This repsonse really has only half to do with Courier, but when you really get down tobusiness, most binary packages distributed with any distribution (especially debian)do not work or behave exactly as the "stock" configure, make,
make install, would have thembehaving. There are few packages that don't include some
sort of tweak, changed"defaults", initscripts, location of conffiles, etc.

If I intend to compile and install a piece of software from source code that is destined for a server with a largenumber of users, it is in my best interest to be intimately familiar
with the operation ofthat software. I usually start by looking at the Debian
packager's diff (even if I'm notcompiling for debian) as they often contain easy and tested
alterations I like. A lot ofthese modifications never go (and don't even belong) upstream.

As far as the problem of a patch causing "an extra iteration of attention from the tech staff", I would have tosuggest looking into getting a tech staff that considers patches
and staged test setupsstandard procedure when installing software. Then, at least
you are paying for an extrahour of their time versus countless money wasted on support staff
trying to BS a response ina situation where the customer is probably right on the
money!

Now the ontoppic part --

Personally, I support Sam coding the software however he wants. If the style is toopurist for you, then patch, or use something else. Start a community
patch to include the"non-standard" features people want that Sam doesnt want to
include. Lots of projects havesomething like this. It seems a lot of the anger here is
misdirected at Sam for notimmediately including a feature that was never there in the first
place, simply becausethere is a hack for it. Mail used to BOUNCE! Error messages were
hard coded in the sourcefile! Now it gets delivered in an ultra-sane and paranoid
manner.

Think writing a web browser. It's easiest to start from a point at which your browser is a 100% correct parserfor perfectly formated and versioned HTML, then add the
features that allow it to deal withthe quirks of webpages in the real world: missing tags,
bugs in popular browsers, mixedhtml DTD versions, etc. This courier situation is no different.

Please use your energy wisely. Instead of complaining, why not expand the 4 line patch into something that isactually complete and includable! Maybe it adds the 8 bit MIME
header and encodes or adds an"X-Warning: Not 7-bit clean" header to messages that violate
the rule if the option is on.Maybe it includes the courierwebadmin code to toggle this
setting, and the documentationto explain it to people. Maybe it puts the entry in the sample conf
files. Maybe it givesoptions for bypassing or skipping other RFC checks too. It's not
like there aren't optionsin courier already that aren't exactly 100% RFC correct (See
the BOFH* options forinstance), It's just not on Sam's agenda right now, but it's not like
you can't ever expectthe option to be there.

I just wish everyone would chill out and stop griping, already. There really is no point, and all you're going to do is piss Sam off to the point at which hedoesn't care anymore. Then you really will have somehting
to complain about.

Kudos to you, Sam for your software. It is definately the best system for me, and the tiny patches Iapply now are nothing compared to the monster hacks I had to
use on other software to achievethe same degree of functionality! Thanks.