[chbot] Chchrobotics Digest, Vol 198, Issue 10 GCC and recursion

[Mark Atherton]

Thanks Charles, and everyone.
 
I was only checking that the compiler (x86) generated re-entrant code before I climbed in too far.
 
The issue turned out to be a bug with pointer management.
 
-Mark
 

> On 03/04/2024 09:16 NZDT Charles Manning <cdhmanning at gmail.com> wrote:
>  
>  
> Recursion can always be mapped to looping and using a manual stack. Some languages, such as Fortran, lack recursion and this is how recursive constructs are achieved.
>  
> But back to the original question: gcc certainly handles recursion properly on ARM, x86 and other "real" CPUs.
>  
> But is this gcc generating code for a tiny hobby CPU? If so, it might not be using a proper stack which would then prevent recursion.
>  
> For example on some 8-bitters like 8051, C compilers and stacks do not map well so, instead, the code that is generated ends up using fixed memory locations for a fixed "stack". Such constructs do not support recursion.
>  
> I would suggest trying out making a small recursive structure such as factorial and have a look at the generated code.
>  
>  
>  
>  
>  
>  
> 
> On Wed, Apr 3, 2024 at 1:30 AM Bevin Brett <bevin_brett at hotmail.com mailto:bevin_brett at hotmail.com> wrote:
> 
> > The alternative to recursion is either a linked list, a fixed size array, or a growing array
> >  
> > In RPN evaluation, the code often looks like this, because you don't need the full power and cost of recursion
> >  
> > ops : array of operators
> > stack : stack of operands
> > top = 0
> > for op in instructionStream do
> >     caseOf nOperands(op.code)
> >         alt add:
> >             stack[top - 1] += stack[top]
> >             top = top - 1
> >         alt int:
> >             top = top+1
> >             stack[top] =http://op.int
> >         ...
> >  
> > Don't forget to check for malformed expressions, which could result in h/w stack corruption due to indexing outside 'stack'
> >  
> > /Bevin
> >  
> >  
> > 
> > ---------------------------------------------
> >  
> > ...
> > 
> > I am building an RPN expression evaluator for the uCPU assembler; I
> > can't see any other way to go other than to use recursion to do the
> > nested processing that I wish.
> > 
> > Currently using "gcc version 12.2.0 (MinGW-W64 x86_64-ucrt-posix-seh,
> > built by Brecht Sanders)"
> > 
> > The first call of the recursed function worked fine, but crashes upon exit.
> > 
> > It could be that the compiler is not generating re-entrant code. The
> > only switch I have used is -O0 to disable compiler optimization.
> > 
> > Are there any other switches for GCC that I need to use/be aware of ?
> > 
> > Thanks, Mark
> > 
> > ...
> > _______________________________________________
> > Chchrobotics mailing list Chchrobotics at lists.ourshack.com mailto:Chchrobotics at lists.ourshack.com
> > https://lists.ourshack.com/mailman/listinfo/chchrobotics
> > Mail Archives: http://lists.ourshack.com/pipermail/chchrobotics/
> > Meetings usually 3rd Monday each month. See http://kiwibots.org for venue, directions and dates.
> > When replying, please edit your Subject line to reflect new subjects.
> > 
> _______________________________________________
> Chchrobotics mailing list Chchrobotics at lists.ourshack.com
> https://lists.ourshack.com/mailman/listinfo/chchrobotics
> Mail Archives: http://lists.ourshack.com/pipermail/chchrobotics/
> Meetings usually 3rd Monday each month. See http://kiwibots.org for venue, directions and dates.
> When replying, please edit your Subject line to reflect new subjects.
> 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ourshack.com/pipermail/chchrobotics/attachments/20240403/9fb26a2a/attachment.html>