comments on pgnus-v0.80

comments on pgnus-v0.80

[SL Baur]

Some miscellaneous comments after trying out the latest edition of
Pterodactyl Gnus.

* Group entry appears to be substantially faster than with Gnus 5.6+tm.

* Escape Quoted is not a wise default for a coding system and can lead
  to lossage[1].  It is probably better to let the editor decide which
  coding system to use.

* The MIME display defaults are unintuitive.  Could you please make
  them more Netscapish?  In particular, every MIME type that can be
  displayed inline, should be displayed inline by default, at least in 
  XEmacs.  It's what most users have come to expect.

* Notwithstanding the previous item and the coolness of Lars'
  presentation, I'm not sure I like handing html parts to W3 as a
  default.  It was pretty cool that tm blanked it out -- html
  doesn't belong in either mail or news.  How fast does this all
  work on slow(er) machines?  W3 flies on Demeter, but that's not
  exactly a fair test yet.

* Displaying html, even with a properly specified charset of
  ISO-2022-JP, loses contained Japanese text.  See
        http://www.xemacs.org/~slb/305
  for a message that exhibits this lossage.
  I put a screen snippet of what I see at:
        http://www.xemacs.org/~slb/mojibake.jpg

* There doesn't appear to be any obvious way to send message/rfc822
  subparts.  The inability of tm to rationally deal with this MIME
  type is its biggest failing IMO.

* There needs to be a more descriptive error message about the
  incompatible change to the `nnmail-spool-file' variable.  A
  pointer to the manual (which contains good detail on this) would
  be fine, but I got a `wrong-type argument listp, procmail' beep
  the first time through.
  
* The broken `Attach file as MIME' function in the Message menu should 
  have a default keybinding.  How about C-c C-x TAB?

* It would be nice to have a preview function as in tm to see what a
  message looks like before sending it out.  If one already exists, it 
  should be added to the Message menu.

Footnotes: 
[1]  In particular, with XEmacs/Mule running under a Japanese locale,
included Japanese text isn't handled properly.  I've submitted a bug
report for this.

Re: comments on pgnus-v0.80

[Alexandre Oliva]

On Apr  7, 1999, SL Baur <steve@xemacs.org> wrote:

> * The MIME display defaults are unintuitive.  Could you please make
>   them more Netscapish?

I don't understand what you mean here...  But then, maybe Lars does :-)

>   presentation, I'm not sure I like handing html parts to W3 as a
>   default.

This is a requirement of the RFCs that define MIME: they say the last
alternate part should be presented by default, since it is supposed to
be the most appealing one, whereas the first ones are supposed to be
closer to plain-text, so that MIME-uncapable readers will get to the
real text before html garbage.  But there could be some way to disable 
html by default, for example (maybe there is already)

> * There doesn't appear to be any obvious way to send message/rfc822
>   subparts.

C-c C-f (optionally followed by cut&paste into another message :-)

> * The broken `Attach file as MIME' function in the Message menu should 
>   have a default keybinding.  How about C-c C-x TAB?

Missing SEMI, huh? :-)

How about C-c C-a?

> * It would be nice to have a preview function as in tm to see what a
>   message looks like before sending it out.

C-c C-d, then enter the drafts group.

-- 
Alexandre Oliva http://www.dcc.unicamp.br/~oliva IC-Unicamp, Brasil
{oliva,Alexandre.Oliva}@dcc.unicamp.br  aoliva@{acm.org,computer.org}
oliva@{gnu.org,kaffe.org,{egcs,sourceware}.cygnus.com,samba.org}
*** E-mail about software projects will be forwarded to mailing lists

Re: comments on pgnus-v0.80

[Lee Willis]

[-- Attachment #1: Type: text/plain, Size: 1444 bytes --]

SL Baur <steve@xemacs.org> writes:

> * The MIME display defaults are unintuitive.  Could you please make
>   them more Netscapish?  In particular, every MIME type that can be
>   displayed inline, should be displayed inline by default, at least in 
>   XEmacs.  It's what most users have come to expect.

Maybe, but are you bearing in mind that pgnus currently defaults to
sending attachments with a disposition of "attachment" as do a large
number of clients so these should *not* be displayed inline even if the
reader is capable of doing so. The only things that should be shown
inline are those that have a disposition of inline.[1]

> * Notwithstanding the previous item and the coolness of Lars'
>   presentation, I'm not sure I like handing html parts to W3 as a
>   default.  It was pretty cool that tm blanked it out -- html
>   doesn't belong in either mail or news.  How fast does this all
>   work on slow(er) machines?  W3 flies on Demeter, but that's not
>   exactly a fair test yet.

Me neither, it doesn't really fly on mine so I like to default to
text/plain when it's there. I do feel that it is *correct* to default in
normal circumstances to the most advanced version (ie quite often
text/html). With this is mind there hould probably be a mention in the
manual about the following code, if there wasn't already (Posted on this
list a while ago by David Goldberg who it seems has contributed a large
portion of my .gnus!)


[-- Attachment #2: .gnus excerpt --]
[-- Type: application/emacs-lisp, Size: 205 bytes --]

[-- Attachment #3: Type: text/plain, Size: 715 bytes --]


> * There needs to be a more descriptive error message about the
>   incompatible change to the `nnmail-spool-file' variable.  A
>   pointer to the manual (which contains good detail on this) would
>   be fine, but I got a `wrong-type argument listp, procmail' beep
>   the first time through.

I think we need a pgnus MIME FAQ  with all the commonly asked questions
like :

  * How do I make pgnus use MIME
  * What needs to change in my .gnus
  * How do I show text/plain instead of text/html (See above!)
  * How do I get w3 not to munge my colours

etc. etc. etc.

2p worth ;)
Lee.
-- 
I was doing object-oriented assembly at 1 year old ...  
For some reason my mom insists on calling it "Playing with blocks"

Re: comments on pgnus-v0.80

[Alexandre Oliva]

On Apr  7, 1999, Lee Willis <lee@gbdirect.co.uk> wrote:

> With this is mind there hould probably be a mention in the manual
> about the following code, if there wasn't already (Posted on this
> list a while ago by David Goldberg who it seems has contributed a
> large portion of my .gnus!)

I'm not sure this is the right thing to do.  I, for one, would prefer
richtext over plain text, so it would probably be better to move
text/html to the lower priority, rather than bringing plain text to
the highest one.

-- 
Alexandre Oliva http://www.dcc.unicamp.br/~oliva IC-Unicamp, Brasil
{oliva,Alexandre.Oliva}@dcc.unicamp.br  aoliva@{acm.org,computer.org}
oliva@{gnu.org,kaffe.org,{egcs,sourceware}.cygnus.com,samba.org}
*** E-mail about software projects will be forwarded to mailing lists

Re: comments on pgnus-v0.80

[Alexandre Oliva]

On Apr  7, 1999, Lee Willis <lee@gbdirect.co.uk> wrote:

> Well maybe not, but it's an example of how to always prefer text/plain
> which is what I wanted, and I imagine moving text/html lower would be
> quite difficult.

Not much: :-)

(eval-after-load
 "mm-decode"
 '(setq mm-alternative-precedence
	(nconc (delete "text/html" mm-alternative-precedence)
	       '("text/html"))))

-- 
Alexandre Oliva http://www.dcc.unicamp.br/~oliva IC-Unicamp, Brasil
{oliva,Alexandre.Oliva}@dcc.unicamp.br  aoliva@{acm.org,computer.org}
oliva@{gnu.org,kaffe.org,{egcs,sourceware}.cygnus.com,samba.org}
*** E-mail about software projects will be forwarded to mailing lists

Re: comments on pgnus-v0.80

[Kai.Grossjohann]

SL Baur <steve@xemacs.org> writes:

  > * The MIME display defaults are unintuitive.  Could you please make
  >   them more Netscapish?  In particular, every MIME type that can be
  >   displayed inline, should be displayed inline by default, at least in 
  >   XEmacs.  It's what most users have come to expect.

Lars figures that the author knew what he was saying when he was
saying `Content-disposition: attachment'.  Are your non-inline parts
like this, or is something else happening?

kai
-- 
Abort this operation?   [Abort]  [Cancel]

Re: comments on pgnus-v0.80

[Kai.Grossjohann]

SL Baur <steve@xemacs.org> writes:

  > * There doesn't appear to be any obvious way to send message/rfc822
  >   subparts.  The inability of tm to rationally deal with this MIME
  >   type is its biggest failing IMO.

I think Lars wants forwarded messages to be MIME-encapsulated in the
future, it has just not happened yet.

Or are you trying to say that one needs to say C-d then C-c C-f on the
right part to forward part of a multipart MIME message?  Maybe this
should be available without going through nndoc first...

kai
-- 
Abort this operation?   [Abort]  [Cancel]

Re: comments on pgnus-v0.80

[Stefanie Teufel]

On Wed, Apr 07, 1999 at 12:23:20 +0100, Lee Willis wrote:

> I think we need a pgnus MIME FAQ  with all the commonly asked questions
> like :

Please:-)

Stefanie
-- 
"Breaking Windows isn't just for kids anymore..."

Re: comments on pgnus-v0.80

[Alan Shutko]

SL Baur <steve@xemacs.org> writes:

>   How fast does this all work on slow(er) machines?  W3 flies on
>   Demeter, but that's not exactly a fair test yet.

After the first time (autoloading everything), I don't really notice
rendering time on my Pentium 60.  

-- 
Alan Shutko <ats@acm.org> - By consent of the corrupted
Sure he's sharp as a razor ... he's a two-dimensional pinhead!

Re: comments on pgnus-v0.80

[Justin Sheehy]

SL Baur <steve@xemacs.org> writes:
 
>   In particular, every MIME type that can be displayed inline,
>   should be displayed inline by default, at least in XEmacs.
 
Not if it was sent with a disposition of 'attachment'.  There are some
pieces of software that are broken in that they ignore the disposition
of MIME parts.  Gnus should not copy their mistakes just because they
are expected.
 
> * There doesn't appear to be any obvious way to send message/rfc822
>   subparts.
 
Forwarded messages are sent this way by default.

> * The broken `Attach file as MIME' function in the Message menu should
>   have a default keybinding.  How about C-c C-x TAB?

How about `C-c C-a'?  This is already present.

> * It would be nice to have a preview function as in tm to see what a
>   message looks like before sending it out.  If one already exists, it
>   should be added to the Message menu.

A real-time preview might be nice, but the drafts group works as a
second-best solution to this.

-- 
Justin Sheehy

In a cloud bones of steel.
  

Re: comments on pgnus-v0.80

[Justin Sheehy]

Kai.Grossjohann@CS.Uni-Dortmund.DE writes:

> I think Lars wants forwarded messages to be MIME-encapsulated in the
> future, it has just not happened yet.

Yes, it has.  When was the last time that you forwarded a message?

-- 
Justin Sheehy

In a cloud bones of steel.
  

Re: comments on pgnus-v0.80

[Per Abrahamsen]

SL Baur <steve@xemacs.org> writes:

> * Notwithstanding the previous item and the coolness of Lars'
>   presentation, I'm not sure I like handing html parts to W3 as a
>   default.  It was pretty cool that tm blanked it out -- html
>   doesn't belong in either mail or news.  How fast does this all
>   work on slow(er) machines?  W3 flies on Demeter, but that's not
>   exactly a fair test yet.

I don't even notice that some articles are in HTML rather than plain
text.  This is on a 166 Mhz UltraSPARC.  I don't think you can buy
PC's slower than that today.

Re: comments on pgnus-v0.80

[Kai.Grossjohann]

Justin Sheehy <justin@linus.mitre.org> writes:

  > Yes, it has.  When was the last time that you forwarded a message?

I think I'm just confused.  Had to get up rather early today...

Sorry.

kai
-- 
Abort this operation?   [Abort]  [Cancel]

Re: comments on pgnus-v0.80

[Jan Vroonhof]

SL Baur <steve@xemacs.org> writes:

> * Notwithstanding the previous item and the coolness of Lars'
>   presentation, I'm not sure I like handing html parts to W3 as a
>   default.  It was pretty cool that tm blanked it out -- html

(eval-after-load "mm-decode"
  '(progn
	;; Do not prefer html
    (nconc (delete "text/html" mm-alternative-precedence) (list
    "text/html"))
	;; Always inline rfc-822
    (nconc mm-attachment-override-types '("message/rfc-822"))))

To attach files use

C-c C-a

to see how to attach messages, try forwarding a messages.

Jan

Re: comments on pgnus-v0.80

[Simon Josefsson]

SL Baur <steve@xemacs.org> writes:

> * It would be nice to have a preview function as in tm to see what a
>   message looks like before sending it out.  If one already exists, it 
>   should be added to the Message menu.

This is what I use.

Several people has suggested using the draft folder. Has anyone
actually tried that?  I just see MML there, which is not really what I
would call a "preview".

(defun mml-preview (&optional raw)
  "Display current buffer with Gnus, in a new buffer. If raw, don't
highlight the article."
  (interactive "P")
  (let ((buf (current-buffer)))
    (switch-to-buffer (get-buffer-create 
		       (concat (if raw "*Raw MIME preview of "
				 "*MIME preview of ") (buffer-name))))
    (erase-buffer)
    (insert-buffer buf)
    (mml-to-mime)
    (unless raw
      (run-hooks 'gnus-article-decode-hook)
      (let ((gnus-newsgroup-name "dummy"))
	(gnus-article-prepare-display)))
    (fundamental-mode)
    (setq buffer-read-only t)
    (goto-char (point-min))))

Re: comments on pgnus-v0.80

[Shenghuo ZHU]

>>>>> "sb" == SL Baur <steve@xemacs.org> writes:

sb> * Escape Quoted is not a wise default for a coding system and can lead
sb>   to lossage[1].  It is probably better to let the editor decide which
sb>   coding system to use.

sb> Footnotes: 
sb> [1]  In particular, with XEmacs/Mule running under a Japanese locale,
sb> included Japanese text isn't handled properly.  I've submitted a bug
sb> report for this.

In GNU Emacs, gnus uses 'emacs-mule' as the coding system for draft.
The question is which coding system is used as auto-save coding-system
and can preserve the characters correctly in XEmacs/Mule.

Variable message-draft-coding-system is used to customize the coding
system.

-- 
Shenghuo ZHU

using agent for a full preview

[Simon Michael]

Simon Josefsson <jas@pdc.kth.se> writes:
> Several people has suggested using the draft folder. Has anyone
> actually tried that?  I just see MML there, which is not really what I
> would call a "preview".

You can get a better preview this way: (gnus-agentize), go unplugged,
send your message, and view it in the "queue" group. 

You can still edit the message (D e ? I think) in which case it goes
back to the "drafts" group - nice.

One problem however: if a copy was saved in your "sent" group, it
won't reflect the edits. Perhaps it should not get archived there
until it's *really* sent, ie in agent mode, when it has left the
queue.

.

Re: comments on pgnus-v0.80

[Hans de Graaff]

SL Baur <steve@xemacs.org> writes:

> * Notwithstanding the previous item and the coolness of Lars'
>   presentation, I'm not sure I like handing html parts to W3 as a
>   default.  It was pretty cool that tm blanked it out -- html
>   doesn't belong in either mail or news.  How fast does this all
>   work on slow(er) machines?  W3 flies on Demeter, but that's not
>   exactly a fair test yet.

Apart from autoloading, rendering HTML using w3 is bearable on my
486/66Mhz with only 32Mb [1]. Then again, anything is bearable on this
machine. Because the whole of things is slow, I also accept the
somewhat slow rendering. It hasn't been slow enough for me to dig in
the manual to do something about it.

Hans

1. Yeah, I know, and I'm going to buy a new machine over the next
   week. :-)

Re: comments on pgnus-v0.80

[SL Baur]

Kai Grossjohann <Kai.Grossjohann@CS.Uni-Dortmund.DE> writes in ding@gnus.org:

> SL Baur <steve@xemacs.org> writes:
>> * The MIME display defaults are unintuitive.  Could you please make
>> them more Netscapish?  In particular, every MIME type that can be
>> displayed inline, should be displayed inline by default, at least in 
>> XEmacs.  It's what most users have come to expect.

> Lars figures that the author knew what he was saying when he was
> saying `Content-disposition: attachment'.  Are your non-inline parts
> like this, or is something else happening?

They were labeled `attachment's.  tm seems to display a lot more things 
inline by default than Gnus MIME does.  

Re: comments on pgnus-v0.80

[SL Baur]

Kai Grossjohann <Kai.Grossjohann@CS.Uni-Dortmund.DE> writes in ding@gnus.org:

 ...
> Or are you trying to say that one needs to say C-d then C-c C-f on the
> right part to forward part of a multipart MIME message?  Maybe this
> should be available without going through nndoc first...

I don't think so.  I'm saying that attaching mail messages shouldn't
be so difficult.  The way I used to this was to type C-c C-x C-t
message/rfc822 and then include the message with C-x i.

As near as I can tell, I totally lost when trying to submit a bug
report in similar fashion a couple days ago.

Re: comments on pgnus-v0.80

[SL Baur]

Per Abrahamsen <abraham@dina.kvl.dk> writes in ding@gnus.org:

> SL Baur <steve@xemacs.org> writes:
>> * Notwithstanding the previous item and the coolness of Lars'
>> presentation, I'm not sure I like handing html parts to W3 as a
>> default.  It was pretty cool that tm blanked it out -- html
>> doesn't belong in either mail or news.  How fast does this all
>> work on slow(er) machines?  W3 flies on Demeter, but that's not
>> exactly a fair test yet.

> I don't even notice that some articles are in HTML rather than plain
> text.  This is on a 166 Mhz UltraSPARC.  I don't think you can buy
> PC's slower than that today.

O.K.  Fair enough.  Objection dropped.

Re: comments on pgnus-v0.80

[SL Baur]

This is nice, but how would I explain it to my mom[1]?  I really want
my mom to use Gnus.

Jan Vroonhof <vroonhof@math.ethz.ch> writes in ding@gnus.org:

> SL Baur <steve@xemacs.org> writes:
>> * Notwithstanding the previous item and the coolness of Lars'
>> presentation, I'm not sure I like handing html parts to W3 as a
>> default.  It was pretty cool that tm blanked it out -- html

> (eval-after-load "mm-decode"
>   '(progn
> 	;; Do not prefer html
>     (nconc (delete "text/html" mm-alternative-precedence) (list
>     "text/html"))
> 	;; Always inline rfc-822
>     (nconc mm-attachment-override-types '("message/rfc-822"))))

> To attach files use

> C-c C-a

> to see how to attach messages, try forwarding a messages.

> Jan


Footnotes: 
[1]  She uses Linux/Windowmaker & Netscape, but not Gnus, just fine to
access her email.

Re: using agent for a full preview

[SL Baur]

Simon Michael <simon@joyful.com> writes in ding@gnus.org:

> Simon Josefsson <jas@pdc.kth.se> writes:
>> Several people has suggested using the draft folder. Has anyone
>> actually tried that?  I just see MML there, which is not really what I
>> would call a "preview".

> You can get a better preview this way: (gnus-agentize), go unplugged,
> send your message, and view it in the "queue" group. 

Come on guys.

Previewing should be a whole lot simpler than this.  Thanks for the
suggestions though.

Lars, could we please have a preview function just like we used to
have with tm?

Re: comments on pgnus-v0.80

[Kai.Grossjohann]

SL Baur <steve@xemacs.org> writes:

  > I don't think so.  I'm saying that attaching mail messages shouldn't
  > be so difficult.  The way I used to this was to type C-c C-x C-t
  > message/rfc822 and then include the message with C-x i.

Righto.  Lars wants more commands which insert <part> tags.

Btw, what happens if you save the msg to a file then type C-c C-a on
the file and choose the right content type?

kai
-- 
Abort this operation?   [Abort]  [Cancel]

Re: comments on pgnus-v0.80

[Kai.Grossjohann]

SL Baur <steve@xemacs.org> writes:

  > This is nice, but how would I explain it to my mom[1]?  I really want
  > my mom to use Gnus.

Well, you could do the Lisp code part :-)

Some time ago, pgnus used to fully buttonize messages, then one could
see the alternate parts and choose the right format.  Try `K b' on
such a message -- maybe there is a way to tell Gnus to always `K b'
all messages.  (Erm.  M-t?)

kai
-- 
Abort this operation?   [Abort]  [Cancel]

Re: comments on pgnus-v0.80

[Lee Willis]

SL Baur <steve@xemacs.org> writes:

> Kai Grossjohann <Kai.Grossjohann@CS.Uni-Dortmund.DE> writes in ding@gnus.org:
> 
> I don't think so.  I'm saying that attaching mail messages shouldn't
> be so difficult.  The way I used to this was to type C-c C-x C-t
> message/rfc822 and then include the message with C-x i.

Which to my mind is more complicated than
C-c C-a filename message/rfc822 RET
Well maybe not more complicated, but certainly no easier.

> As near as I can tell, I totally lost when trying to submit a bug
> report in similar fashion a couple days ago.

Works fine for me in p0.80

Lee.
-- 
I was doing object-oriented assembly at 1 year old ...  
For some reason my mom insists on calling it "Playing with blocks"

Re: comments on pgnus-v0.80

[Lee Willis]

SL Baur <steve@xemacs.org> writes:

> Kai Grossjohann <Kai.Grossjohann@CS.Uni-Dortmund.DE> writes in ding@gnus.org:
> 
> > Lars figures that the author knew what he was saying when he was
> > saying `Content-disposition: attachment'.  Are your non-inline parts
> > like this, or is something else happening?
> 
> They were labeled `attachment's.  tm seems to display a lot more things 
> inline by default than Gnus MIME does.  

In which case tm if tm displays parts marked as attachments inline then
tm is broken, an attachment is just that, an attachment, it should not
be displayed inline ...

Lee.
-- 
I was doing object-oriented assembly at 1 year old ...  
For some reason my mom insists on calling it "Playing with blocks"

Re: comments on pgnus-v0.80

[SL Baur]

Kai Grossjohann <Kai.Grossjohann@CS.Uni-Dortmund.DE> writes in ding@gnus.org:

> SL Baur <steve@xemacs.org> writes:
>> This is nice, but how would I explain it to my mom[1]?  I really want
>> my mom to use Gnus.

> Well, you could do the Lisp code part :-)

> Some time ago, pgnus used to fully buttonize messages, then one could
> see the alternate parts and choose the right format.  Try `K b' on
> such a message -- ...

That's nice, but there is no way I could teach her to do that.  She
can point and click and do little else.  Nor should she[1] have to do
anything else.

Footnotes: 
[1]  or anyone else

Re: comments on pgnus-v0.80

[Kai.Grossjohann]

SL Baur <steve@xemacs.org> writes:

  > That's nice, but there is no way I could teach her to do that.
  > She can point and click and do little else.  Nor should she[1]
  > have to do1 anything else.

Some people prefer to see all alternatives by default, some people
prefer to see only one alternative by default.  There is no way pgnus
can please both groups.  Thus, at least one group has to do something
to get what they want.

Some people prefer to see html by default, others prefer text.  See
above.

I was only trying to show you that the buttonized look exists, and if
you want your Mom to have it, you can hack her ~/.emacs so that it
happens automagically.

kai
-- 
Abort this operation?   [Abort]  [Cancel]

Re: comments on pgnus-v0.80

[Jan Vroonhof]

SL Baur <steve@xemacs.org> writes:

> This is nice, but how would I explain it to my mom[1]?  I really want
> my mom to use Gnus.

I think that you should tell your mother not use alpha-grade software
:-). It simply isn't finished yet.

Jan

Re: comments on pgnus-v0.80

[Karl Kleinpaste]

Kai.Grossjohann@CS.Uni-Dortmund.DE writes:
> maybe there is a way to tell Gnus to always `K b' all messages.

Yup.

        `gnus-unbuttonized-mime-types' is a variable declared in Lisp.
          -- loaded from "gnus-art"
        ...
        Documentation:
        List of MIME types that should not be given buttons when rendered.

Default value is '(".*/.*").

(setq gnus-unbuttonized-mime-types nil)

Re: comments on pgnus-v0.80

[SL Baur]

Jan Vroonhof <vroonhof@math.ethz.ch> writes in ding@gnus.org:

> SL Baur <steve@xemacs.org> writes:
>> This is nice, but how would I explain it to my mom[1]?  I really want
>> my mom to use Gnus.

> I think that you should tell your mother not use alpha-grade software
> :-). It simply isn't finished yet.

:-)  I know that.

If I didn't care, I wouldn't be commenting now.  I have been using
Gnus+tm for almost four years now.  It works (quite well) for me and
many others.  I wish to see native Gnus MIME work as well.

Re: comments on pgnus-v0.80

[Per Abrahamsen]

SL Baur <steve@xemacs.org> writes:

> [1]  She uses Linux/Windowmaker & Netscape, but not Gnus, just fine to
> access her email.

Netscape includes a nice mail agent, why not suggest her to use that?

Re: comments on pgnus-v0.80

[Rupa Schomaker (list)]

SL Baur <steve@xemacs.org> writes:

> Per Abrahamsen <abraham@dina.kvl.dk> writes in ding@gnus.org:
> 
> > SL Baur <steve@xemacs.org> writes:
> >> * Notwithstanding the previous item and the coolness of Lars'
> >> presentation, I'm not sure I like handing html parts to W3 as a
> >> default.  It was pretty cool that tm blanked it out -- html
> >> doesn't belong in either mail or news.  How fast does this all
> >> work on slow(er) machines?  W3 flies on Demeter, but that's not
> >> exactly a fair test yet.
> 
> > I don't even notice that some articles are in HTML rather than plain
> > text.  This is on a 166 Mhz UltraSPARC.  I don't think you can buy
> > PC's slower than that today.
> 
> O.K.  Fair enough.  Objection dropped.


Where W3 really breaks on my machine is on the only piece of HTML
email that I actually read.  It is about 60K and is a multi-level
table.  Rendering it takes about 30s - 1min on my PPro 200.  Very
slow.  Regular "HTML Email" that most people generate renders
immediately.  I think that tables just take forever to render...


-- 
Rupa (rupa@rupa.com for normal email)
Please don't email duplicate replies.

Re: using agent for a full preview

[Colin Rafferty]

SL Baur writes:

> Lars, could we please have a preview function just like we used to
> have with tm?

Simon Josefsson's previewer (message <iluwvzov1go.fsf@squid.pdc.kth.se>)
does exactly what we need.  It even gets the X-Face, if you have it in.

Lars just needs to include that (or something like that) in Gnus.

-- 
Colin

Re: comments on pgnus-v0.80

[William M. Perry]

"Rupa Schomaker (list)" <rupa-list@rupa.com> writes:

> SL Baur <steve@xemacs.org> writes:
> 
> > Per Abrahamsen <abraham@dina.kvl.dk> writes in ding@gnus.org:
> > 
> > > SL Baur <steve@xemacs.org> writes:
> > >> * Notwithstanding the previous item and the coolness of Lars'
> > >> presentation, I'm not sure I like handing html parts to W3 as a
> > >> default.  It was pretty cool that tm blanked it out -- html
> > >> doesn't belong in either mail or news.  How fast does this all
> > >> work on slow(er) machines?  W3 flies on Demeter, but that's not
> > >> exactly a fair test yet.
> > 
> > > I don't even notice that some articles are in HTML rather than plain
> > > text.  This is on a 166 Mhz UltraSPARC.  I don't think you can buy
> > > PC's slower than that today.
> > 
> > O.K.  Fair enough.  Objection dropped.
> 
> 
> Where W3 really breaks on my machine is on the only piece of HTML email
> that I actually read.  It is about 60K and is a multi-level table.
> Rendering it takes about 30s - 1min on my PPro 200.  Very slow.  Regular
> "HTML Email" that most people generate renders immediately.  I think that
> tables just take forever to render...

Tables are very slow.... if anyone wants to volunteer to rewrite the
rendering of them, feel free. :)

-bp

Re: comments on pgnus-v0.80

[Alan Shutko]

Kai.Grossjohann@CS.Uni-Dortmund.DE writes:

> Lars figures that the author knew what he was saying when he was
> saying `Content-disposition: attachment'.

Unfortunately, that's a baseless assumption, given the number of
mailers out there that _only_ do attachments, and couldn't send a file
inline to save their user's life.

-- 
Alan Shutko <ats@acm.org> - By consent of the corrupted
Six bottles of beer. And quickly please, the world's about to end...

Re: comments on pgnus-v0.80

[Steinar Bang]

>>>>> SL Baur <steve@xemacs.org>:

> If I didn't care, I wouldn't be commenting now.  I have been using
> Gnus+tm for almost four years now.  It works (quite well) for me and
> many others.  I wish to see native Gnus MIME work as well.

I've been a tm user since 1995 (first with MH until 1997 and with Gnus
from then on) and I prefer the pgnus way.  At least for reading.

It was a real bitch making tm do anything other than the defaults.

Re: comments on pgnus-v0.80

[lconrad]

>>>>> "sb" == SL Baur <steve@xemacs.org> writes:

    sb> Jan Vroonhof <vroonhof@math.ethz.ch> writes in ding@gnus.org:
    >> SL Baur <steve@xemacs.org> writes:
    >>> This is nice, but how would I explain it to my mom[1]?  I
    >>> really want my mom to use Gnus.

    >> I think that you should tell your mother not use alpha-grade
    >> software :-). It simply isn't finished yet.

    sb> :-) I know that.


I think this is an important point -- yes, pgnus is alpha software
which isn't yet usable by lots of people who use netscape, but it it a 
design goal that when it is "ready" it will be? 

I would think that one thing that would have to happen before this 
would be the case would be that the documentation would have to be
written differently.

For instance, I make my living being able to figure out software from
reading the documentation, and I still can't get the gnus agent to
work.  (This is one of the features gnus allegedly has that
netscape users might want.)  A couple of sentences that I think need
some amplification:

   * First, set up Gnus as you would do if you were running it on a
     machine that has full connection to the net.  

What does this mean?  What needs to work when I'm connected to the
net, before I add the (gnus-agentize) line?

     The main way to control what is to be downloaded is to create a
     "category" and then assign some (or all) groups to this category.

It might be more useful to the new user to explain what the default
category does and why, and then explain how and why to create one that 
does something different.

-- 
Laura (mailto:lconrad@world.std.com , http://www.world.std.com/~lconrad/ )
(617) 661-8097	 
233 Broadway, Cambridge, MA 02139

Re: comments on pgnus-v0.80

[Hrvoje Niksic]

SL Baur <steve@xemacs.org> writes:

> * Escape Quoted is not a wise default for a coding system and can
>   lead to lossage[1].  It is probably better to let the editor
>   decide which coding system to use.

Can you elaborate on this?  I'm genuinely interested to hear what
"escape quoted coding system" and "let the editor decide" mean.
Specifically, why does it matter which coding system the buffer is
using, when Gnus will recode everything prior to sending it?

> * The MIME display defaults are unintuitive.  Could you please make
>   them more Netscapish?  In particular, every MIME type that can be
>   displayed inline, should be displayed inline by default, at least
>   in XEmacs.  It's what most users have come to expect.

This is what is currently being done, with two exceptions:

* Parts specifically marked as "attachments";

* Images larger than the window, which would be unscrollable in
  XEmacs.

I think it's a reasonable default (I haven't changed it.)

> * Notwithstanding the previous item and the coolness of Lars'
>   presentation, I'm not sure I like handing html parts to W3 as a
>   default.  It was pretty cool that tm blanked it out -- html
>   doesn't belong in either mail or news.  How fast does this all
>   work on slow(er) machines?  W3 flies on Demeter, but that's not
>   exactly a fair test yet.

It's very slow with default W3 settings.  Per is probably not a good
judge of the speed, as he turns off all the fonts and images stuff
(and he runs under FSF which is a good deal faster anyway).

I think presenting HTML to W3 makes sense.  I turn it off like this:

(require 'mm-decode)
(setq mm-alternative-precedence (delete "text/html"
					mm-alternative-precedence)
      mm-alternative-precedence (nconc mm-alternative-precedence
				       (list "text/html")))

I don't know if your mom would grok that.  Mine wouldn't.

> * There doesn't appear to be any obvious way to send message/rfc822
>   subparts.  The inability of tm to rationally deal with this MIME
>   type is its biggest failing IMO.

Gnus doesn't deal with it rationally either.

> * The broken `Attach file as MIME' function in the Message menu
>   should have a default keybinding.  How about C-c C-x TAB?

`C-c C-a' should work, and is compatible with what VM had for ages.
Even the prompts are VM-like.  Why is it broken, BTW?

Re: comments on pgnus-v0.80

[Lars Magne Ingebrigtsen]

SL Baur <steve@xemacs.org> writes:

> * Escape Quoted is not a wise default for a coding system and can lead
>   to lossage[1].  It is probably better to let the editor decide which
>   coding system to use.

I wasn't aware that Gnus defaulted to Escape Quoted anywhere.  Er.  I
don't even know what that is.  :-)

> * The MIME display defaults are unintuitive.  Could you please make
>   them more Netscapish?  In particular, every MIME type that can be
>   displayed inline, should be displayed inline by default, at least in 
>   XEmacs.  It's what most users have come to expect.

Everything that Gnus (or really, mm) knows how to display inline is
usually displayed inline.  Unless the Content-Disposition says
otherwise. 

> * Notwithstanding the previous item and the coolness of Lars'
>   presentation, I'm not sure I like handing html parts to W3 as a
>   default.  It was pretty cool that tm blanked it out -- html
>   doesn't belong in either mail or news.  How fast does this all
>   work on slow(er) machines?  W3 flies on Demeter, but that's not
>   exactly a fair test yet.

Loading w3 takes a long time, but after having done that, I typically
see no speed difference bewteen displaying the text/plain part and the
text/html part in a multipart/alternative.  Typically.  Of course, if
the html is complex (tables, etc), w3 may take arbitrarily long.  But
the same is true for text/plain with lots of highlighting going on. 

> * There doesn't appear to be any obvious way to send message/rfc822
>   subparts.  The inability of tm to rationally deal with this MIME
>   type is its biggest failing IMO.

`S o m' sends out a multipart/mixed with a message/rfc822 part...

> * There needs to be a more descriptive error message about the
>   incompatible change to the `nnmail-spool-file' variable.  A
>   pointer to the manual (which contains good detail on this) would
>   be fine, but I got a `wrong-type argument listp, procmail' beep
>   the first time through.

Yes.  But I think that this really should be backwards compatible.
Someone said to me (in Japan) that it would be a good idea to add a
new variable that defaults to nil, but which takes precedence over
nnmail-spool-file.  So the code would check first the new var, and if
that doesn't exist, it checks nnmail-spool-file, but it better well
grok the latter perfectly.

What values doesn't the current code understand?

> * It would be nice to have a preview function as in tm to see what a
>   message looks like before sending it out.

Yes, that would be nice.

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

Re: comments on pgnus-v0.80

[Lars Magne Ingebrigtsen]

Alexandre Oliva <oliva@dcc.unicamp.br> writes:

> This is a requirement of the RFCs that define MIME: they say the last
> alternate part should be presented by default, since it is supposed to
> be the most appealing one, whereas the first ones are supposed to be
> closer to plain-text, so that MIME-uncapable readers will get to the
> real text before html garbage.

Yup:

5.1.4.  Alternative Subtype

[...]

   Systems should recognize that the content of the various parts are
   interchangeable.  Systems should choose the "best" type based on the
   local environment and references, in some cases even through user
   interaction.  As with "multipart/mixed", the order of body parts is
   significant.  In this case, the alternatives appear in an order of
   increasing faithfulness to the original content.  In general, the
   best choice is the LAST part of a type supported by the recipient
   system's local environment.

Hm.  But Gnus doesn't really respect this.  It only looks on
`mm-alternative-precedence', and ignores the part order.  Er.  Hm.

Gnus could choose the last part it's able to display, but the user
must also be able to say that some types are discouraged.  How about
getting rid of `mm-alternative-precedence', and instead have a
`mm-alternative-discouraged-precedence', that would say in which order 
parts are unwanted.  For instance, if the user says

("text/html" "text/enriched")

then that would mean that text/html is really discouraged, but
"text/enriched" is less discouraged.  If you get my drift.

This variable would default to nil, meaning that the precedence by
default would be the order of the parts.

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

Re: comments on pgnus-v0.80

[Lars Magne Ingebrigtsen]

Alan Shutko <ats@acm.org> writes:

> > Lars figures that the author knew what he was saying when he was
> > saying `Content-disposition: attachment'.
> 
> Unfortunately, that's a baseless assumption, given the number of
> mailers out there that _only_ do attachments, and couldn't send a file
> inline to save their user's life.

Yes, it certainly looks that way.  But, at least for the time being, I 
think it's reasonable to respect the RFCs here, and let users fiddle
with mm-attachment-override-types to display things inline.

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

Re: comments on pgnus-v0.80

[Lars Magne Ingebrigtsen]

SL Baur <steve@xemacs.org> writes:

> I don't think so.  I'm saying that attaching mail messages shouldn't
> be so difficult.  The way I used to this was to type C-c C-x C-t
> message/rfc822 and then include the message with C-x i.

Yes.  There should be oodles of commands for inserting different
parts.  The template command looks like:

(defun mml-attach-file (file &optional type description)
  "Attach a file to the outgoing MIME message.
The file is not inserted or encoded until you send the message with
`\\[message-send-and-exit]' or `\\[message-send]'.

FILE is the name of the file to attach.  TYPE is its content-type, a
string of the form \"type/subtype\".  DESCRIPTION is a one-line
description of the attachment."
  (interactive
   (let* ((file (mml-minibuffer-read-file "Attach file: "))
	  (type (mml-minibuffer-read-type file))
	  (description (mml-minibuffer-read-description)))
     (list file type description)))
  (mml-insert-tag 'part 'type type 'filename file 'disposition "attachment"
		  'description description))

`mml-attach-message' and other commands should be written.  Anyone
wants to mail me a patch?

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

Re: comments on pgnus-v0.80

[Lars Magne Ingebrigtsen]

lconrad@world.std.com writes:

> For instance, I make my living being able to figure out software from
> reading the documentation, and I still can't get the gnus agent to
> work.  (This is one of the features gnus allegedly has that
> netscape users might want.)  A couple of sentences that I think need
> some amplification:
> 
>    * First, set up Gnus as you would do if you were running it on a
>      machine that has full connection to the net.  
> 
> What does this mean?  What needs to work when I'm connected to the
> net, before I add the (gnus-agentize) line?

Well, the reason for this nebulous phrasing is that what you have to
do depends so much on what you want to do.  What I'm trying to say is
that you should have Gnus working as you want, and after that, then
you can start thinking about using the Agent.

>      The main way to control what is to be downloaded is to create a
>      "category" and then assign some (or all) groups to this category.
> 
> It might be more useful to the new user to explain what the default
> category does and why, and then explain how and why to create one that 
> does something different.

That's handled in the "The Category Buffer" node, isn't it?

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

Re: comments on pgnus-v0.80

[Hrvoje Niksic]

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

> Gnus could choose the last part it's able to display, but the user
> must also be able to say that some types are discouraged.  How about
> getting rid of `mm-alternative-precedence', and instead have a
> `mm-alternative-discouraged-precedence', that would say in which
> order parts are unwanted.  For instance, if the user says
> 
> ("text/html" "text/enriched")
> 
> then that would mean that text/html is really discouraged, but
> "text/enriched" is less discouraged.  If you get my drift.

Yes.  Yes.  YES!

> This variable would default to nil, meaning that the precedence by
> default would be the order of the parts.

...and meaning that Steve's Mom could do

    (setq mm-alternative-discouraged-precedence '("text/html"))

Re: comments on pgnus-v0.80

[Per Abrahamsen]

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

>   (mml-insert-tag 'part 'type type 'filename file 'disposition "attachment"
> 		  'description description))

Wouldn't it be more lispy to use :keywords, like

    (mml-insert-tag 'part :type type :filename file :disposition "attachment"
                    :description description))

or did I totally misunderstood the syntax above?

Re: comments on pgnus-v0.80

[Lars Magne Ingebrigtsen]

Hrvoje Niksic <hniksic@srce.hr> writes:

> > then that would mean that text/html is really discouraged, but
> > "text/enriched" is less discouraged.  If you get my drift.
> 
> Yes.  Yes.  YES!

I'll take that as a positive.  :-)

> > This variable would default to nil, meaning that the precedence by
> > default would be the order of the parts.
> 
> ...and meaning that Steve's Mom could do
> 
>     (setq mm-alternative-discouraged-precedence '("text/html"))

Yup.  Although I called the variable `mm-discouraged-alternatives'. 

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

Re: comments on pgnus-v0.80

[Lars Magne Ingebrigtsen]

Per Abrahamsen <abraham@dina.kvl.dk> writes:

> >   (mml-insert-tag 'part 'type type 'filename file 'disposition "attachment"
> > 		  'description description))
> 
> Wouldn't it be more lispy to use :keywords, like
> 
>     (mml-insert-tag 'part :type type :filename file :disposition "attachment"
>                     :description description))
> 
> or did I totally misunderstood the syntax above?

Keywords would be the thing in `mml-insert-part' recognized certain
keywords, but what the function does is just insert the name of the
symbols.  So it would make sense to say

(mml-insert-tag "part" "type" type "filename" file ...)

instead, but I kinda like symbols for stuff like that.  

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

Re: comments on pgnus-v0.80

[Hrvoje Niksic]

Per Abrahamsen <abraham@dina.kvl.dk> writes:

> Lars Magne Ingebrigtsen <larsi@gnus.org> writes:
> 
> >   (mml-insert-tag 'part 'type type 'filename file 'disposition "attachment"
> > 		  'description description))
> 
> Wouldn't it be more lispy to use :keywords, like
> 
>     (mml-insert-tag 'part :type type :filename file :disposition "attachment"
>                     :description description)

The difference between symbols and keywords here is that symbols can
be anything, but keywords are usually enumerations.  So you can do:

(mml-insert-tag 'foo 'bar "qux")

and <!#foo bar=qux> will be inserted into the buffer.

You can replace symbols with strings, e.g.
(mml-insert-tag "foo" "bar" "qux") will do the same as the above.

I would reserve keywords for options, such as `:close-tag t' or stuff
like that.


Hrvoje, author of mml-insert-tag.

Re: comments on pgnus-v0.80

[Hrvoje Niksic]

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

> (mml-insert-tag "part" "type" type "filename" file ...)
> 
> instead, but I kinda like symbols for stuff like that.

You should be able do that, if you want to.