[PD] Crasher bug found...

[Luke Iannini]

On Mon, Feb 2, 2009 at 7:44 PM, Miller Puckette
<mpuckett at imusic1.ucsd.edu> wrote:
> Thanks -- that was just what I needed to find it... here's a patch that
> should fix it.  I'll be puttint up a new bugfix release after I manage to
> clean up another couple of problems.
Hi Miller, a little late, but just wanted to say a big thanks for
finding this and fixing it!  I'm very happy to be able to continue
work on my sequencers crash-free : )
Best
Luke

>
> Miller
>
> diff --git a/src/g_template.c b/src/g_template.c
> index 34541c0..f216b10 100644
> --- a/src/g_template.c
> +++ b/src/g_template.c
> @@ -106,7 +106,7 @@ t_template *template_new(t_symbol *templatesym, int argc, t_
>     bad:
>         argc -= 2; argv += 2;
>     }
> -    if (templatesym->s_name)
> +    if (*templatesym->s_name)
>     {
>         x->t_sym = templatesym;
>         pd_bind(&x->t_pdobj, x->t_sym);
>
>
> On Sun, Feb 01, 2009 at 12:43:48AM -0800, Luke Iannini wrote:
>> Well, I'll let you all confirm that before I truly say so : ).
>>
>> But, by sheer luck, I'm hoping I've found the memory-corruption bug
>> causing my pd abstractions to be so unstable (as I've lamented quite a
>> few times).
>>
>> Hopefully this is narrowed down enough to be useful Miller! (I'm happy
>> to pare it down further if there are too many extraneous objects)
>>
>> Basically, open ds-slider-o-help.pd (which contains 10 copies of the
>> GOP ds-slider-o.pd abstraction) and click the [CLICK TO CRASH( message
>> which is connected to a [send] object with no destination.
>>
>> Assigning a destination prevents the crash.
>>
> [...]
>