On Thu, Apr 26, 2018 at 05:14:27PM -0700, Ammon Riley wrote: > > The =b/=B/=h are explicity mentioned because of their IMAP behavior. > > I did copy the =b/=B. I hadn't considered IMAP for this feature, as I'm > not using it. Since we have to parse the message to match content-type, > how would this behave under IMAP? Would it work on the server, or > does it have to be local? If it can work on the server, then perhaps I > should distinguish that -- I can imagine an IMAP user might not want to > download large PDF-containing messages while performing this limit.
It will download the message. I'll have to check myself if there is
some way to do it server-side.
> Updated patch attached.
Your new patch was too fast. :-) I realized I forgot to include one
other comment, below. I have to run, but I'll take another closer look
at the revised patch later tonight.
> diff --git a/pattern.c b/pattern.c
> +static int match_content_type(const pattern_t* pat, BODY *b)
> +{
> + char buffer[STRING];
> + if (!b)
> + return 0;
> +
> + if (snprintf(buffer, STRING, "%s/%s", TYPE (b), b->subtype) >= STRING)
> + buffer[STRING-1] = '\0';
snprintf (unlike strncpy) will always add the terminating null byte.
--
Kevin J. McCarthy
GPG Fingerprint: 8975 A9B3 3AA3 7910 385C 5308 ADEF 7684 8031 6BDA
signature.asc
Description: PGP signature
