Re: bug with the Message-ID ?

[Brian Stafford <brian stafford uklinux net>]

On Mon,  3 December 15:12 M. Thielker wrote:
> On 2001.12.03 15:41 Brian Stafford wrote:
>> Anyway, maybe if EHLO responds with "(desws061)" in the first response, 
>> auto deletion of troublesome headers could be enabled.  But even if I 
>> implement this, I would add a --enable- option to the configure script and 
>> it would always be off by default - I don't want to make this sort of thing 
>> easy.
>> 
> 
> Don't bank on that response. It's the internal name of the one host out of 
> the cluster they have doing mail that responded to your connection request. 
> Other's results _will_ vary.

I can live with Compuserve in the EHLO response meaning "don't bank on 
standards compliance".

> I, personally, after seeing all this talk, would prefer to have that 
> pseudo-protocol "CIS Mail" in there. If and when CIS mends it's ways it can 
> be dropped.

Autodetection is less obtrusive - especially if it can be made reliable.

> No configure option, please. The class of users who will have this problem 
> in the future will use binary packages, I'm sure! Some may not even install 
> the compiler on their system, they'll regard Linux as much as a black box as 
> they do windows.

But would such users see this?  Those making the binary distros would.  Then 
they would know that enabling the option probably breaks something else that 
*is* compliant and that I, the package's author, disapprove strongly of the 
code they are enabling.  (Assuming I wrote it, of course.)

>>> If the WIndows monopoly is to be broken, we need to think as users would, 
>>> not as programmers do.
>> 
>> Indeed.  And this is why I consider that standards are so important.  When 
>> everyone conforms all those niggling little things that need to be manually 
>> configured by confused punters simply just go away.
>> 
> No. Someone will always try to modify a standard. A program, from a "dumb" 
> user's point of view, just has to work. He doesn't care how, or why, it has 
> to work! If it doesn't he'll call support, e.g. nag this mailing list.

But surely all those options that tweak things away from standards are highly 
obscure and unlikely to be meaningful to punters anyway.

> It's that simple. We have found the reason for the problem with CIS, if we 
> refuse to fix it, Balsa will be that much less usable.

Well, there's not necessarily a refusal to fix it.  However it is *very* 
important that the actual fix, if attempted, is is the `correct' one.

RFC 2821/2822 are an attempt to get what was becoming a bad situation back 
under control.  RFC 2476 was also created for this reason.  For that reason, I 
am reluctant to ignore any SHOULD in those standards, even though I might have 
done with the old 821/822/974/1123 requirements.

Given that 2821/2822 were only published in April, not many implementations 
will have moved towards compliance yet.  Bear in mind we are amongst the 
first.  Eventually the situation will improve.  With a bit of luck CIS might 
even be motivated to fix things now the new standards are published - I doubt 
they react quickly.

> The reason this hasn't come up until now is also very simple: Not too many 
> simple users are on Linux yet, and the percentage of those who use CIS is 
> small. Anyone who knows a bit about computers and/or has a bit of common 
> sense will avoid CIS/AOL et al like the plague.

Indeed.  For now, though, I prefer to fix problems with standards compliance, 
not that I seem to have too many.  If only one or two users have CIS problems, 
that's tough - its not my priority.  AFAIK, nothing stops them getting a 
Yahoo! mail or other free mailbox account to post their mail through.  
Alternatively, run a local MTA on the workstation.  There are those on this 
list who have had good results with various MTAs and can probably offer better 
advice than I can on this.

> But, another class of users, without computer knowledge/common sense, is 
> part of the group of users that will use Linux/Gnome/Balsa in the future.

Personally, I feel that's a while off yet, though it will happen.

> They need to be able to handle it - without configure, without a text editor 
> and without CIS changing their software.

I doubt that class of user will ever see a configure script.

As to changing software, running broken non-compliant stuff is not exactly 
user friendly.  As it is, developers have a hard enough time handling all the 
requirements and inconsistencies in standards (think recent IMAP discussion) 
without having to second guess what insanity maverik implementations inflict 
on the world.  The problem here is you can never know when all the bugs are 
emulated.  There will always be users for whom things don't work.  The end 
result is that the developer (i.e. me) releases frequently, fails to make a 
universally workable implementation and draws even more abusive flaming.  It's 
a viscious circle.  Anyway, I give the code away for free.  I don't see why I 
should have to take flak over non-conformant implementations that have nothing 
to do with me.

>> I'm half inclined to advise CIS users to get a Yahoo! account.  Nothing
> 
> They don't know what Yahoo is, some of them....

Surely not?  Yahoo! has even got a mention on the BBC Six O'Clock News which 
these days is soo dumbed down that the reporters think M$ invented the 
internet in 1995.

>> - qmail's author is even more pedantic about standards than I am.  (Qmail
> 
> That's saying a lot! :)

Maybe, but few people experience problems with qmail.  And almost no punter 
knows of it, so that's a good recommendation for a MTA.

--
Brian