On Sun, Aug 25, 2019 at 07:01:01AM +0200, Martijn van Duren wrote:
> Right now all we get is "Misbehaving filter", which doesn't tell the
> developer a lot.
>
> Diff below does the following:
> - Make the register_hooks actually fail on misbehaviour.
> - Change some *ep = 0 to a more semantically correct ep[0] = '\0'
> - Hoist some checks from lka_filter_process_response to processor_io
> - Make lka_filter_process_response void (it fatals now)
> - strtoull returns ULLONG_MAX on error, not ULONG_MAX
> - change str{,n}casecmp to str{,n}cmp for consistency.
> - change an fugly "nextline:" "goto nextline" loop to an actual while.
> - restructure lka_filter_process_response to be shorter.
> and most importantly
> - Add a descriptive fatal() minefield.
>
> -12 LoC.
>
> Tested with filter-dnsbl with both a successful and rejected connection.
>
> OK?
just one remark.
> Index: lka_filter.c
> ===================================================================
> RCS file: /cvs/src/usr.sbin/smtpd/lka_filter.c,v
> retrieving revision 1.41
> diff -u -p -r1.41 lka_filter.c
> --- lka_filter.c 18 Aug 2019 16:52:02 -0000 1.41
> +++ lka_filter.c 25 Aug 2019 04:47:47 -0000
> @@ -429,88 +429,68 @@ lka_filter_process_response(const char *
> struct filter_session *fs;
>
> (void)strlcpy(buffer, line, sizeof buffer);
> - if ((ep = strchr(buffer, '|')) == NULL)
> - return 0;
> - *ep = 0;
> + ep = strchr(buffer, '|');
> + ep[0] = '\0';
is it possible to buffer to not have '|' ? if yes, you could deference NULL.
--
Sebastien Marie
