[Article display mess up]

[Article display mess up]

[Danny Siu]

The headers line in Article buffer mess up when I enter a nnfolder mail group
(GNUS) and view the first unread article the _first time_.  Subsequence views
of the same articleis are displayed OK.  When the Article buffer is mess up,
the text "nnfolder:GNUS" is inserted in the beginning and the end of all the
headers lines (To:, Subject:, From:, etc).

I am using rgnus-0.81 with picon and x-face (all display in article buffer)
enabled.

Any Help?

Thanks.
-- 
                                                                ^..^
 '/..\"                                                        ( oo )  )~
m( oo )m                                                         ,,  ,,
/=-=-=-=-=-=-=-=-=-=-=-==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=\
|  Danny Dick-Fung Siu                http://www-leland.stanford.edu/~dsiu |
|  Reference System Group             http://www.csua.berkeley.edu/~dsiu   |
|  Adobe Systems Incorporated       E-Mail:<dsiu@alumni.eecs.berkeley.edu> |
|       "What do you expect from a         <dsiu@leland.stanford.edu>      |
|          Cal. Berkeley grad.?"           <dsiu@adobe.com>                |
\=-=-=-=-=-=-=-=-=-=-=-==-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=/

Re: [Article display mess up]

[Jan Vroonhof]

Danny Siu <dsiu@Adobe.COM> writes:

> The headers line in Article buffer mess up when I enter a nnfolder mail group
> (GNUS) and view the first unread article the _first time_.  Subsequence views
> of the same articleis are displayed OK.  When the Article buffer is mess up,
> the text "nnfolder:GNUS" is inserted in the beginning and the end of all the
> headers lines (To:, Subject:, From:, etc).
> 
> I am using rgnus-0.81 with picon and x-face (all display in article buffer)
> enabled.

I just want confirm that I seeing this too (and I have been seeing
this for a few versions now). I haven't had the time to investigate
further to send a proper bug report yet. (For me it is a mangeled
version of the newsgroup name that is inserted, sometimes multiple times).

Jan

Re: [Article display mess up]

[Lars Magne Ingebrigtsen]

Danny Siu <dsiu@Adobe.COM> writes:

> The headers line in Article buffer mess up when I enter a nnfolder mail group
> (GNUS) and view the first unread article the _first time_.  Subsequence views
> of the same articleis are displayed OK.  When the Article buffer is mess up,
> the text "nnfolder:GNUS" is inserted in the beginning and the end of all the
> headers lines (To:, Subject:, From:, etc).
> 
> I am using rgnus-0.81 with picon and x-face (all display in article buffer)
> enabled.

Guessing wildly, I think this sounds like a glyph glitch.  Is the text
really there, or does it just look like it's there?  

Try switching x-face and picons off and see whether that helps.

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@ifi.uio.no * Lars Ingebrigtsen

Re: [Article display mess up]

[Wesley.Hardaker]

>>>>> "Lars" == Lars Magne Ingebrigtsen <larsi@ifi.uio.no> writes:

Lars> Guessing wildly, I think this sounds like a glyph glitch.  Is the text
Lars> really there, or does it just look like it's there?  

I'm guessing something similiar...  A lot of people (ok, 5 or 6) have
reported similar problems too me but I've never been able to reproduce
it or see it happen so I'm a bit confused really...  I guessed it was
probably a bug with XEmacs not picons or gnus, but I can't be sure...

Lars> Try switching x-face and picons off and see whether that helps.

Just swicthing off picons should do it...  Its the group name thats
getting inserted by the picons stuff...

Wes

Re: [Article display mess up]

[Jan Vroonhof]

[-- Attachment #1: Type: text/plain, Size: 592 bytes --]

Wesley.Hardaker@sphys.unil.ch writes:

> >>>>> "Lars" == Lars Magne Ingebrigtsen <larsi@ifi.uio.no> writes:
> 
> Lars> Guessing wildly, I think this sounds like a glyph glitch.  Is the text
> Lars> really there, or does it just look like it's there?  
> 
> I'm guessing something similiar...  A lot of people (ok, 5 or 6) have
> reported similar problems too me but I've never been able to reproduce
> it or see it happen so I'm a bit confused really...  I guessed it was
> probably a bug with XEmacs not picons or gnus, but I can't be sure...

Just so you guys can see what it looks like:



[-- Attachment #2: mangeld.gif --]
[-- Type: image/gif, Size: 25358 bytes --]

[-- Attachment #3: Type: text/plain, Size: 275 bytes --]



> Lars> Try switching x-face and picons off and see whether that helps.
> 
> Just swicthing off picons should do it...  Its the group name thats
> getting inserted by the picons stuff...

I will try that. (Is there any way to view the extents explitly for
debugging?)

Jan

Re: [Article display mess up] SOLVED! (-> FAQ ???)

[Jan Vroonhof]

 writes:

[From a discussion about a problem where the group name gets insert by
the picons all over the headers on the very firt message you read]

> Lars> Guessing wildly, I think this sounds like a glyph glitch.  Is the text
> Lars> really there, or does it just look like it's there?  
> 
> I'm guessing something similiar...  A lot of people (ok, 5 or 6) have
> reported similar problems too me but I've never been able to reproduce
> it or see it happen so I'm a bit confused really...  I guessed it was
> probably a bug with XEmacs not picons or gnus, but I can't be sure...

I solved it.. It is a RTFM type operator error. It is produced by
something like the following in .gnus

(setq 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)

You get this if you copy the example out of the Info file and then
change the first line to say 'article without reading the
documentations for the other functions. I did exactly that and I think
there are some others with the above lines in their .gnus's

It SHOULD read of course

(setq gnus-picons-display-where 'article)
(add-hook 'gnus-article-display-hook 'gnus-article-display-picons t)
(add-hook 'gnus-article-prepare-hook 'gnus-group-display-picons t)

which works fine.

The junk is generated when upon entering a group
gnus-summary-prepare-hook gets run. With the above setiings it trys to
insert an annotation in the article buffer. If this is the first
message you read, there is none. Therefore the following code in
gnus-group-display-picons

    (save-excursion
      (set-buffer (get-buffer-create
		   (gnus-get-buffer-name gnus-picons-display-where)))
      (gnus-add-current-to-buffer-list)


creates the article buffer and inserts an annoation with the group
name. This annotation has a "begin-glyph" property set to the group
icon/name. This wouldn't be a problem, BUT the gnus
article-fancification code does not expect this and creates
annotations/extents itself each one inheriting the begin-glyph
property. This gives a group "icon" for every extent which creates the
mess.

So I suggest that

1.  the example in the Info files is extended by comments explaining
    that this choice of hooks corresponds to only this choice for
    g-p-d-where
2. The picon hook functions refrain from creating one of the Gnus
   special buffers (i.e. *Article*) itself but instead give an error.
and/or
3. The picon functions are made more intelligent with a special
   gnus-picon-summary-hook-function and similar
   g-p-article-hook-function which will do the right thing
   automagically based on the value of g-p-d-where. 

Jan

P.S. This extent stuff is an horror to debug.

Re: [Article display mess up] SOLVED! (-> FAQ ???)

[Lars Magne Ingebrigtsen]

Jan Vroonhof <vroonhof@math.ethz.ch> writes:

> This annotation has a "begin-glyph" property set to the group
> icon/name. This wouldn't be a problem, BUT the gnus
> article-fancification code does not expect this and creates
> annotations/extents itself each one inheriting the begin-glyph
> property.

Is there a way to avoid inheriting this begin-glyph?  Could it be
marked as "non-inheritable" or something?

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@ifi.uio.no * Lars Ingebrigtsen

Re: [Article display mess up] SOLVED! (-> FAQ ???)

[Jan Vroonhof]

Lars Magne Ingebrigtsen <larsi@ifi.uio.no> writes:

> > This annotation has a "begin-glyph" property set to the group
> > icon/name. This wouldn't be a problem, BUT the gnus
> > article-fancification code does not expect this and creates
> > annotations/extents itself each one inheriting the begin-glyph
> > property.
> 
> Is there a way to avoid inheriting this begin-glyph?  Could it be
> marked as "non-inheritable" or something?

The best way would be make sure the picon annoation extent never
overlaps the headers as it should be. You could do that by making sure
one inserts the article headers BEFORE the extent.
The easiest way to do this by creating the picon annotations after the
article texts had bin inserted which is why one should put the picon
stuff in article-display-hook.

JAn

Re: [Article display mess up] SOLVED! (-> FAQ ???)

[Lars Magne Ingebrigtsen]

Jan Vroonhof <vroonhof@math.ethz.ch> writes:

> The best way would be make sure the picon annoation extent never
> overlaps the headers as it should be. You could do that by making sure
> one inserts the article headers BEFORE the extent.

Could we make the picons `start-open'?  Shouldn't that do the trick? 

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@ifi.uio.no * Lars Ingebrigtsen