[PD] Fwd: What exactly is a "stack overflow" ?

[Patrice Colet]

  Hello Mike, everyone,

Mike McGonagle a écrit :
> 
> On Dec 19, 2007 2:57 PM, Charles Henry < czhenry at gmail.com 
> <mailto:czhenry at gmail.com>> wrote:
> 
> 
>     I seem to be tuning in a little late, in this discussion, but if it's
>     a bad problem, couldn't you change the method of until to use only
>     floats?
> 
> 
> But this is the whole point of discussions like this. It is NOT to fix a 
> FEATURE of PD, but to give warnings as what NOT to do to shoot yourself 
> in the foot.
> 
> If you know NOT to connect up a [bang( -> [until] without some sort of 
> termination means, then you won't do it... Or if you really need to 
> experience it first hand, then by all means do it. You probably will 
> never do it again...

  Well, even with experiencing the [until] machine gun fire crashing de 
computer once, with simply testing what's happening, the user could get 
this problem by accident, because of (let's say) another bug in the 
patch or in the externals, unless it's used through an abstraction with 
protections against every kinds of overflow issues.

> Besides, this object is so old that to make a change like this would 
> probably break a large portion of PD patches that use [until].
> 
> Mike
> 

  Do you think it would be possible to define internaly a limit to 
[until] for the case if it has been banged? The sub-machine gun would 
stop after this limit has been reached. This limit number would be 
determined by the processor in use, right?
  During PureData building the compiler could fix this limit following 
which cpu it is used for.