On Wednesday 15 October 2008 23:00:39 Wietse Venema wrote: > Gaute Amundsen: > > On Wednesday 15 October 2008 14:03:06 Wietse Venema wrote: > > > Gaute Amundsen: > > <snip> > Postfix looks at the first input line. There currently is no > way to override this, so your best bet is to use the same line > terminator consistently (having a first line with CRLF might > work "best" for hybrid mail, but that behavior is not promised).
Ah! Good. As I just said: >With some luck that would mean that only my application was inconsistent, and >that all the other old stuff running on the box _could_ be ok. > The current behavior originates from the time when binary tranparency > was considered a good thing. Sort of suspected that. <snip> > > At this point I could vent some hostile comments about banging away > at the keyboard and declaring victory when the last error message > disappears, but let's leave that for a different discussion. Thick books are even sold, advocating this practice, I believe, so trying to put a stop to that may be sort of an uphill battle.. If engineers ruled the world... :-) <snip> > > The familiar complaint: it "works" with Sendmail, so it must be right. > Not that it must be right, but that is it unreasonable to make every user suffer for other users oversights, and with no way out. Having to discover that you had to enable the "sendmail_EOL_workalike" flag would be plenty of punishment for using other peoples less than consistent code I think. <snip> > > After several years of deployment Postfix has some clout, and I am > no longer bending over backwards in order to forgive every > inconsistency in other mail handling software. > > Wietse I see. I found a patch that said: "The idea is to be liberal in what we accept, strict in what we send." But that was from way back in 2004. personally I'm not much good at C, I stick to python as much as I am able, but I guess I ought to head over to php.net and post this info to the relevant places at least. Thanks for the clarification. Regards Gaute Amundsen
