26904 - Format=Flowed is generated for message body with QP

Status: VERIFIED INVALID

(MailNews Core :: MIME, defect, P3)

Description

[nhottanscp]

Separated from a bug 26734.

The following is from RFC2546 page 5,

   [Quoted-Printable] encoding SHOULD NOT be used with Format=Flowed
   unless absolutely necessary (for example, non-US-ASCII (8-bit)
   characters over a strictly 7-bit transport such as unextended SMTP).
   In particular, a message SHOULD NOT be encoded in Quoted-Printable
   for the sole purpose of protecting the trailing space on flowed lines
   unless the body part is cryptographically signed or encrypted (see
   Section 4.6).

   The intent of Format=Flowed is to allow user agents to generate
   flowed text which is non-obnoxious when viewed as pure, raw
   Text/Plain (without any decoding); use of Quoted-Printable hinders
   this and may cause Format=Flowed to be rejected by end users.

We are putting Format=Flowed with bodies containing quoted-printable.
My proposal for beta1 is to check the strict_mime pref value in addition to 
check charset names ("us-ascii" and "ISO-8859-1") to decide whether to generate 
format=flowed. Not sure about the case which server converts to QP, need more 
info.

Comment 1

[John G. Myers]

The sentence is "SHOULD NOT be encoded ... for the sole purpose of protecting 
the trailing space..."

If there is a purpose for using QP other than protecting the trailing space, 
there is no problem with using QP and format=flowed together.

Comment 2

[Daniel Bratell]

[lowering severity]

Do the quoted printable-encoder touch trailing spaces?

Comment 3

[Gervase Markham [:gerv]]

Test comment (related to bug 26936). Apologies for the spam (but that's sort of
what Bug 26936 is about - big CC fields).

Gerv (using Mozilla M13 Win95)

Comment 4

[John G. Myers]

When the quoted-printable encoder encounters trailing spaces, it is required by
the spec to encode them.  That is because trailing spaces are ignored in the
quoted-printable encoding.

Comment 5

[John G. Myers]

I think this bug is INVALID.

Comment 6

[Daniel Bratell]

I'm marking it INVALID now. The specification says that we SHOULD NOT use 
format=flowed for QP which means that we shouldn't do _unless_ we really have 
considered what it means to mix the two. I would say we have done that now. If 
any further complications turns up, we just reopen the bug.

Comment 7

[Katsuhiko Momoi]

I'm going to verify this resolution. Given the frequency with which
body could be encoded in QP in ISO-8859-x text, the wording in the
RFC, i.e. "...unless absolutely necessary...", seems to be too
strong. I would hope that the author would soften the wording.

Comment 8

[Randall Gellens]

The wording means that QP encoding shouldn't be used solely to 
protect the trailing space.  If you have any legitimate reason to use 
QP (such as 8-bit content without 8BITMIME support, or cryptographic 
considerations), then there is no reason to avoid QP.

The wording does not say there are any problems using QP and F=F 
together.  The wording is only saying that use of F=F is not in and 
of itself cause to also use QP.