Romain Francoise <[EMAIL PROTECTED]> wrote:

> Sorry for the occasional breakage -- I do my best to make emacs-snapshot
> as stable as any Emacs release but even though I read emacs-diffs daily,
> read the debdiff of each snapshot, and test each version thoroughly
> (with my setup), some changes still slip under my radar (like the
> hi-lock bug last week, or this filling issue).

<p>Some oversights are inevitable, and they are, in my opinion, more
then compensated for by responsiveness.  I have actually found this
emacs to be quite stable.</p>

> Back to this bug: I grepped through your configuration, and it could be
> caused by any (or both) of the following customizations:
> - next-line-add-newlines set to t (it's nil here)
> - this substitution in q-message.el:
>
>   (substitute-key-definition
>    'next-line 'mail-abbrev-next-line
>    message-mode-map global-map
>    )

<p>I don't actually see why this would make a difference -- I'd think
using the regular next-line function would still result in the same
buffer contents before sending.  The buffer contents look right to me
before I send the message.  It's only after I get my copy back that I
see the problem.</p>

<p>Oddly, I am having some trouble reproducing this bug now.  One
message I sent out had the problem.  I did a dist-upgrade after
sending that message, but no emacs-related packages changed.  The
locales package changed as did libc6, but nothing else did.  It would
be really surprising if this were a locale or libc related issue.</p>

<p>Anyway, I tried removing the substitute-key-definition and
restarted emacs.  The problem was not there.  Then I put the offending
lines back, and the problem was still not there.  There was an
intervening restart of emacs with the new libc and locales package,
but that's it.  Nothing else changed.  Maybe this bug is like the last
one: something in my environment that gets triggered over time breaks
it.  I'll have to keep an eye out for this.  I've wrapped paragraphs
here with "p tags" to indicate where my paragraph boundaries are in
case this is the lucky message that gets mangled. :-)</p>

-- 
Jay Berkenbilt <[EMAIL PROTECTED]>


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Reply via email to