Mail Wrapping

Dr. H. Nikolaus Schaller hns at goldelico.com
Sat Aug 14 15:41:06 CEST 2010


Am 14.08.2010 um 10:14 schrieb rixed at happyleptic.org:

> Please people stop spamming about line length.
> If you MUA is so good then ask it to automatically split long lines :-p

I agree that we should not spam - but IMHO this was raised as a serious
problem.

I could live with the idea that everyone uses a MUA that can
wrap lines when reading and displaying long lines. But it appears there
are systems out there that can't (which I did not yet know). 

And I am asked to solve their display problem on the senders side (although
I have much better things to do)...

Am 14.08.2010 um 09:28 schrieb Matthias Apitz:

> El día Saturday, August 14, 2010 a las 08:34:51AM +0200, Dr. H. Nikolaus Schaller escribió:
> 
>>> Nikolaus, couldn't you wrap your lines to something more standard (72 or
>>> so ?) Thanks, I like reading your prose, but those long lines are really
>>> irritating.
>> 
>> Hi Steve,
>> 
>> there are different opinions if the 80 char line wrapping of mail is standard or some
>> old-fashioned relict from the 80ies. I have tried to find out what it is but it appears
>> to be a problem with some MUAs not correctly handling RFC 2646.
>> 
>> Here is also some discussion about mailman being responsible or not:
> 	....
> 
> Hi Nikolaus,
> 
> It is the responsibility of your MailUserAgent to wrap lines correctly
> around column 72. You are using Apple Mail (2.1081). If this can not do
> this, just use another MUA or another system providing correct software.

Before we start fingerpointing on any client we are using, let's do a little more research.

http://mailformat.dan.info/body/linelength.html

quotes RFC 2822 (the successor to RFC 822):

  "There are two limits that this standard places on the number of characters in a line. Each line of characters MUST be no more than 998 characters, and SHOULD be no more than 78 characters, excluding the CRLF."

I.e. lines more than 78 characters are *not* forbidden. From this I conclude
that a mail recipient must cope with that. If not, the client is broken.

And, I conclude that it is not a rule for writing or displaying mails - just
for transferring them over SMTP without making buffer overflows.

Now let's look into the plain code my MUA is sending. Here is an example:

> Hi Steve,
> 
> there are different opinions if the 80 char line wrapping of mail is standa=
> rd or some
> old-fashioned relict from the 80ies. I have tried to find out what it is bu=
> t it appears
> to be a problem with some MUAs not correctly handling RFC 2646.
> 
> Here is also some discussion about mailman being responsible or not:
> 

So Apple Mail *is* correclty sending wrapped lines according to RFC.

I do not excactly know what the rules are to interpret the = signs at the
end of the line. I guess it has to do with 

Mime-Version: 1.0 (Apple Message framework v1081)
Content-Type: text/plain; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

A little more search shows this comes from RFC 2045 (MIME) on page 19 (Soft Line Breaks).

From this I conclude that the mails my client sends are correct (according
to the standard).

So it appears to me that we are trying to treat a corectly implemented
feature as a bug because there is some habit of expecting that lines
are always *displayed* with wrapping (even if they are sent as non-wrapped
lines using MIME quoted-printable). But I may be missing something.

If somebody can point me to the RFC that *requires* that lines must
be *visibly* wrapped by the sender so that no client ever shows
long lines, I am happy to file a bug report for Apple Mail (wouldn't
be the first one :).

If there is no such RFC, please report bugs for your clients that
don't wrap long (MIME encoded) lines they receive to the width of
the display window.

Nikolaus




More information about the community mailing list