From: Bruce Lilly (blilly@erols.com)
Date: Mon Apr 28 2003 - 14:41:13 CDT
Henry Spencer wrote:
> Like it or not, trailing whitespace is *not* reliably
> preserved in Usenet transmission and storage, and it's best not to depend
> on it, either for continuation parsing or for the obligatory SP.
The critical question on the table is whether there is in fact any
software still in use which *does* depend on either the existence of
the SP after colon or on something between that and the first line folding
of a continued field.
> The
> son-of-1036 rules in this area were designed to guarantee that trailing
> whitespace was never significant in headers.
There's no such significance in messages *other* than the 1036
restrictions. Let's not deviate from 822/2822 unless there's still
a valid reason for doing so.
So far nobody's piped up and said "software package XYZ prints 'the
end is nigh' and reformats the hard disk on encountering something
other than a SP after the colon". I can only assume -- in the absence
of some such report -- that it's time for the sun to set on that
particular set of restrictions.