Testing gnus-html here and there

Testing gnus-html here and there

[Lars Magne Ingebrigtsen]

I've now tested gnus-html under Emacs 22 on a "clean" machine, and I
found a missing require, but nothing major...

Oh, and I found that you have to say

(require 'gnus-load)

to get the autoloads rights.  Is that supposed to be that way?  I kinda
imagined that Gnus would load the gnus-load file automatically.

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

gnus-load.el (was: Testing gnus-html here and there)

[Reiner Steib]

On Tue, Aug 31 2010, Lars Magne Ingebrigtsen wrote:

> Oh, and I found that you have to say
>
> (require 'gnus-load)
>
> to get the autoloads rights.  Is that supposed to be that way?  

At least it is documented this way in GNUS-NEWS and the README. :-)

,----[ (info "(gnus)Oort Gnus") ]
|         * `(require 'gnus-load)'
| 
|           If you use a stand-alone Gnus distribution, you'd better add
|           `(require 'gnus-load)' into your `~/.emacs' after adding the
|           Gnus lisp directory into load-path.
| 
|           File `gnus-load.el' contains autoload commands, functions and
|           variables, some of which may not be included in distributions
|           of Emacsen.
`----

(Probably it should be in some other node.)

> I kinda imagined that Gnus would load the gnus-load file
> automatically.

It is not necessary for the Gnus version bundled with Emacs.

Bye, Reiner.
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo---  |  PGP key available  |  http://rsteib.home.pages.de/

Re: Testing gnus-html here and there

[Katsumi Yamaoka]

[-- Attachment #1: Type: text/plain, Size: 306 bytes --]

Lars Magne Ingebrigtsen wrote:
> I've now tested gnus-html under Emacs 22 on a "clean" machine, and I
> found a missing require, but nothing major...
[...]

`set-process-plist' and `process-plist' that gnus-html.el uses
are not available in Emacs 21 and XEmacs.  Aren't they something
like the following?


[-- Attachment #2: Type: application/emacs-lisp, Size: 952 bytes --]

[-- Attachment #3: Type: text/plain, Size: 108 bytes --]


I haven't yet reached to displaying images using `gnus-article-html',
so I'm not sure it is really useful.

Re: Testing gnus-html here and there

[Katsumi Yamaoka]

Some html messages contain image data referenced to by cid tags.
I wish `gnus-article-html' could display them as well.  Here's
an example message:

ftp://ftp.jpl.org/pub/tmp/cid-example.gz

This is what nnshimbun (bundled with emacs-w3m) created.  The same
message can be found as the article 2866 in the gwene.jp.asahi-com
newsgroup (only a link is there, though).

Re: Testing gnus-html here and there

[Steinar Bang]

>>>>> Katsumi Yamaoka <yamaoka@jpl.org>:

> `set-process-plist' and `process-plist' that gnus-html.el uses
> are not available in Emacs 21 and XEmacs.  Aren't they something
> like the following?

FWIW I still have some RHEL4 machines that has emacs 21, and I use git
gnus on them (which I have to update if I want Ted's gnus-sync.el
improvements (which I do)).

So... I would be happier if the default gnus configuration would still
build and run on emacs21.

Re: Testing gnus-html here and there

[Ulf Stegemann]

Katsumi Yamaoka <yamaoka@jpl.org> wrote:

> Some html messages contain image data referenced to by cid tags.
> I wish `gnus-article-html' could display them as well.

1+

> This is what nnshimbun (bundled with emacs-w3m) created.  The same
> message can be found as the article 2866 in the gwene.jp.asahi-com
> newsgroup (only a link is there, though).

Another issue with nnshimbun is that shimbun articles may contain
relative link targets (as their source is a web page, not an HTML mail).
Those are presented by gnus-html `as they are', i.e. they are not
prefixed with host/path of the original source page.  This frequently
creates broken links with nnshimbun (for examples take a look at nearly
any article in `nnshimbun+heise:news').

I don't know if doing link target re-writing is beyond the scope of the
intended /simple/ HTML viewer but without that feature gnus-html would
probably not be usable for nnshimbun.  I'm not so much into RSS, but I
can imagine that relative link targets may appear there, too, which in
turn would also lead to problems with gwene articles.

Ulf

Re: Testing gnus-html here and there

[Lars Magne Ingebrigtsen]

Katsumi Yamaoka <yamaoka@jpl.org> writes:

> Some html messages contain image data referenced to by cid tags.
> I wish `gnus-article-html' could display them as well. 

That's easy enough.  Pushed out now.

I'm a bit confused by some of the output from w3m, though.  In the test
mail I tried (work-related), the output was the following:

-----------
<img_alt src="cid:image012.png@01CB485E.8BA6A0F0" title="https://www.saxobank.com/Advertisements/email/2010/partnerNotification/AUGUST/images/header_inst_notitfication.png">https://www.saxobank.com/Advertisements/email/2010/</img_alt>                                         
<IMG_ALT SRC="cid:image012.png@01CB485E.8BA6A0F0">partnerNotification/AUGUST/images/header_inst_notitfication.png</img_alt>                             
-----------

So the cid: picture appeared twice?  But once with lower-case "img_alt"
and once with upper-case "IMG_ALT"?  gnus-html ignores the latter, and
just renders the text between the tags, but should probably delete it
altogether.  But I just wonder what the logic is here?

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

Re: gnus-load.el

[Lars Magne Ingebrigtsen]

Reiner Steib <reinersteib+gmane@imap.cc> writes:

> |           If you use a stand-alone Gnus distribution, you'd better add
> |           `(require 'gnus-load)' into your `~/.emacs' after adding the
> |           Gnus lisp directory into load-path.

Hm.  But why?  Can't the development version of Gnus just do a

(require 'gnus-load)

in gnus.el?  It can be removed before syncing with Emacs.

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

Re: Testing gnus-html here and there

[Lars Magne Ingebrigtsen]

Katsumi Yamaoka <yamaoka@jpl.org> writes:

>   (defun gnus-set-process-plist (process plist)
>     "Replace the plist of PROCESS with PLIST.  Returns PLIST."
>     (put 'gnus-process-plist process plist))

Hm.  Can I just use `put' directly on a process object?  That didn't
occur to me...

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

Re: Testing gnus-html here and there

[Lars Magne Ingebrigtsen]

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

>>   (defun gnus-set-process-plist (process plist)
>>     "Replace the plist of PROCESS with PLIST.  Returns PLIST."
>>     (put 'gnus-process-plist process plist))
>
> Hm.  Can I just use `put' directly on a process object?  That didn't
> occur to me...

Oh, never mind.  The other way around.

But aren't you just using this as a global variable here, sort of?
Which makes sense, I guess.  I'll install your code.

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

Re: Testing gnus-html here and there

[Ted Zlatanov]

On Tue, 31 Aug 2010 10:17:46 +0200 Steinar Bang <sb@dod.no> wrote: 

>>>>>> Katsumi Yamaoka <yamaoka@jpl.org>:
>> `set-process-plist' and `process-plist' that gnus-html.el uses
>> are not available in Emacs 21 and XEmacs.  Aren't they something
>> like the following?

SB> FWIW I still have some RHEL4 machines that has emacs 21, and I use git
SB> gnus on them (which I have to update if I want Ted's gnus-sync.el
SB> improvements (which I do)).

SB> So... I would be happier if the default gnus configuration would still
SB> build and run on emacs21.

You have to admit that's a big maintainer burden, though, for the sake
of very few users in 2010.  I would make 22 the minimum.

Ted

Re: Testing gnus-html here and there

[Lars Magne Ingebrigtsen]

Ted Zlatanov <tzz@lifelogs.com> writes:

> You have to admit that's a big maintainer burden, though, for the sake
> of very few users in 2010.  I would make 22 the minimum.

True.  I would guess that none of the mainstream Linux distributions
haven't shipped Emacs 21 in a few years.

However, the code Katsumi posted also made the gnus-html code work in
XEmacs, which is worthwhile.

I think...  :-)

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

Re: Testing gnus-html here and there

[Ted Zlatanov]

On Tue, 31 Aug 2010 20:29:13 +0200 Lars Magne Ingebrigtsen <larsi@gnus.org> wrote: 

LMI> Ted Zlatanov <tzz@lifelogs.com> writes:
>> You have to admit that's a big maintainer burden, though, for the sake
>> of very few users in 2010.  I would make 22 the minimum.

LMI> However, the code Katsumi posted also made the gnus-html code work in
LMI> XEmacs, which is worthwhile.

I agree, XEmacs is definitely worthwhile.

Ted

Re: Testing gnus-html here and there

[Katsumi Yamaoka]

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

>>>   (defun gnus-set-process-plist (process plist)
>>>     "Replace the plist of PROCESS with PLIST.  Returns PLIST."
>>>     (put 'gnus-process-plist process plist))
>>
>> Hm.  Can I just use `put' directly on a process object?  That didn't
>> occur to me...

> Oh, never mind.  The other way around.

> But aren't you just using this as a global variable here, sort of?
> Which makes sense, I guess.  I'll install your code.

Thank you for the installation.  Er, I don't see what you suspect,
but it's a mere impromptu idea.  There will be a couple of ways to
do that.  For instance:

(defvar gnus-process-plist-alist nil)
(set-alist 'gnus-process-plist-alist PROCESS PLIST)

P.S. I'll add `gnus-process-get' later for the compatibility with
Emacs 21 and XEmacs.

Re: Testing gnus-html here and there

[Katsumi Yamaoka]

Lars Magne Ingebrigtsen wrote:
> Katsumi Yamaoka <yamaoka@jpl.org> writes:

>> Some html messages contain image data referenced to by cid tags.
>> I wish `gnus-article-html' could display them as well.

> That's easy enough.  Pushed out now.

Excellent.  Thank you very much.
In addition to this, I've made `gnus-article-html' decode raw html
contents by the charset that the Content-Type header specifies
before passing to w3m -halfdump.  At least Japanese html articles
got human-readable.  Cf.: ftp://ftp.jpl.org/pub/tmp/cid-example.gz

> I'm a bit confused by some of the output from w3m, though.  In the test
> mail I tried (work-related), the output was the following:

> -----------
> <img_alt src="cid:image012.png@01CB485E.8BA6A0F0" title="https://www.saxobank.com/Advertisements/email/2010/partnerNotification/AUGUST/images/header_inst_notitfication.png">https://www.saxobank.com/Advertisements/email/2010/</img_alt>
> <IMG_ALT SRC="cid:image012.png@01CB485E.8BA6A0F0">partnerNotification/AUGUST/images/header_inst_notitfication.png</img_alt>
> -----------

> So the cid: picture appeared twice?  But once with lower-case "img_alt"
> and once with upper-case "IMG_ALT"?  gnus-html ignores the latter, and
> just renders the text between the tags, but should probably delete it
> altogether.  But I just wonder what the logic is here?

I'll look into this.

Re: Testing gnus-html here and there

[Katsumi Yamaoka]

Katsumi Yamaoka wrote:
>> I'm a bit confused by some of the output from w3m, though.  In the test
>> mail I tried (work-related), the output was the following:

>> -----------
>> <img_alt src="cid:image012.png@01CB485E.8BA6A0F0" title="https://www.saxobank.com/Advertisements/email/2010/partnerNotification/AUGUST/images/header_inst_notitfication.png">https://www.saxobank.com/Advertisements/email/2010/</img_alt>
>> <IMG_ALT SRC="cid:image012.png@01CB485E.8BA6A0F0">partnerNotification/AUGUST/images/header_inst_notitfication.png</img_alt>
>> -----------

>> So the cid: picture appeared twice?  But once with lower-case "img_alt"
>> and once with upper-case "IMG_ALT"?  gnus-html ignores the latter, and
>> just renders the text between the tags, but should probably delete it
>> altogether.  But I just wonder what the logic is here?

> I'll look into this.

I don't know why the upper-case IMG_ALT tag appears, but I recalled
it's been existing since the early days of emacs-w3m.  It seems to
appear if the alt string of an image consists of two or more lines.
Emacs-w3m does a workaround anyway.  Gnus may want to do something
similar (deleting "<IMG_ALT ...>...</img_alt>" ?).  Sorry for the
helpless information.

Re: Testing gnus-html here and there

[Lars Magne Ingebrigtsen]

Katsumi Yamaoka <yamaoka@jpl.org> writes:

> In addition to this, I've made `gnus-article-html' decode raw html
> contents by the charset that the Content-Type header specifies
> before passing to w3m -halfdump.  At least Japanese html articles
> got human-readable.  Cf.: ftp://ftp.jpl.org/pub/tmp/cid-example.gz

Great!

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

Re: Testing gnus-html here and there

[Lars Magne Ingebrigtsen]

Katsumi Yamaoka <yamaoka@jpl.org> writes:

> Gnus may want to do something similar (deleting "<IMG_ALT
> ...>...</img_alt>" ?).  Sorry for the helpless information.

Yeah, I'll just have it delete the IMG_ALT region.

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

Re: Testing gnus-html here and there

[Malcolm Purvis]

>>>>> "Ted" == Ted Zlatanov <tzz@lifelogs.com> writes:

Ted> I agree, XEmacs is definitely worthwhile.

As someone whose mail client at home and work is bleeding-edge gnus on
XEmacs, I agree.

Malcolm


-- 
		     Malcolm Purvis <malcolmp@xemacs.org>