From: Daniel Stenberg (daniel.stenberg@autodiagnos.se)
Date: Fri Jun 04 1999 - 01:07:00 CDT
Otis Gospodnetic wrote:
> > Yep, I'd say you found it. These later versions of hypermail are kinda
> > picky about message ids: any messages with duplicate or missing ids are
> > just simply dropped, with no warning.
> I saw that this was already talked about quite a bit, but I'm not sure I
saw
> a definite conclusion about how HM will deal with it in the future.
Paul has already written code that deals with these problems in a "nicer"
way. You're free to get it from the CVS server. I haven't considered it
major enough to release a new tar package.
> I know
> having no Message-ID is wrong to begin with, so one way to look at it
would
> be to say 'hey, it's not HM's fault, fix your mail reader/sender app', but
> that wouldn't help the maintainer of the archives much...
So how did you generate the mails in the first place? Real mails have
Message-IDs and they must have them according to RFC822. If you don't
have real mails, the program that generates the faked mails should
output message-ids as well as the other headers hypermail will look
for.
Ok, you may find some old mails from the past that already were parsed and
made with 1.02 and don't have message-ids. So, then you need that fix that
adds a random message-id to each mail without one.
> Also, I see that HM archive at landfield.com is using HM 2.0b4 - is that
> version better at handiling null message IDs?
Yes. Because it ignores empty message-IDs completely. I don't even think
you can get that 2.0b4 and you won't get any bugs fixed in it if you
report such...
> What is the safest solution? Go back to HM 1.02?
No, adjust the latest code to do as you want. Distribute the patch, discuss
solutions on this list and you'll take both your own project forward at the
same time as hypermail develops.
And, while I'm at it. I'll go away for a long summer vacation now. I won't
answer any mails or do any developing until August. If there's anything you
want from me before I leave, tell me now or wait! :-)
/ Daniel