From: John Moreno (phenix@interpath.com)
Date: Thu May 13 1999 - 13:30:49 CDT
Charles Lindsey <chl@clw.cs.man.ac.uk> wrote:
> John Moreno <phenix@interpath.com> writes:
>
> >I've recently seen a problem with folding and Netscape -- instead of
> >replacing the space with CRLFSP it instead *inserts* CRLFSP. This was
> >happening in the subject and really screwing things up, as it introduced
> >undefined behavior -- specifically what to do with trailing white space
> >and how to unfold if there is more than one space at the beginning of
> >the line.
>
> >What I'd suggest is that if there is more than one space at the point of
> >folding that a HTAB be used instead (and vice-versa of course). So
> >that:
>
> >X-Example: This is an example
> > of a line with 2 SP's to be preserved.
>
> >Can be correctly turned into
> >X-Example: This is an example of a line with 2 SP's to be preserved.
>
>
> Hmmm! I am not over-convinced, nor quite sure exactly what you are
> suggesting. Is the number of Tabs supposed to correspond to the number of
> spaces that are supposed to be there, or merely to indicate that the
> trailing SPs on the preceding line (n-1 of them) are to be preserved?
No, what I meant to indicate (the trailing space was a typo which
confuses things, sorry) was that LEADING spaces would be preserved on
the continuation line by stripping off the tabs (and vice-versa).
Let's try it without the actual characters (\s=space, \t=tab)
X-Example: This is an example
\t\s\sof a line with 2 SP's to be preserved.
That get's turned into:
X-Example: This is an example\s\sof a line with 2 SP's to be preserved.
> Would it not be simpler to say that they should be preserved regardless
> (and if some transport loses them, well that is tough)?
I'd prefer something more definite, that didn't rely upon transports not
changing things. This is actually easier to check -- I read the current
rules as meaning:
X-Example: This
\s\t\s\t\s\tline was folded
should be unfolded to either:
X-Example: This\sline was folded
OR
X-Example: This\tline was folded.
Let me see if I can word this like I'd like it in the draft.
When unfolding, a run of SPs or TABs are collapsed into a single SP
or TAB unless the next character is a SP or TAB, in which case the
run is simply deleted.
That's not quite right because it doesn't make clear that I don't mean
SPTABSPTAB as a run, but it's close.
> >-snip->
> >> 4.3.2. Body Conventions
> >>
> >-snip-
> >> 4.4.2. Character Sets within Article Bodies
> >>
> >-snip-
> >> precise spacing is essential. Formfeed signifies a point at which a
> >> reading agent SHOULD pause and await reader interaction before
> >> displaying further text.
>
> >I'd suggest that this be moved up to the Body Convention section and
> >that the fact that it's called a "Spoiler Char" in that context be
> >mentioned.
>
> I have actually move that _whole_ paragraph to the "Body Format Issues"
> section, and it now says
>
> "Formfeed (which is sometimes referred to as the "spoiler character")
> signifies a point ..."
Sounds fine.
> >> 4.5. Size Limits
-snip-
> >We should have already defined HOW to fold, no reason to go into it
> >again, except possible to give a section number. And there are headers
> >where this isn't generally desirable for one reason or another (Subject
> >and References in particular).
>
> On reviewing that section, I see that I have described the limits on
> headers after the limits on bodies, and that the wording for the bodies
> bit is better than the header wording.
>
> So I have reversed the order, and rewritten the header paragraph as
> follows:
>
> Posting agents SHOULD endeavour to keep all header lines, so far as is
> possible, within 79 characters by folding them at suitable places (see
> ...). However, posting agents MUST permit the poster to include longer
> headers if he so insists, and compliant software MUST support headers of
> at least 998 octets. Likewise, injecting agents SHOULD fold any headers
> generated automatically by themselves. Relaying agents MUST NOT fold
> headers (i.e. they must pass on the folding as received).
How about a mention in the section on folding that some headers allow
folding but recommend against it or have different recommended minimum
lengths?
-snip-
-- John Moreno