Supercite/Message woes

Supercite/Message woes

[Fabrice Gamberini]

Hello,

(NT-emacs 20.4, Gnus v5.8.7, Supercite.el 3.1)

I think I have a problem now with supercite, or maybe not, and I have been unable to find
any searchable mailing-list archive for this package. Maybe someone will
recognize  the symptoms. 

The problem appears after I kill/abort an email (reply/followup) I was
writing. If I compose a new message after that, each time the line wraps automatically,
the next line is indented, as if it was a citation (with the
attribution to the previous reply's message author).

here's an example: suppose I started a reply to 'Dodo', then C-c C-k the
message, whenever I type a new message, even if it is NOT a citation -- like
this message I'm typing -- it gets indented like this:

     Dodo> this is an example of this annoying problem that happens whenever
     Dodo> lines wrap...

instead of 

this is an example of this annoying problem that happens whenever
lines wrap...

Actually this is happening right now, and I'm not in SC mode, the status line
reads: (Message MML Filladapt Abbrev Fill)


thanks


-- 
Fabrice Gambérini
  -- = Wavecom S.A. = -- 
Email: fabrice.gamberini@wavecom.fr

Re: Supercite/Message woes

[Kai Großjohann]

On 04 Oct 2000, Fabrice Gamberini wrote:

> here's an example: suppose I started a reply to 'Dodo', then C-c C-k
> the message, whenever I type a new message, even if it is NOT a
> citation -- like this message I'm typing -- it gets indented like
> this:

Hm?  You start a reply to Dodo, then kill it, then start a completely
unrelated message, and it inherits the fill-prefix "Dodo> "?  Strange.

Is this an Emacs 20.4 bug, maybe?  Does it still happen with 20.7?

kai
-- 
I like BOTH kinds of music.

Re: Supercite/Message woes

[Fabrice Gamberini]

>>>>> "Kai" == Kai Großjohann <Kai.Grossjohann@CS.Uni-Dortmund.DE> writes:

    Kai> On 05 Oct 2000, Fabrice Gamberini wrote:
    >> I switched to 20.7, same problem. My description was not entirely
    >> correct, though. It only happens when the sentence I'm typing is
    >> "wrapping" at the end of Dod> the buffer line (just like this phrase,
    >> see that indent Dod> ?).

    Kai> I'm pretty sure the problem is that fill-prefix has the wrong value,
    Kai> just like I said.

    Kai> Hm.

    Kai> When you compose a message, take note of the buffer name.  Then, have
    Kai> a look at the list of buffers after doing `C-c C-k' -- is the message
    Kai> buffer gone?  And when you compose a new message, does that reuse the
    Kai> old buffer?

Kai,

  I have attempted a finer analysis of the problem, it really seems to be
  lying somewhere in-between bbdb and supercite.  I tried breakpointing on the
  filladapt paragraph functions and the 'fill-prefix' variable *has* the right
  value when I compose a reply, but it is not used by whatever piece of code
  flows the lines of text around... I think I'll send the problem to the
  supercite mailing-list, though it seems this package hasn't been updated in
  ages, maybe they'll have a clue.


  Thanks for trying anyway. I'm beginning to wonder if there's not too much
  elisp packages to play with in emacs... certainly too much for me to grasp
  at the moment. 


fabrice

-- 
Fabrice Gambérini
Email: fabrice.gamberini@wavecom.fr

Re: Supercite/Message woes

[Kai Großjohann]

On 05 Oct 2000, Fabrice Gamberini wrote:

>   I have attempted a finer analysis of the problem, it really seems
>   to be lying somewhere in-between bbdb and supercite.  I tried
>   breakpointing on the filladapt paragraph functions and the
>   'fill-prefix' variable *has* the right value when I compose a
>   reply, but it is not used by whatever piece of code flows the
>   lines of text around... I think I'll send the problem to the
>   supercite mailing-list, though it seems this package hasn't been
>   updated in ages, maybe they'll have a clue.

Hm.  Hmmm...  That's quite strange.  I just did the following test: I
typed "xxx" at the beginning of a line, then hit `C-x .'.  This set
the fill-prefix to "xxx".  Then I deleted the line and started typing
on a fresh line.  And sure enough, when I reached the end of the line
and auto-fill kicked in, the second line began with "xxx".

When you turn on adaptive fill and fill-prefix is nil, then Emacs
tries to figure out a fill prefix to use from the current buffer
contents.  In particular, it looks at the beginning of the current
paragraph for a fill prefix to use.  (Normally at the second line, but
for single-line paragraphs, it also looks at the first line.)  You can
type `ESC {' and `ESC }' to go backward/forward a paragraph.  This
also tells you where Emacs thinks a paragraph starts.

For me, blank lines delimit paragraphs, so when I type just below a SC
citation, the SC citation will be reused.  Like so:

  Foo> This is an SC citation line.
  Foo> The SC citation must be at least two lines long.
And here is what I type just below that citation.  The next line will
  Foo> inherit the fill prefix from the second line of this paragraph,
  Foo> and since this paragraph starts with two SC citation lines, the
  Foo> auto-filled lines also have SC citation prefixes.

See?

Maybe this is what happens to you?

kai
-- 
I like BOTH kinds of music.

Re: Supercite/Message woes

[Fabrice Gamberini]

>>>>> "Kai" == Kai Großjohann <Kai.Grossjohann@CS.Uni-Dortmund.DE> writes:

     FGA>> problem description --snipped--

    Kai> Hm.  Hmmm...  That's quite strange.  I just did the following test: I
    Kai> typed "xxx" at the beginning of a line, then hit `C-x .'.  This set
    Kai> the fill-prefix to "xxx".  Then I deleted the line and started typing
    Kai> on a fresh line.  And sure enough, when I reached the end of the line
    Kai> and auto-fill kicked in, the second line began with "xxx".

Did you try that with sc-auto-fill-region-p on ?? (I think it's C-c C-p C-t f)

Here the wrong-fill-prefix-problem seems to have disappeared after setting
sc-auto-fill-region-p to nil. This comes from re-reading the SuperCite
manual and experimenting with the stuff I didn't understand at first. Not that
I can claim to have a clear understanding of everything now, but at least it
works. Plus it doesn't auto-cite on text _between_ cited paragraphs, which I
think is a wrond thing to do, so it's actually better than what you describe,
I don't have to backspace on bl**dy indented paragraphs anymore. Me Happy.


    Kai> When you turn on adaptive fill and fill-prefix is nil, then Emacs
    Kai> tries to figure out a fill prefix to use from the current buffer
    Kai> contents.  In particular, it looks at the beginning of the current
    Kai> paragraph for a fill prefix to use.  (Normally at the second line,
    Kai> but for single-line paragraphs, it also looks at the first line.)
    Kai> You can type `ESC {' and `ESC }' to go backward/forward a paragraph.
    Kai> This also tells you where Emacs thinks a paragraph starts.

    Kai> For me, blank lines delimit paragraphs, so when I type just below a SC
    Kai> citation, the SC citation will be reused.  Like so:

    Foo> This is an SC citation line.
    Foo> The SC citation must be at least two lines long.
    Kai> And here is what I type just below that citation.  The next line will
    Foo> inherit the fill prefix from the second line of this paragraph,
    Foo> and since this paragraph starts with two SC citation lines, the
    Foo> auto-filled lines also have SC citation prefixes.

    Kai> See?
 Yes, this doesn't seem quite right to me though, it gets confusing to read,
isn't it ?
 The more I understand, the more I think this "feature" is A Bad
Thing. (sc-auto-fill, I mean)


    Kai> Maybe this is what happens to you?

Well, It might be something like this, but this happened in 2 _different_
messages, which was a source of extreme confusion and bafflement. All it took
was a single -documented though quite badly IMHO- variable  to solve this
misbehaviour.


    Kai> kai
    Kai> -- 
    Kai> I like BOTH kinds of music.


Thanks for looking this in.

Fabrice


-- 
Fabrice Gambérini
  -- = Wavecom S.A. = -- 
Email: fabrice.gamberini@wavecom.fr

Re: Supercite/Message woes

[Kai Großjohann]

On 06 Oct 2000, Fabrice Gamberini wrote:

> Did you try that with sc-auto-fill-region-p on ?? (I think it's C-c
> C-p C-t f)

I don't use SC at all!  Hence, the value of that variable is not
important.  (It is nil, if you must know.)

sc-auto-fill-region-p controls a behavior which does not have anything
to do with auto-fill.  The variable name is unfortunate.

Normally, auto-fill changes the behavior of the space key when you
type it to the right of fill-column.  But sc-auto-fill-region-p
means that SC hits M-q for you.  Ie, it is about normal filling, not
auto-fill.

Does anybody know of a good synonym for `auto' so that the ambiguity
can be resolved?

> Here the wrong-fill-prefix-problem seems to have disappeared after
> setting sc-auto-fill-region-p to nil.

Quite strange.  Hm.  I don't claim to understand what's going on.  In
particular, I gather, that you got a "Dodo>" prefix which did not
occur in that message at all.  Whee.

kai
-- 
I like BOTH kinds of music.