On Mon, 2006-03-13 at 11:13 +0100, Michael Haardt wrote:
> Once I had a mail with a broken MIME structure. I sent it to a couple
> people, curious what their client would show. I was very surprised
> when someone said: The MIME structure is fine, what's the point? Indeed
> Evolution repaired it, and saying "show message source" showed the
> repaired version - not the source present on the IMAP server.
Ah, right. Yeah, that's another manifestation of the same problem
really. Evolution mangles stuff on the way in, and it shouldn't.
> And since clients should show headers unfolded in the regular display,
> folding with a tab causes ugly results. Evolution implements *that*
> right and many other clients don't.
I disagree with this, because I think you're misreading RFC2822. AFAICT
it doesn't say that MUAs should _display_ unfolded headers. It says
(§2.2.3) that "[each] header field should be treated in its unfolded
form for further syntactic and semantic evaluation". That's all very
well, but doesn't mean you should _display_ it unfolded.
Evolution is the only client which I know of that does so, and it's
merely a symptom of a known bug -- a bug which also causes the other
problems you've described, and which has also in the past caused
Evolution to reduce the following:
To: Some people I invited to my party : ;
... to an empty 'To:' header.
Certain parts of the mail may well be _syntactically_ and _semantically_
irrelevant (or erroneous in your example), but that doesn't mean that it
shouldn't be displayed to the user as it arrived.
> That's why I say: Let's use spaces to fold headers, not tabs, being
> nice to users. And an amazing amount of users looks at "Received:"
> these days.
That would be a logical conclusion _if_ I were to accept that it makes
sense for a MUA to mangle what it receives. But I believe that an MUA
should display its input faithfully, without gratuitous mangling. So
there's no need to pander to Evolution by using spaces instead of tabs.
When displayed properly, tabs look better.
--
dwmw2
