/OS9/6809 (CoCo)
    17-Jan-91  00:31:53
Sb: #memory
Fm: Everett Chimbidis 76370,1366
To:  76703,4227 (X)
Any help puting boot strap in the right order? I have pmap and dirm but don't
know how to use them! Can I rearange the mods in memory then cobble back to
disk? Want to make the most out of my memory (running bbs) sometimes errors
out!
Any help?
There is 1 Reply.

#: 9177 S10/OS9/6809 (CoCo)
    18-Jan-91  17:31:05
Sb: #9170-memory
Fm: Bill Dickhaus 70325,523
To: Everett Chimbidis 76370,1366 (X)
The only way to reorder your boot file is to use os9gen or an equivalent
utility. Its just a matter of making up a file of all system modules (an "ident
-s" of OS9Boot is one way to get the current list of modules) and redirecting
the file as input to os9gen.
If you are having system memory problems, rearranging your boot won't help. Any
(or all) of the following will help. Unless you really need it for something,
remove VDGInt from your boot. Remove any unecessary device descriptors (if your
just running the BBS you will only need a couple of window descriptors, for
example). If you are using serial drivers with large buffers, make sure you
really need all the buffer space (which comes from system memory). Keep the
number of processes to a minimum, don't start up a bunch of shells if you don't
really need them. Keep the number of open paths to a minimum, you can check
this with Kevin's paths command. Make sure you don't have any duplicate modules
in your OS9Boot file (this can't happen if you use cobbler to create the boot
file). Make sure that both GrfInt and WindInt are not in your boot, and unless
you need WindInt functions, you could use GrfInt (which is smaller).
Anything you can do to reduce the size of the OS9Boot file without eliminating
required functions will help.
You can use Kevin's smap utility to display current system memory utilization.
The pmap utility will show which processes are using how much memory, and how
much (in 8K chunks), the system process (PID 1) is using. The dirm utility is
similar to "mdir e", but includes additional information. It is invaluable in
seeing where in memory a system module is relative to another system modules
(and therefore helps in shuffling modules around to fix the BLOB).
Bill

#: 9174 S10/OS9/6809 (CoCo)
    18-Jan-91  16:48:18
Sb: #Link shell command
Fm: Denise Tomlinson 71021,3274
To: All
Is it necessary to use a "link shell" command in the startup file? Also, if I
want to autoboot a os9 program that is on my system disk, what is the name of
the file I must use? I tryed booting "ultimusic" by inserting the line in the
startup file but it bombs out. If I type "ulitmusic <enter>" after the os9
prompt everything works fine. I read somewhere that I have to rename
"ultimusic" something else to autoboot it? Thanks, Denise
There are 2 Replies.

#: 9178 S10/OS9/6809 (CoCo)
    18-Jan-91  17:38:33
Sb: #9174-Link shell command
Fm: Kevin Darling (UG Pres) 76703,4227
To: Denise Tomlinson 71021,3274
Denise - no need to "link shell" in the startup file... but it speeds things
up.  The CC3Go program first runs "shell startup", and then "shell". So linking
itself in startup saves having to load it again.
Trick to starting programs in scripts (such as Startup): their standard input
is the script itself, which means normal window calls using path #0 won't work
in the programs and they stop.
So you redirect the stdinput in the script to be the same as the std output
(which is usually the window), like this:
 ultimusic </1
The "<" redirects std input. The "/1" means from stdoutput (path 1). You might
have had to do this same trick with setime "setime </1". Let me know if this
doesn't work. best - kev

#: 9179 S10/OS9/6809 (CoCo)
    18-Jan-91  18:01:27
Sb: #9174-Link shell command
Fm: Bill Dickhaus 70325,523
To: Denise Tomlinson 71021,3274
Its only necessary to put "link shell" in your startup file if you are going to
use the shell after booting. To boot directly into a program, rename it to
"autoex", and make sure its in the current execution directory. This may not
always work for all programs, but its easy to try!
Bill

#: 9185 S10/OS9/6809 (CoCo)
    18-Jan-91  22:16:38
Sb: #12-bit timer: help!
Fm: PaulSeniura 76476,464
To: all
I am thinking about using the 12-bit timer in the CoCo3 in a Timer Driver or
SVC for OS9 Level 2.  It could help the Standard MIDI File player and other
projects to do critical timing more easily.
I came across some recent messages on Delphi by Eddie Kuns that relates to how
the Sleep SVC isn't accurate when multitasking.  This might explain why Mike
Knudsen's UME is bogging down at times during thick (vertical) parts of the
score (even after he rewrote the Play module for more efficiency).  I had asked
Knudsen way back in 1989 to consider precompiling the score before playing it,
to offload some of the "gathering" logic beforehand.  But using the hardware
timer in the CoCo3 would be much more accurate and easier, permitting the
application to carry on to generate the next chunk of notes etc.
Has someone already come up with an OS9 Driver/SVC for the 12-bit timer
register in the ACVC/GIME?  If so, could you document how to use it and how to
order/buy/download it?  (I'll gladly go on to the next step in developing our
SMF Player.)
If not, have any of the OSK developers come up with a similar function for the
MM/1, Tomcats, Delmar, etc., machines?  Again if so, could you document how to
use it so I can make our timer somehow compatible?  This would encourage OSKers
to make a CoCo3/OS9 version of their software, too.  Hint hint.
In the meantime, I'll type my notes & thoughts of how I will write a Timer
Driver in an uploaded text file TMRDRV.TXT in the CoCo/OS9 section.  I'm hoping
with your inputs we can make a standard one with enough functionality to let
Knudsen and others adopt it.  (I'll keep these message threads separate from
the other stuff on SMFs for sanity's sake.)
-- Thx, Paul Seniura.
(CIS: 76476,464.  Delphi: PAULSENIURA.  GEnie: PSENIURA.)
                //^^^^^^^^^^^^^^^^^^^\\
               <preferred contact point>
There is 1 Reply.

#: 9190 S10/OS9/6809 (CoCo)
    19-Jan-91  00:33:58
Sb: #9185-12-bit timer: help!
Fm: Kevin Darling (UG Pres) 76703,4227
To: PaulSeniura 76476,464
I believe it's pretty easy to use the timer, Paul.  Just write a 12-bit
countdown value to the timer (forgot address: read gime docs here), turn on the
IRQ enable bit, and that's it.  It reloads and interrupts each time it passes
down through 0000.  Oh, and you can set the counter freq in the GIME to either
horizontal scan rate (15khz) or something very much faster (again, check the
docs.  GIME.TXT I think?).
Long time ago I did an IRQ-driven timer-based sound output driver (crude
effort)... and found that around 3KHz the interrupts overwhelmed the coco. (I
could be wrong about rate: long time ago).
For MIDI use, you'd probably want it to simply increment a global system
variable, and have the serial input driver read that for a timestamp. Dunno
about output.

#: 9194 S10/OS9/6809 (CoCo)
    19-Jan-91  10:29:36
Sb: #EPROMs
Fm: LUTE MULLENIX 70721,2230
To:  76703,4227 (X)
Kevin:
I have a question for you, I noticed that someone asked about putting the boot
in a EPROM. Do you think that it would be possible to burn every thing up to
the reading of the startup file in an EPROM? Some way of booting without a
floppy. I now have a SCSI HD using the Disto adapter and Ken Scales drivers.
And I have noticed that on bootup, it reads from /D0 up to the point of the
startup file, which is read from /DD (/H0).
By the way, does OS9 automaticly change to a hard drive as the primary drive
when one is on line?
     >Lute<
There are 2 Replies.

#: 9199 S10/OS9/6809 (CoCo)
    19-Jan-91  17:47:41
Sb: #9194-EPROMs
Fm: Pete Lyall 76703,4230
To: LUTE MULLENIX 70721,2230 (X)
Lute -
If you have a large enough EPROM (typically around 32K), you can burn the whole
mess into EPROM. Most coco's are setup with 8K ROM slots though.
The thing that causes /D0 to be read is the preboot (part of track 34), and
then the Boot module. When the Boot module completes, it hands off to Os9,
which in turn runs Sysgo, CC3Go, or whatever is specified in the Init module.
The Init module also controls who the Root directory will be at startup, so
this is the place you need to tweak for HD startup.
Once you have customized your INIT module, cobbled a bootfile, and burned it
into ROM, you can do what you like.
Pete

#: 9205 S10/OS9/6809 (CoCo)
    19-Jan-91  22:41:13
Sb: #9194-EPROMs
Fm: Kevin Darling (UG Pres) 76703,4227
To: LUTE MULLENIX 70721,2230 (X)
Well, let's review the relevant boot sequences:
 1. DOS command loads in kernel track (REL,BOOT,OS9P1), executes it.
 2. REL relocates (copies) kernel track code to $ED00, sets GIME etc.
    REL then jumps to OS9P1.
 3. OS9P1 inits vars, verifies kernel modules, tries F$Link to INIT.
    Link fails, so does F$Boot which calls BOOT.
 4. BOOT loads in os9boot file from floppy.  OS9P1 can now find INIT.
 5. OS9P1 tries link to OS9P2, jumps to it (else does F$Boot).
 6. OS9P2 inits vars, checks INIT for default device, CHD/CHX to it
    (else F$Boot)
 7. OS9P2 checks INIT for default term, opens it.
 8. OS9P2 starts up CC3GO.
 9. CC3GO tries to CHD/CHX to /H0.  Does startup shell, etc.
To boot without a floppy, the best method is to make an EPROM which loads in a
kernel "track" off the hard disk... and also make a BOOT module which reads in
an OS9boot file from hard disk.  This leaves everything updateable, and the
only thing you have to really change is the BOOT module.
Still, you could burn everything (bootfile and all) into EPROM which copied
itself into RAM... with a BOOT module that just returned "okay".  Or change
OS9P1 slightly so that it verifies _all_ the EPROM'd modules (then BOOT
wouldn't get called at all!).

#: 9196 S10/OS9/6809 (CoCo)
    19-Jan-91  14:14:08
Sb: #9190-12-bit timer: help!
Fm: Bill Dickhaus 70325,523
To: Kevin Darling (UG Pres) 76703,4227 (X)
Kev,
Probably the only workable way to deal with MIDI output would be to have the
driver handle timing. This means passing timing information with the outgoing
MIDI data to the driver. It would end up being fairly complex. I think the most
realistic place for all this timing and stuff is out on a hardware interface,
rather than bogging down the CoCo with all those interrupts.
Bill

#: 9198 S10/OS9/6809 (CoCo)
    19-Jan-91  17:39:09
Sb: #9190-#12-bit timer: help!
Fm: Pete Lyall 76703,4230
To: Kevin Darling (UG Pres) 76703,4227 (X)
Kev -
On the internet mailing address, those used for UUCP don't translate directly.
Example:
  Internet: pete@wlbr.imsd.contel.com
The imsd.contel.com is a domain name assigned to my company out here. It is
registered at SRI, where the internet numbers are controlled. Saying
'imsd.contel.com' is the same as using a Internet address (TCP/IP) of
 192.26.147.???. In our case, 'wlbr' is a local host with a node number of 8,
so the resolved Internet address for wlbr.imsd.contel.com is 192.26.147.8. The
'pete@' part just says that when it gets to that machine, further direct it to
user id 'pete'. Now - how this message gets routed from its source to
pete@wlbr.imsd.contel.com is a complex issue, and one I'm not slick enough to
explain.
The UUCP addresses however are VERY literal. In my case, the UUCP address would
be {major backbone sites}!jplgodo!wlbr!pete, where the message will actually
pass through each of the specified sites, until it reaches wlbr, and is
ultimately mailed to me.
Pete
There is 1 Reply.

#: 9204 S10/OS9/6809 (CoCo)
    19-Jan-91  22:40:11
Sb: #9198-12-bit timer: help!
Fm: Kevin Darling (UG Pres) 76703,4227
To: Pete Lyall 76703,4230 (X)
Ah! Thanks.  I knew it wouldn't/couldn't be that simple ;-).

#: 9200 S10/OS9/6809 (CoCo)
    19-Jan-91  18:08:43
Sb: FOR SALE
Fm: John Colegrove 76467,36
To: ANYONE
FOR SALE!COCO3 512K,2 drives,CM8mon,modem-pak,music-pak,del-joysticks,lots of
software+XtraCOCO3for parts-Leave Message-J.Colegrove76467,36

#: 9202 S10/OS9/6809 (CoCo)
    19-Jan-91  19:48:17
Sb: #9117-#Eliminator repair
Fm: Bruce Isted (UG VP) 76625,2273
To: JOERG SATTLER 74016,631 (X)
Joerg,
  I've heard that Emerald may not be able to help with WD 1002-05 repair at
the moment, but here's their address & phone number again anyway:
Emerald Microware
P.O. Box 1726
Beaverton, OR   97075
Phone:  (503) 641-0347
There is 1 Reply.

#: 9231 S10/OS9/6809 (CoCo)
    22-Jan-91  07:14:43
Sb: #9202-Eliminator repair
Fm: JOERG SATTLER 74016,631
To: Bruce Isted (UG VP) 76625,2273
Spoke to Brian just this past Friday. He' going to try and see what he can do.
He's got the controller, so we'll have to hope and see.     Thanks  Joerg

#: 9206 S10/OS9/6809 (CoCo)
    19-Jan-91  23:00:28
Sb: #Eliminator RTC
Fm: james pottage 71750,2012
To: [F] Bruce Isted 76625,2273 (X)
Bruce, I've been using the eliminator with the real time clock for about one
month now and I have noticed that the system seems to speed up about every 15
seconds (like a power burst) when I am running Pacos9. The momentary speed
increase only happens when I am using Pacos9 as far as I can tell. Any ideas on
why this is happening and how to solve it would be greatly appreciated
                                                           Jim Pottage exit
There are 3 Replies.

#: 9207 S10/OS9/6809 (CoCo)
    20-Jan-91  10:50:47
Sb: #9206-Eliminator RTC
Fm: James Jones 76257,562
To: james pottage 71750,2012 (X)
You know, that is sort of odd--there are times when I run pacos9 that things do
seem to speed up in short bursts, and the creatures seem to dash about the maze
faster.  (The glyph that I'm controlling doesn't seem to...sigh. :-) I wonder
what, if anything, pacos9 is doing?

#: 9211 S10/OS9/6809 (CoCo)
    20-Jan-91  13:58:09
Sb: #9206-Eliminator RTC
Fm: Bruce Isted (UG VP) 76625,2273
To: james pottage 71750,2012 (X)
Jim,
  I haven't seen a problem such as the momentary speed up that you describe,
but it sounds like a problem that could be caused by a marginal GIME chip.
Since the 60 Hz tick count of the GIME's VBORD IRQ is used by the RTC clock
module (same as the software clock module) that'd be the first thing to check.
  Do you happen to know what version (1986, 1987, or 1987A - I think those are
the versions the GIME has gone through) GIME chip you have?  If your CoCo 3 is
pretty old I'd suspect it had a questionable 1986 version GIME.
  Oh... before going out & replacing the GIME, sometimes all it takes is a
gentle press to seat it in its socket properly... if you don't mind opening the
case.
  Also, are you using a Multi-Pak Interface, and if so, is it upgraded for use
with the CoCo 3?  If its not upgraded, do that first before you do anything
else.
  Bruce

#: 9220 S10/OS9/6809 (CoCo)
    21-Jan-91  04:44:36
Sb: #9206-Eliminator RTC
Fm: LARRY OLSON 72227,3467
To: james pottage 71750,2012 (X)
James,
     Put your tools and testers away <grin>, there is nothing wrong with the
hardware. The speedup you see in pacos9 is all software related. Kevin was
right about part of what you are noticeing, with the new GRFDRV installed the
ghosts will sometimes line up on a byte boundry which makes the gets/puts
faster, this is part of the speed increase. The other part of the speed
increase you see is if pacman is not moving, when pacman is stopped, the
processing time that would be devoted to moving pacman is now available for the
ghosts, which causes them to speed up.
   If Bruce Isted's VRN was available at the time, it could have been used to
fix that problem by setting the program up to move the objects under control of
interrupts.
   I had some code in the program that would waste the processing time that
pacman wasn't using, so that the ghosts didn't speed up noticeably, but I took
it out. It seemed terrible to be throwing away processing time and anyway the
more I thought about it, I thought that the speedup of the ghosts when pacman
stops might add a little more of a challenge to the game.
  So James....Do Not Attempt To Adjust Your Set
           We Are Controlling What You See....
Larry (guilty party) Olson

#: 9209 S10/OS9/6809 (CoCo)
    20-Jan-91  13:14:06
Sb: #9178-Link shell command
Fm: Denise Tomlinson 71021,3274
To: Kevin Darling (UG Pres) 76703,4227 (X)
Thanks, I'll try that. I am using a real time clock for "setime". I am using
the smart watch in my controller. Works real good for time keeping under os9,
by the way. Huggs, Denise

#: 9210 S10/OS9/6809 (CoCo)
    20-Jan-91  13:15:02
Sb: #9179-Link shell command
Fm: Denise Tomlinson 71021,3274
To: Bill Dickhaus 70325,523 (X)
Thanks for the info Bill, I'll try that. Huggs, Denise

#: 9212 S10/OS9/6809 (CoCo)
    20-Jan-91  15:36:40
Sb: #Eliminator RTC
Fm: james pottage 71750,2012
To: Bruce Isted
Bruce, I am not using a multi pak interface. As well the problem does not seem
confined just to the one coco. I have a second newer coco3 that also has the
same problem as the older coco3. The problem I encounter is the same as James
Jones described in message 9206. The newer COCO3 was boughtin 1989 during the
coco clearance sales. The eliminator system I am running is composed of the
eliminator, on seagate st124 drive (20meg), one 40meg imprimis wren II drive,
one 3 1/2 720K drive and one 360K 5 1//4 drive. Also I am using a monochrome
monitor with the disto monocrome adapter on this system, along with one
joystick. Could the problem be with either the RTC or with pacos9?
                                                         Jim Pottage
There are 2 Replies.

#: 9213 S10/OS9/6809 (CoCo)
    20-Jan-91  15:48:22
Sb: #9212-Eliminator RTC
Fm: Bruce Isted (UG VP) 76625,2273
To: james pottage 71750,2012 (X)
Jim,
  I really don't know what the problem is... I have pacos9 (CRC $FE5BF0) but
I've never noticed any problem.  Of course, I don't play too many games, except
when I need a break and probably wouldn't notice a problem if it bit me.
  I doubt the problem is with the RTC, but you could check by making up a new
boot file that has the software clock module in it.
  Bruce

#: 9215 S10/OS9/6809 (CoCo)
    20-Jan-91  19:43:40
Sb: #9212-Eliminator RTC
Fm: Kevin Darling (UG Pres) 76703,4227
To: james pottage 71750,2012 (X)
Jim - I've forgotten whether Pacos9 used the windowing system or not... if it
did, then perhaps it hits on the fast Grfdrv PUT boundaries now and again, and
that's the reason for the speedups you see.
He wrote it before that grfdrv was widely available, if I recall.
 best - kev

#: 9214 S10/OS9/6809 (CoCo)
    20-Jan-91  16:22:44
Sb: #Eliminator RTC
Fm: james pottage 71750,2012
To: Bruce Isted
Bruce, after I sent my last message I discovered two interesting things. First,
the old grfdrv driver seems to work fine with the RTC. Only the new grfdrv
seems to cause the problem. Also, the system works fine with the new grfdrv if
the RTC is not used and the clock driver sent with the eliminator package is
used (or the old clock driver that came with OS9 Level II - I am presently
using your clock with my disto system and it works fine). Is it possible that
the new grfdrv is turning off the interupts?       Jim Pottage
There is 1 Reply.

#: 9216 S10/OS9/6809 (CoCo)
    20-Jan-91  19:44:26
Sb: #9214-Eliminator RTC
Fm: Kevin Darling (UG Pres) 76703,4227
To: james pottage 71750,2012 (X)
Aha. Yep, see previous message. It must be the fast PUT locations. - kev

#: 9218 S10/OS9/6809 (CoCo)
    20-Jan-91  21:17:58
Sb: Eliminator RTC
Fm: james pottage 71750,2012
To: Kevin Darling
Kevin - Thanx very much for the reply.

#: 9219 S10/OS9/6809 (CoCo)
    20-Jan-91  21:51:19
Sb: #HELP
Fm: John Colegrove 76467,36
To: ANYONE
I am fairly new at OS9 and have a problem. I try to use Deskmate 3's Telecom
with a Direct connect modem pak(junk I know)and it won't work. When I run ident
on the different modules, I get MODULE HEADER IS INCORRECT! on DMAPPS and
DMALT. I asked people in Fort Worth and they don't know what I am talking
about. Can anyone help?
There is 1 Reply.

#: 9223 S10/OS9/6809 (CoCo)
    21-Jan-91  16:18:46
Sb: #9219-HELP
Fm: Steve Wegert 76703,4255
To: John Colegrove 76467,36 (X)
John,
I wouldn't get too concerned over the ident error report, just yet. It's
possble that Deskmate just can't find the right serial device. You did tell it
to use /m1 and you have /m1 and modpak in your boot, jes?
Steve

#: 9225 S10/OS9/6809 (CoCo)
    21-Jan-91  18:35:47
Sb: #Sending nulls to T2
Fm: Paul Rinear 73757,1413
To: Anyone
    I use an old printer that has trouble keeping up at 1200 baud.
 If I use xmode to set null=120, there is enough delay for it to
 work. However, consecutive line feeds are unbearably slow.
    I tried setting null back to zero and writing a filter that
 counts output characters until a 'return' and then sends a number
 of nulls based on this line length. The shorter the line the shorter
 the delay after a return.
    Using this command :
                         list filename ! filter > /t2
           I get no delay no matter how many null characters the
 filter sends out.
           Should this work, or is there something in the device
 descriptor that ignores these null characters ?
                                      ??????,
                                              Paul R.
There is 1 Reply.

#: 9229 S10/OS9/6809 (CoCo)
    21-Jan-91  22:52:42
Sb: #9225-#Sending nulls to T2
Fm: Kevin Darling (UG Pres) 76703,4227
To: Paul Rinear 73757,1413 (X)
Paul - your filter would have to do I$Read and I$Write (vs WriteLn which would
ignore nulls, you see).  What language are you using?  (if you don't
understand, that is ;-).  best - kev
There is 1 Reply.

#: 9232 S10/OS9/6809 (CoCo)
    22-Jan-91  12:40:48
Sb: #9229-#Sending nulls to T2
Fm: Paul Rinear 73757,1413
To: Kevin Darling (UG Pres) 76703,4227 (X)
     I wrote it in C using getchar and putchar. No good?
  While on the subject of languages, I have another question: I got
  the Level I assembler (asm) so that I can assemble code generated
  by the Pascal translator program. It seems to work with no errors,
  but it leaves no object code anywhere. ??
     I ordered a copy of your book, Inside OS9 Level II, and there
  is a three week wait. Maybe it will help answer the last question
  I left (about renegade windows?).
                                       Thanks,
                                                Paul
There are 2 Replies.

#: 9233 S10/OS9/6809 (CoCo)
    22-Jan-91  13:34:07
Sb: #9232-Sending nulls to T2
Fm: Kevin Darling (UG Pres) 76703,4227
To: Paul Rinear 73757,1413 (X)
I can't recall if Asm automatically leaves a module in your current CHX dir or
not.  If not, try something like "asm file o=filename"... been a while since I
did that.
On the filter, try redirecting its output to a file, and dumping it to see
what's actually coming out of your filter.  Might give some clues.

#: 9236 S10/OS9/6809 (CoCo)
    22-Jan-91  18:30:00
Sb: #9232-#Sending nulls to T2
Fm: Pete Lyall 76703,4230
To: Paul Rinear 73757,1413 (X)
Paul -
Getchar and putchar use buffered I/O, which will intelligently select
I$writ/readln or I$write/read.... You'd be better off using read()/write()...
There's some skulduggery you can use to force binary operations, but this is
cleaner...
Re: asm no generating code... hmmm. Did you try:
 asm l o=/d0/cmds/woof woof #20k
The 'o' argument tells it to generate the binary... actually, you could have
omitted the filename and it would have defaulted to the execution dir and the
same name as the source.
Pete
There is 1 Reply.

#: 9237 S10/OS9/6809 (CoCo)
    22-Jan-91  19:17:16
Sb: #9236-#Sending nulls to T2
Fm: Paul Rinear 73757,1413
To: Pete Lyall 76703,4230 (X)
  The 'o' argument works, and I get an output file. If I set the attributes and
then load the file, it shows up on an 'mdir' but will always load from disk
again when I try to run it. It also will load into memory without the execute
attribute set. This is utterly useless.
  I will try your suggestions on the C code.
                                             Thanks,
                                                        Paul
There is 1 Reply.

#: 9246 S10/OS9/6809 (CoCo)
    22-Jan-91  23:43:44
Sb: #9237-#Sending nulls to T2
Fm: Pete Lyall 76703,4230
To: Paul Rinear 73757,1413 (X)
If it loads from disk if it is already in memory, then very likely your MODULE
name is not the same as the FILENAME. Make sure that they are the same.
Also - I _dare_ you  to load an executable without execute permissions... are
you SURE it's actually in memory?
Pete
There is 1 Reply.

#: 9255 S10/OS9/6809 (CoCo)
    23-Jan-91  09:45:19
Sb: #9246-#Sending nulls to T2
Fm: Paul Rinear 73757,1413
To: Pete Lyall 76703,4230 (X)
    My mistake on the attributes; I had the source code file named the same as
the executable and I attr'd that one.
    On the other point, it doesn't seem to matter if the module name is the
same as the disk file name or not, it still goes to the disk. The module is
definitely in memory. Makes me think that the module itself ia is accessing the
disk. It is only a test program of the 'hello world' sort.