Subj : Re: more chess
To   : Rolf Jentsch
From : Jasen Betts
Date : Mon Nov 26 2001 09:41 pm

Hi Rolf.

25-Nov-01 15:47:00, Rolf Jentsch wrote to Jasen.Betts@xspamp42.f531.n640.z3.f


 RJ> From: email@rolfjentsch.de (Rolf Jentsch)

 RJ> Jasen.Betts@xspamp42.f531.n640.z3.fidonet.org meinte am 24.11.01 zum
 RJ> Thema
 RJ> "more chess"

 RJ> Hi Jasen

 RJ> some comments on your batch?

ok, it's basically finished (just needs a little polish, and some final
tweaks. and changes to get round a couple of bad bugs I found in option...

 RJ> Here we go

>> setdos /b0
>> setdos /b1
 RJ> How about one line?

because it doesn't work here (possibly a bug in STM when it changes the
display mode it also sets the blinking on but doesn't inform 4dos, so if
blinking was already off, it takes too rounds of setdos /b to turn blinking
off, this difficulty is not evident on the command-line bacause 4dos activates 
the
blink setting before every prompt.

>> if %_env ge 799 goto enough
>> :moreenv
>> c:\4dos\4dos /E:800 /C %0
>> goto end_of_batch

 RJ> revert question, use comspec, and get rid of goto f.e
 RJ> if %_env lt 800 (%comspec /e:800 /c %0 %+ quit)
 RJ> seems to achieve the same goal.

here I end up with %_env = 799 after /e:800 so the above would eat all ram
then crash. other than that,

if comspec isn't pointing to 4dos.com it also won't work
near the start of the batch I detect non-4dos using
if not 4dos == 4DOS goto moreenv

>> alias bad_move=goto get_move

 RJ> no real need for an alias. A set would have done the same and is probably
 RJ> faster....

It shouldn't make much difference as the batch is 100% internal commands.
the aliases are kept in lower memory (so there's no swapping from ems/xms/disk)
a new version actually makes used of some of the alias features to print
"error" messages.

also %vars are done when 4dos sees the line, but think aliases aren't
processed until 4dos get to that part of the line. so there's no accessing
of the alias contents until it's needed

also aliases should be faster as 4dos doesn't need to use the O/S to
access them.

>> :redraw
>> for x in (0 1 2 3 4 5 6 7) for y in (0 1 2 3 4 5 6 7) set v%x%y=.
>> drawbox 0 16 41 73 2 7 on 8
 RJ> Make it a subroutine

I don't like single-use subroutines.

>> :drawboard
>> for x in (0 3 6 1 4 7 2 5) do (
>>   scrput 0 %@eval[%x*7+20] 1 on 14 %@char[%@eval[%x+65]]
>>   scrput 41 %@eval[%x*7+20] 1 on 14 %@char[%@eval[%x+65]]
>>   scrput %@eval[%x*5+3] 16 1 on 14 %@eval[8-%x]
>>   scrput %@eval[%x*5+3] 73 1 on 14 %@eval[8-%x]
>> )
 RJ> this one too.

I'm rearranging that stuff a bit now.

>> if "%mov" == "" bad_move
>> switch %mov
>> case debug
>>              set debug=%@if[ "%debug" == "1" ,,1]

>> case quit .or. resign
>>              goto resign
 RJ> instead of goto use the (set fh.... %+ quit)

that has merit... but why is it better than a goto?
I was putting the exit-point at the end of the batch...

>>              if %move ge %fs bad_move
>>              set move=%@inc[%move]
>>              gosub read_move %move
>>              goto drawboard
>> case redraw .or. refresh
>>              goto redraw
>> endswitch

 RJ> drawboard is done in any way, so pull it behind the swich as a
 RJ> subroutine.

It's not done as the default action, i'd have to put a goto to skip over
it. drawboard is slow on this machine (100Mhz 486) so I only do it when
neccessary,

the lastest (unreleased) version of the batch has a one-line drawboard
to update minor changes in the display resulting in some speed improvements
the spiral drawing is still used at startup as it's not significanlty
slower than drawing the pieces in a more mundane order and it's more fun
to wait for! (takes about 8 seconds)

>>   if %y lt 58 (if %[y%b] ne - bad_move %+set y%b=%@eval[56-%y] %+iterate)

 RJ> does this work? Yes, but it is not easy to understand.
 RJ> That iff %y lt 58
 RJ> then
 RJ>  if %[y%b] ne - bad_move
 RJ>  set y%b=%@eval[56-%y]
 RJ>  iterate
 RJ> endiff
 RJ> seems more readable.

but it's about 6 times longer, probably about as fast though.
that whole section needs rearranging for readibility not sure what the best
approach is though, if I split it out into long iffs to gets really long and
doesn't fit on the screen.

 RJ> Here some figure suggestions

 They got blanked out somewhere en-route. (you might have to uuencode them,
 or worse, use %@char)

 base-64 also works, but quoted printable I have to undo by hand :(

 -=> Bye <=-

---
 * Origin: Keep grandma off the streets -- legalize bingo. (3:640/531.42)

.