13 replies to this topic

[Mike Nicholson]

Found an interesting bit of code and am just trying to work out what actually happens. Thought I'd share it with people...

   CASE $VAR

     WHEN 'CASE1' : ... CODE ...

     WHEN 'CASE2' : ... CODE ...

                    STEP 1



1                   IF #VAR THEN 

                       STEP 2

                    ENDIF

                    ... CODE ...

                    STEP 2



2                   IF #VAR2 THEN STEP 1



      OTHERWISE    : ... CODE

   ENDCASE

Note that the code has been changed to protect the innocent...

[Dan Shannon]

#VAR will evaluate to TRUE if it has a non-zero value.

So it looks like a conditional loop repeating until #VAR2 is zero, skipping the code if #VAR is not zero?

Sounds completely insane to me, good job it was written by the innocent, otherwise I'd blame them.

Cheers

Dan Shannon

[Mike Nicholson]

The pseudo code I put in was a vast simplification. The IF #VAR statements are really a couple of lines long and there are a lot more of them. It was the jumping around with the steps that got me.

Probably shouldn't reveal the author though ;-)

Cheers

Mike

[Chris Pepper]

Thanks for this example.
I'm amazed that the parser processed it!
Perhaps we should have a competition for the strangest pieces of code?

[Neil Hunter]

Who uses step logic these days anyway,

why not rather use a for loop or while loop

Much easier to follow

[Surajit]

Hi all,

I have rarely seen PROIV's code which use 'STEP' keyword, it looks like 'GOTO' in BASIC which make the code lost its structure.

If we have to share the same line of codes in the same logic then use global logic.

regards,

[Richard Bassett]

An obfuscated ProIV competition - excellent idea!

Here is my entry.
In case it's not clear (!) it is a one-line global-logic that takes an alpha parameter and returns an alpha.

 001   A   PARMS($]#' =LEN($]$ =$[1,4 ^4)#' =LEN($]$ =$[1,#')RETURN[$[1,#']) 

PS. I am amazed you were amazed at anything the parser would process after all these years!

[Mike Nicholson]

Well it took me , Cleve and Rob for that one but now it's

PARMS($]

#' =LEN($]

$ =$[1,4 ^4)

#' =LEN($]

$ =$[1,#')

RETURN[$[1,#'])

It just space pads a string to 256 characters.

Simple really ;-)

P.S. Is this a standard example of your code, in which case please don't submit your CV here ;-)

[Chris Pepper]

Speaking as someone who has to look at Richard's code now and then he probably simplified the example a bit!!!

...and I had to write a test function to prove that you could read alternate records from two sequential files using one time LU's and LSCALL. It worked of course. So that was two things I learned in one day!

[Mike Nicholson]

Try this one...

## =1 #$ =0-1 FOR #1 = 0 TO ## 1-1(#1,0) UMSG(CONV(0),#$) ENDFOR

1-1
P PARMS(##,#$) #= = ## IF #= # 1 #= = 1; #$ =(## *-1) +#=

[Richard Bassett]

No, it doesn't just pad a string to 256 characters.
I wouldn't post something so boring!
It's much more useful than that.
Did you test it properly?

[Richard Bassett]

Of course it worked (cough)..
Glad someone tried it because I didn't

[Richard Bassett]

Well, that is more readable as:

#A = 1

#B = -1

FOR #i = 0 TO #A

  GLOGIC(#i, 0)

  UMSG(CONV(0), #

ENDFOR



procedure GLOGIC(#one, #two)

  #X = #one

  IF #X <> 1

    #X = 1 ;

  #two = (#one * -1) + #X

Which in turn simplifies to:

FOR #i = 0 TO 1

  GLOGIC(#i, 0)

  UMSG(CONV(0), -1)

ENDFOR



procedure GLOGIC(#one. #two)

  #two  = 1 - #one

So the paused output is 1 followed by blank (CONV of zero) because the call to GLOGIC reassigns the value of the pseudo-constant 0 (nice of you to reset it though).

It occurs to me that maybe one could save workspace by identifying sequences of code within which certain constants were not needed and using those constants therein as variables.

[Mike Nicholson]

>So the paused output is 1 followed by blank (CONV of zero) >because the call to GLOGIC reassigns the value of the >pseudo-constant 0 (nice of you to reset it though).

You're welcome ;-) although if I hadn't it could possibly have caused people a few problems

> It occurs to me that maybe one could save workspace by identifying sequences of code within which certain constants were not needed and using those constants therein as variables.

Now that's a scary thought. We do need to trim the function size of a few of our larger functions but not quite sure I'd fancy maintaining that ;-)

Rechecked the one you posted, it does pad out, but only to the length that you defined your variable in the first place. Perhaps Friday afternoons not the best time to try and read something like that - the pub was calling . . .

Anybody else - come on Jan, I know you must have a few !!

Cheers

Mike