Pending patches for INN?

[Richard Kettlewell]

On 2015-05-23 13:06, Julien ÉLIE wrote:
> Hi Richard,
> 
>> I would suggest NOT delaying any release for my benefit - I devote time
>> to this in very unpredictable ways.
> 
> OK, noted.  Thanks again for your investment in improving
> the robustness of INN.

No problem!

> Regarding your first patch for innxmit, shouldn't
>   static char *buff;
> be changed to
>   static char *buff = NULL;
> as you no longer use xmalloc, but only xrealloc?

Objects with static storage duration are initialized automatically, in
this case to a null pointer (C99 6.7.8#10) - so the two lines are
equivalent.

> If *buff is not initialized to NULL, it seems that it could
> trigger an issue off with xrealloc trying to expand memory
> from a wrong location.
> 
> 
> 
> 
> As for actsync.c, maybe we could add the following test at the beginning
> of get_active()?
> 
> @@ -742,7 +742,9 @@
>      if (len == NULL)
>          die("internal error #1: len is NULL");
>      if (errs == NULL)
> -        die("internal error #2: errs in NULL");
> +        die("internal error #2: errs is NULL");
> +    if (host == NULL)
> +        die("internal error #10: host is NULL");
>      if (D_BUG)
>          warn("STATUS: obtaining active file from %s", host);

Yes, that wouldn't hurt.

ttfn/rjk