[Mailman-Developers] ASCII Prefixes

[Barry A. Warsaw]

>>>>> "BAW" == Barry A Warsaw <barry@zope.com> writes:

    BAW> OTOH, we could use a simple heuristic to try to keep the
    BAW> prefix unencoded in the common case: if the prefix contains
    BAW> only ascii characters, we could choose to leave it in the
    BAW> default us-ascii (or really the None) encoding.  If it
    BAW> contains non-ascii characters, we'd have no choice but to
    BAW> encode it.

I just realized there's one unfortunate side effect of this choice.
As Ben is fond of pointing out, the RFCs are vague at what should
happen to spaces between non-encoded text and encoded words.  It isn't
specified whether that whitespace is significant (as in unencoded
text) or ignored (as inbetween encoded words).  So this Subject: line

    Subject: [Foo] =?iso-8859-1?Q?accentu=E9_that_n?= =?iso-8859-1?Q?eeds_m=F4re_sp=E0ces?=

may be displayed in your MUA as

    Subject: [Foo] accentué that needs môre spàces

or

    Subject: [Foo]accentué that needs môre spàces

I'm not sure what to do about that.
-Barry