Picon related bugs--anyone else see these?

Picon related bugs--anyone else see these?

[Sudish Joseph]

There's a bug in the picon stuff that only occurs when you visit the
very first article in a session...and only once in every 10 sessions
or so.  I feel quite sure that this will turn out to be a heisenbug of
the worst type.

The article buffer comes up with the picons corresponding to the group
(not the address stuff) displayed at the beginning *and* end of
*every* header line.  The only useful data I have is that when this
occurs the address picons aren't displayed at all.

Once in a *very* rare while gnus'll hang when displaying an article
that contains picons.  This occurs after the headers have been
highlighted but before they've been hidden.  It's not an infloop,
which leads me to believe that it's a bug in XEmacs.

I'm guessing that both these are XEmacs bugs, given their infrequency
and given that I'm seeing similar problems in w3/tm as well.  Anyone
else seen these?

-Sudish

Re: Picon related bugs--anyone else see these?

[Wes Hardaker]

Sudish Joseph <sudish@mindspring.com> writes:

> The article buffer comes up with the picons corresponding to the group
> (not the address stuff) displayed at the beginning *and* end of
> *every* header line.  The only useful data I have is that when this
> occurs the address picons aren't displayed at all.

Yikes...  I can't imagine how this would happen.  Nor can I imagine
how I would reproduce it.  What do you have gnus-picons-display-where
set to?  I assume 'article?

Very odd...

Of course, its even harder for me to test now that I'm in switzerland
and haven't installed the picons stuff here (**gasp**) and am not even
using all that gnus-picons code I wrote (**big gasp**).

Wes

Re: Picon related bugs--anyone else see these?

[Sudish Joseph]

In article <qk2d90xjf0y.fsf@sphys.unil.ch>,
Wes Hardaker <Wesley.Hardaker@sphys.unil.ch> writes:
> What do you have gnus-picons-display-where set to?  I assume
> 'article?

Yup.  My picon settings are:

;; picons
(setq gnus-use-picons t
      gnus-picons-database "/usr/lib/picons"
      gnus-picons-display-where 'article)
(add-hook 'gnus-article-display-hook 'gnus-article-display-picons t)
(add-hook 'gnus-summary-prepare-hook 'gnus-group-display-picons t)
(add-hook 'gnus-article-display-hook 'gnus-picons-article-display-x-face)

I guess I shouldn't use gnus-group-display-picons.  Still, it doesn't
explain why I see this bug (possibly in XEmacs, not your stuff) so
infrequently. 

-Sudish

Re: Picon related bugs--anyone else see these?

[Lars Magne Ingebrigtsen]

Sudish Joseph <sudish@mindspring.com> writes:

> I guess I shouldn't use gnus-group-display-picons.  Still, it doesn't
> explain why I see this bug (possibly in XEmacs, not your stuff) so
> infrequently. 

I've seen various XEmacs glyph bugs once in a while.  It seems that if
there's nore glyphs on a single line than there's room for, XEmacs
goes wild and puts the glyph *all* over the place.

-- 
  "Yes.  The journey through the human heart 
     would have to wait until some other time."

Re: Picon related bugs--anyone else see these?

[John Griffith]

I just started using picons.  I see this bug as well (XEmacs 19.14).
I have essentially the same settings as Sudish.

I also notice the picon display is slow.  When selecting an article,
the buffers get messed up for about half a second before stabilizing.
It could be because of searching for picon files across NFS, or it
could just be slow drawing on XEmacs' part.

Re: Picon related bugs--anyone else see these?

[Wesley.Hardaker]

John Griffith <griffith@sfs.nphil.uni-tuebingen.de> writes:

> I also notice the picon display is slow.  When selecting an article,
> the buffers get messed up for about half a second before stabilizing.
> It could be because of searching for picon files across NFS, or it
> could just be slow drawing on XEmacs' part.

Well, XEmacs isn't really much of a picture editor, so its not on its
high speed priority list.  However, yes, loading it over NFS will
definately make it worse.  Also, elisp is not really the best language
for speed either, so you'll loose some there as well.  However, there
is not much that can be done to speed things up, so I'm afraid you're
stuck for now...

Wes

Re: Picon related bugs--anyone else see these?

[Wesley.Hardaker]

Wesley.Hardaker@sphys.unil.ch writes:

> Well, XEmacs isn't really much of a picture editor, so its not on its
> high speed priority list.  However, yes, loading it over NFS will
> definately make it worse.  Also, elisp is not really the best language
> for speed either, so you'll loose some there as well.  However, there
> is not much that can be done to speed things up, so I'm afraid you're
> stuck for now...

Another speed up is to display the picons buffer in a seperate emacs
frame, not controlled by gnus.  This way gnus draws the
summary/article (and whatever else) really quickly and the icons seem
to actually get drawn after everything else, allowing you to read the
article quickly.

Wes

Re: Picon related bugs--anyone else see these?

[Wesley.Hardaker]

Sudish Joseph <sudish@mindspring.com> writes:

> Yup.  My picon settings are:
> 
> ;; picons
> (setq gnus-use-picons t
>       gnus-picons-database "/usr/lib/picons"
>       gnus-picons-display-where 'article)
> (add-hook 'gnus-article-display-hook 'gnus-article-display-picons t)
> (add-hook 'gnus-summary-prepare-hook 'gnus-group-display-picons t)
> (add-hook 'gnus-article-display-hook 'gnus-picons-article-display-x-face)

Looks about right, but I haven't displayed picons in the article
buffer in a while...  I'll have to switch to that for a bit and see if
I see the problem as well.  I finally installed the picons database
yesterday, so I can actually see them now (yay).  I suppose I should
re-submit my face for my new address now!

> I guess I shouldn't use gnus-group-display-picons.  Still, it doesn't
> explain why I see this bug (possibly in XEmacs, not your stuff) so
> infrequently. 

Well, I'm not sure what the problem could be.  The code is barely mine
anymore, as Lars almost completely re-wrote it as he was working on it
(which is good, as he's a much more proficient elisp programmer than
I).  Of course, he did reverse the icon order...  I should go change
it back and send him a patch and see if he notices ;->.  

I'd doubt the bug would be in XEmacs, but you never know...

Wes