Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.lang.forth > #34720 > unrolled thread

A useful trick in recursive debugging.

Back to article view | Back to comp.lang.forth

Contents

  A useful trick in recursive debugging. albert@spenarnc.xs4all.nl (Albert van der Horst) - 2014-11-30 17:10 +0000
    Re: A useful trick in recursive debugging. Paul Rubin <no.email@nospam.invalid> - 2014-11-30 09:54 -0800
      Re: A useful trick in recursive debugging. albert@spenarnc.xs4all.nl (Albert van der Horst) - 2014-11-30 20:04 +0000
        Re: A useful trick in recursive debugging. Paul Rubin <no.email@nospam.invalid> - 2014-11-30 12:36 -0800

#34720 — A useful trick in recursive debugging.

Normally .S looks something like

: .S
    CR &S EMIT &[ EMIT SPACE
    DSP@ S0 @ BEGIN 2DUP - 0< WHILE 0 CELL+ - DUP @ . REPEAT 2DROP
    &] EMIT ;

With output like
S[ 0 0 0 0 0 0 0 896 -420 ]

Stack printouts for recursive programs can be hard to read.
A little indenting works wonders. What you need is some way
to find out the return stack depth. In the example the actual
stack pointe RSP@ is compared to the initial one.
You may even have RDEPTH in your favourite Forth.

: .S
    CR   R0 @ RSP@ - 1 CELLS / SPACES     &S EMIT &[ EMIT SPACE
    DSP@ S0 @ BEGIN 2DUP - 0< WHILE 0 CELL+ - DUP @ . REPEAT 2DROP
    &] EMIT ;


Now the output for a recursive program is
...
          S[ 3 1 ]
             S[ 4 3 0 ]
                   S[ 4 0 2 0 ]
                      S[ 4 1 2 9 ]
                   S[ 4 1 1 -20 ]
                      S[ 4 1 1 0 ]>
             S[ 4 2 ]
             S[ 6 3 9 ]
                   S[ 6 0 2 9 ]
                      S[ 6 0 2 18 ]
                   S[ 6 0 1 -11 ]
                      S[ 6 0 1 9 ]>
             S[ 6 0 ]>

For
    : doit .s aap noot .... mies .s ;
the input and the output of a Forth word now line up
nicely.

Groetjes Albert
-- 
Albert van der Horst, UTRECHT,THE NETHERLANDS
Economic growth -- being exponential -- ultimately falters.
albert@spe&ar&c.xs4all.nl &=n http://home.hccnet.nl/a.w.m.van.der.horst

[toc] | [next] | [standalone]

#34722

albert@spenarnc.xs4all.nl (Albert van der Horst) writes:
>           S[ 3 1 ]
>              S[ 4 3 0 ]

How do you locate the boundaries between DS frames without
pushing/popping DSP on the return stack with every function call?

[toc] | [prev] | [next] | [standalone]

#34724

In article <7xr3wk8thd.fsf@ruckus.brouhaha.com>,
Paul Rubin  <no.email@nospam.invalid> wrote:
>albert@spenarnc.xs4all.nl (Albert van der Horst) writes:
>>           S[ 3 1 ]
>>              S[ 4 3 0 ]
>
>How do you locate the boundaries between DS frames without
>pushing/popping DSP on the return stack with every function call?

With a .S at the start and the entry of a recursive call, all
intermediate stack prints are more indented. So those two line
up vertically. That is often enough. After all .S is not much
more than a simple debugging tool.

In other words, DS stack frames are just the two or three last
stack items from a less indented line
    S[ 0 0 0 0 0 1 2 ]
         ..
         s[ ..
         S[ ..
         ..
    S[ 0 0 0 0 0 3 ]

With a stack diagram of ( n n -- n )
we see that some call did ( 1 2 -- 3 )

Groetjes Albert
-- 
Albert van der Horst, UTRECHT,THE NETHERLANDS
Economic growth -- being exponential -- ultimately falters.
albert@spe&ar&c.xs4all.nl &=n http://home.hccnet.nl/a.w.m.van.der.horst

[toc] | [prev] | [next] | [standalone]

#34725

albert@spenarnc.xs4all.nl (Albert van der Horst) writes:

> In article <7xr3wk8thd.fsf@ruckus.brouhaha.com>,
> Paul Rubin  <no.email@nospam.invalid> wrote:
>>albert@spenarnc.xs4all.nl (Albert van der Horst) writes:
>>>           S[ 3 1 ]
>>>              S[ 4 3 0 ]
>>
>>How do you locate the boundaries between DS frames without
>>pushing/popping DSP on the return stack with every function call?
>
> With a .S at the start and the entry of a recursive call, all
> intermediate stack prints are more indented. So those two line
> up vertically. That is often enough. After all .S is not much
> more than a simple debugging tool.

Oh I see, that output was from a separate call to .S at each level.

Does the .S somehow distinguish between recursive calls and ordinary
calls?  I.e. do you get indentation if you're 4 levels deep in normal,
non-recursive calls?  I suppose it's fine either way.

[toc] | [prev] | [standalone]

Back to top | Article view | comp.lang.forth

csiph-web