[PD] Mysterious bug with [text]

[Miller Puckette]

Cool - pushed it to git repo.

thanks
M

On Sun, Mar 06, 2016 at 08:52:59PM +0100, IOhannes m zmölnig wrote:
> On 03/06/2016 08:21 PM, IOhannes m zmölnig wrote:
> > On 03/04/2016 11:22 AM, Christof Ressi wrote:
> >> > So you're probably right with your guess about memory corruption
> > actually, valgrind already proved that this was not a guess.
> 
> and with the help of valgrind it was also very easy to spot the actual
> problem: a pointer was not updated after calling binbuf_resize().
> 
> attached is a one-liner fix for the problem.
> 
> gfmards
> IOhannes

> From ff63300dc38fb5d659c4bf9831ea5ddd05680a2c Mon Sep 17 00:00:00 2001
> From: IOhannes m zmoelnig <zmoelnig at umlautQ.umlaeute.mur.at>
> Date: Sun, 6 Mar 2016 20:49:34 +0100
> Subject: [PATCH] Make sure that 'vec' is up-to-date.
> 
> of we call `binbuf_resize()` the address of the atom-vector might change
> (even if we shrink), so we need to update the 'vec' variable that's caching
> this address.
> ---
>  src/x_text.c | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/src/x_text.c b/src/x_text.c
> index 4b25983..eed93a5 100644
> --- a/src/x_text.c
> +++ b/src/x_text.c
> @@ -706,6 +706,7 @@ static void text_set_list(t_text_set *x,
>                      sizeof(*vec) * (oldn - end));
>                  if (n < oldn)
>                      (void)binbuf_resize(b, n);
> +                vec = binbuf_getvec(b);
>              }
>          }
>          else
> -- 
> 2.7.0
> 





> _______________________________________________
> Pd-list at lists.iem.at mailing list
> UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list