Hello Milan, I got some more information regarding a proper fix for this bug.
Regards, Martin ---------- Forwarded message ---------- Subject: Re: [Crm114-general] option to disable changing the message-id? Date: Donnerstag 25 Januar 2007 22:24 From: Paolo [...] To: [EMAIL PROTECTED] [...] > > OTOH, if it turns out that the *MSG-ID* itself is changed than we've > > a bad bug. > > Ah okay, it didn't see that initially... at here CRM114 just *adds* a > Message-Id. I thought it would overwrite it as I did not notice that > the original message-id header is still in place. it's supposed (I think...) to add a comment (the sfid) to the MSG-ID iif such field is already present. Since msg-id isn't required, mail*.crm should refrain to insist to have one (perhaps void) anyway. IIRC the idea was to look at other 'safe' places to tag with sfid, namely the in-reply-to, resent-msg-id, references, which MUA aren't supposed to touch nor discard. Unfortunatly some MUA drop RFC-comments from such fields, while other drop X-* fields altogether. Turns out that relying on presence of opt headers isn't safe unless you've control over which MUA are used. mail*.crm try some supposedly harmless redundancy as a workaround. > > so Debian shouldn't ship crm114 with patch suggested in #406207. > > Hmmm... I did not read those RFCs and can't comment on this. rfc822 and rfc2822. hmm, but it actually *adds* (void) msg-id, so we need a fix... not adding a comment shouldn't be the default, likewise adding msg-id anyway. So the right (imo) fix is: - drop patch in #406207 - ie leave maillib.crm:434 as shipped - comment out maillib.crm:374-378 (the 'alius' in func :mungmail_add_comment:) that way, if header field is present, it gets a comment, else nothing is done. ---------- Forwarded message ---------- -- Martin 'Helios' Steigerwald - http://www.Lichtvoll.de GPG: 03B0 0D6C 0040 0710 4AFA B82F 991B EAAC A599 84C7
pgpcMOuPmc0L9.pgp
Description: PGP signature
