delay when quitting digest

delay when quitting digest

[John H Palmieri]

Pterodactyl Gnus v0.76, GNU Emacs 20.3.1 (sparc-sun-solaris2.5.1, X
toolkit) of Thu Aug 27 1998 on darwin

Admittedly, my computer is slow in general, but depending on the
circumstances, there is a noticeable delay when quitting digests.
Here's the situation:

Background: I have gnus-auto-select-first set to nil in
gnus-select-group-hook, so I don't automatically get an *Article*
buffer when I enter a group.

Case 1: I enter a mail group which contains a digest, then go to the
digest and hit `C-d' to look at it.  Then when I hit `q', I return to
the *Summary* buffer almost instantaneously.

Case 2: I enter a mail group which contains a digest, read a
non-digest article first (so there is an *Article* buffer around),
then go to the digest and hit `C-d' to look at it.  Then when I hit
`q', there is a delay (perhaps several seconds--the length depends on
the size of the digest), and then I return to the *Summary* buffer,
which now shows the digest in a single *Article* buffer.

First, what's causing this delay?  Probably reading the article from
disk and parsing it for highlighting, etc.

Is there a hook I can set to fix this (maybe so that when I enter or
leave a digest, the *Article* buffer associated to the parent
*Summary* buffer is deleted, or maybe to turn off article highlighting
just for digests)?  Or is there some other fix?  Is there a good
reason to automatically display the digest in an *Article* buffer
after having read it as a digest, or when quitting a digest should the
default be to just go to the *Summary* buffer?

-- 
John H. Palmieri                       e-mail: palmieri.2@nd.edu
$400 million in gold bullion Honduras Ft. Bragg genetic assassination
CIA KGB supercomputer Semtex quiche plutonium counter-intelligence
class struggle smuggle FBI

Re: delay when quitting digest

[Lars Magne Ingebrigtsen]

John H Palmieri <John.H.Palmieri.2@nd.edu> writes:

> First, what's causing this delay?  

Try `(setq debug-on-quit t)' and then `C-g' while it's working.

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: delay when quitting digest

[John H Palmieri]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

> John H Palmieri <John.H.Palmieri.2@nd.edu> writes:
> 
> > First, what's causing this delay?  
> 
> Try `(setq debug-on-quit t)' and then `C-g' while it's working.

I did this several times, just to check.  Each time I hit C-g, the
backtrace looks like this (with variations in the regexp being
searched for):

  re-search-forward("\\bmailto:\\([^ \n	]+\\)" nil t)
  gnus-article-add-buttons()
  gnus-treat-article(nil 1 1)
  gnus-display-mime()
  gnus-article-prepare-display()
  gnus-article-prepare(1574 nil)
  gnus-summary-display-article(1574 nil)
  gnus-summary-select-article(nil force)
  gnus-summary-show-article()
  gnus-handle-ephemeral-exit((#<buffer *Summary nnml:mail.linux*> . article))
  gnus-summary-exit()
  call-interactively(gnus-summary-exit)

So I guess if I've just looked at a digest, I want it not to be
displayed in the Article buffer, so that it doesn't get "prepared" or
"treated" or something.  So maybe I want to change the state of the
old summary buffer when leaving the ephemeral group?


By the way, there is also a delay of several seconds when entering a
digest.  I'm talking about digests of at least 20 articles, totaling
at least 1000 lines.  In this case, I get essentially the same
backtrace:

  re-search-forward("\\bmailto:\\([^ \n	]+\\)" nil t)
  gnus-article-add-buttons()
  gnus-treat-article(nil 1 1)
  gnus-display-mime()
  gnus-article-prepare-display()
  gnus-article-prepare(1574 nil)
  gnus-summary-display-article(1574 nil)
  gnus-summary-select-article()
  gnus-summary-enter-digest-group(nil)
  call-interactively(gnus-summary-enter-digest-group)

This may be unavoidable, and it may just be a consequence of a slow
computer, slow disk access, and long digests.  Or is it avoidable?  (I
guess I could unsubscribe from the mailing lists that send me the long
digests, but that wasn't what I had in mind...)

-- 
John H. Palmieri                         palmieri.2@nd.edu

I pretend I'm living in a styrofoam packing crate, high in th'
 SWISS ALPS, still unable to accept th' idea of TOUCH-TONE DIALING!!

Re: delay when quitting digest

[Lars Magne Ingebrigtsen]

John H Palmieri <John.H.Palmieri.2@nd.edu> writes:

> I did this several times, just to check.  Each time I hit C-g, the
> backtrace looks like this (with variations in the regexp being
> searched for):
> 
>   re-search-forward("\\bmailto:\\([^ \n	]+\\)" nil t)
>   gnus-article-add-buttons()

It was always in `gnus-article-add-buttons'?  Hm.  Well, if it was a
big article, buttonizing it might take a while...

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: delay when quitting digest

[John H Palmieri]

Back to my original question:


John H Palmieri <John.H.Palmieri.2@nd.edu> writes:

> There is a noticeable delay when quitting digests.
> Here's the situation:
> 
> Background: I have gnus-auto-select-first set to nil in
> gnus-select-group-hook, so I don't automatically get an *Article*
> buffer when I enter a group.
> 
> Case 1: I enter a mail group which contains a digest, then go to the
> digest and hit `C-d' to look at it.  Then when I hit `q', I return to
> the *Summary* buffer almost instantaneously.
> 
> Case 2: I enter a mail group which contains a digest, read a
> non-digest article first (so there is an *Article* buffer around),
> then go to the digest and hit `C-d' to look at it.  Then when I hit
> `q', there is a delay (perhaps several seconds--the length depends on
> the size of the digest), and then I return to the *Summary* buffer,
> which now shows the digest in a single *Article* buffer.

The delay is apparently caused by buttonizing the digest (to show it
in the *Article* buffer).  My question:

What's the easiest way to avoid this delay?  What can I add to which
hook so that when I leave the digest, the *Article* buffer associated
to the parent *Summary* buffer is deleted, so that the article is
never treated for buttons or anything else?  Or should I take some
other approach?

-- 
John H. Palmieri                         palmieri.2@nd.edu

Being a BALD HERO is almost as FESTIVE as a TATTOOED KNOCKWURST.

Re: delay when quitting digest

[Lars Magne Ingebrigtsen]

John H Palmieri <John.H.Palmieri.2@nd.edu> writes:

> What's the easiest way to avoid this delay?  What can I add to which
> hook so that when I leave the digest, the *Article* buffer associated
> to the parent *Summary* buffer is deleted, so that the article is
> never treated for buttons or anything else?  Or should I take some
> other approach?

When leaving an nndoc group, Gnus tries to go to the next article, and 
if there is no next article, it reselects the current one...

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: delay when quitting digest

[John H Palmieri]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

> John H Palmieri <John.H.Palmieri.2@nd.edu> writes:
> 
> > What's the easiest way to avoid this delay?  What can I add to which
> > hook so that when I leave the digest, the *Article* buffer associated
> > to the parent *Summary* buffer is deleted, so that the article is
> > never treated for buttons or anything else?  Or should I take some
> > other approach?
> 
> When leaving an nndoc group, Gnus tries to go to the next article, and 
> if there is no next article, it reselects the current one...

But this only seems to be true if there was an article selected before
entering the nndoc group.  (I have gnus-auto-select-first set to nil,
in case that makes a difference.)  If I don't select any articles, but
go straight to the nndoc group with C-d, then when I quit the nndoc
group, it doesn't select the next article.  How do I get this behavior
even if I have read an article before entering the nndoc group?

Select an article, enter the nndoc group, leave the nndoc group -->
  long delay when leaving the nndoc group (because it's buttonizing
  the long article containing the digest)

Don't select any articles (just enter the mail group from the Group
  buffer with SPC, with gnus-auto-select-first set to nil), enter the
  nndoc group, leave the nndoc group --> no delay, because it's not
  selecting and buttonizing any article. 

[I feel thoroughly confused, so pardon me if I'm asking the same thing
over and over again...]

-- 
John H. Palmieri
e-mail: palmieri.2@nd.edu             205 Computing/Math Bldg.
URL:http://www.nd.edu/~jpalmier/      University of Notre Dame
(219) 631-8846                        Notre Dame, IN 46556

Re: delay when quitting digest

[Lars Magne Ingebrigtsen]

John H Palmieri <John.H.Palmieri.2@nd.edu> writes:

> But this only seems to be true if there was an article selected before
> entering the nndoc group. 

You have always selected an article before entering an nndoc group,
because `C-d' selects the current article.  :-)

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: delay when quitting digest

[John H Palmieri]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

> John H Palmieri <John.H.Palmieri.2@nd.edu> writes:
> 
> > But this only seems to be true if there was an article selected before
> > entering the nndoc group. 
> 
> You have always selected an article before entering an nndoc group,
> because `C-d' selects the current article.  :-)

Then why is there a delay (when quitting a digest) only if I've read
another article in that group before opening up the digest?

-- 
John H. Palmieri
e-mail: palmieri.2@nd.edu             205 Computing/Math Bldg.
URL:http://www.nd.edu/~jpalmier/      University of Notre Dame
(219) 631-8846                        Notre Dame, IN 46556

Re: delay when quitting digest

[Lars Magne Ingebrigtsen]

John H Palmieri <John.H.Palmieri.2@nd.edu> writes:

> Then why is there a delay (when quitting a digest) only if I've read
> another article in that group before opening up the digest?

Uhm...  I don't know.  :-)

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen