CW%APG.PH.UCL.AC.UK@pucc.PRINCETON.EDU (04/01/91)
Hi again,
Well I've been playing around with the routine at 371Dh all
weekend and it seems rather better than I imagined. So far
I have tried it on
i) Array of String (Varying length)
ii) Array of Global Name (varying length)
iii) Array of XLIB name
iv) Array of System Binary
v) Array of Real Array (Varying sizes and dimensions)
PLUS vi) Arrays of mixed types!! (More later)
and each time it is capable of correctly extracting the n'th object
from an array. An example now follows of vi) above with an array
that contains
[ Array_of_Real Array_of_System_Binary Array_of_XLIB_name ]
This will demonstate the abilities of the routine at 371Dh quite
nicely.
This array looks like this in memory ( as it would be viewed in the
memory scanner (i.e. reversed))
8E920 [ Array prolog
68000 Length of array in nybbles (inc these 5)
8E920 Here are 5 nybbles that dictate
what is contained in the array
In this case 029E8-> array of arrays
10000 Number of dimensions
30000 Number of points in that dimension
(i.e. one dimensional array of three arrays)
Then begins the first object without its header
93000 Length of 1st array in nybbles (inc these 5)
33920 Prolog of what is contained in this array
(real numbers)
20000 Number of dimensions
20000 Number of numbers in first dimension
10000 Number of numbers in second dimension
0000000000000010 First real number (1)
0000000000000020 second real number (2)
Then begins the second object without its header
91000 Length of this array in nybbles (inc these 5)
11920 Prolog of whats in this array (system binary)
10000 Number of dimensions
10000 Number of system binaries in that dimension
(1x1) array
55555 The system Binary itself <55555h>
Then begins the third object without its header
02000
29E20 Prolog of whats in this array (XLIB names)
10000
20000
B01000 XLIB 267 0
B01100 XLIB 267 1
If the 48sx was programmed to display this it would look something
like,
[ [[ 1 ] [ <55555h> ] [ XLIB 267 0 XLIB 267 1 ] ]
[ 2 ]]
but if you do (*this point is marked for further reference)
"8E920680008E9201000030000930003392020000200001000000000000000000
10000000000000002091000119201000010000555550200029E201000020000B
01000B0110026CF" ASC->
(which is equivalent to putting the aforementioned array on the stack)
you get
1: Array of Real Array
but don't be put off! All the information is in there.
Incase you didn't read my last posting I found that if I had the following
on the stack
2: <N>
1: Array of String
(Where <N> is a system binary i.e. <1h> <3d> etc....)
And typed the following
#371Dh SYSEVAL
then the N'th string in the array would be extracted.
And you get
2: "N'th string"
1: External
where I am not sure what the external is (probably True or False for
success or failure?). I have now found that it can be generalized further
to if you have the following on the stack
1: <N>
2: Array of Any Object
and you type
#371Dh SYSEVAL
then the N'th Object of the array will be extracted. (When I say
array of any object I am assuming it will work for any object
although I've only tried it on the ones listed above, but I see no
reason why it should not work on any object!!)
So lets write a program called GET2 (typed in by hand).
\<<
R\->B
#5A03h SYSEVAL @Binary to System Binary
SWAP
#371Dh SYSEVAL @Extract N'th object from array
DROP @Get rid of external
\>>
So if you create the array mentioned above (see point marked for
further reference above) then you can play with. I shall refer to it
as ARRAY
If you have ARRAY (it must be the array and NOT a local or global name
specifying where it is stored) in level one of the stack the following
can be done
A) 1 GET2 produces
1: [[ 1 ]
[ 2 ]]
and then Ai) 1 GET2 produces
1: 1
or Ai) 2 GET2 produces
1: 2
B) 2 GET2 produces
1: Array of System Binary
and then Bi) 1 GET2 produces
1: <55555h>
C) 3 GET2 produces
1: Array of XLIB name
and then Ci) 1 GET2 produces
1: XLIB 267 0
or Cii) 2 GET2 produces
1: XLIB 267 1
Which is what was stored in the array.
Well I'll leave it there for the moments. If anybody finds an object it
doesn't work on please let me know.
As for a use for all of this: Hard to say! Mixed arrays are rather like
lists, and so might not be necessary, and arrays of just one type of
object might not be useful, but as Arthur C Clarke says (I think it was
him), 'We are only limited by our imagination'.
One final caveat
This SEEMS to work for 2-dimensional arrays but you still only
apply one number. eg if you have
1: [ [ 3 7 ]
[ 4 5 ] ]
then 1 GET2 gives 3, 2 GET2 gives 7, 3 GET2 gives 4, and 4 GET2 gives 5.
Also the routine does check to see if an array has been passed to it
and justs returns an external (probably a fail flag).
Talk to you soon
Conradcloos@acsu.buffalo.edu (James H. Cloos) (04/02/91)
I did find the routine at 371d listed in my symbol tables (from Jake). Looks like it was contributed to the collective list be Rick Grevelle. He called it get_1array2short. Based on how the other routines work, it would be safe to say that all this routine does is take the array, jump ahead to the arry_type field, put that into a register, jump ahead to the first element, skip (short - 1) elements, and make an object out of the stored prolog and the data it finds at the (short)'th element. (It would most likely figure out how far to skip for each previous element based on the prolog: if the prolog is of a fixed size object, that many nybbles would be skiped each time, otherwise the size field for each element in turn would be read and that many nybbles skipped each time; the same criterion would be used to determine how many nybbles make up the wanted element. Please note that I HAVE NOT (yet) looked at the dissassembly of this routine & for that matter don't even know if it is in ML or RPL.) What this means is that any data could be used for this routine, allowing the programmer to, ah, have some fun, as it were. ;^) (BTW, Conrad, even though it has been previously documented, and a handful of us would think it obvious (I refuse to use trivial w/o the original src!), the net at large is grateful for the tutorial; the more of these we can get "out there," the more fun and interesting programs we will find available to us! Danke.) Happy hacking! -JimC -- James H. Cloos, Jr. Phone: +1 716 673-1250 cloos@ACSU.Buffalo.EDU Snail: PersonalZipCode: 14048-0772, USA cloos@ub.UUCP Quote: <>