problem with https and w3-fetch ... sometimes

problem with https and w3-fetch ... sometimes

[prayner]

Quite recently my access to google searches using emacs/w3 has broken.
This is using the emacspeak auditory feedback package but I don't
believe the problem is there.
If I run w3-fetch on the following url
https://www.google.com/search?source=hp&q=%3E+lequere+2014+earth+system+data&num=25
I get an incomplete web page (no results) and a warning from gnutls
that the connection was improperly terminated. If I feed the same url
to google-chrome or firefox I get the complete page. 
I think I've chased this down to tls.el so I suspect something has
changed in gnutls-cli.
This is with emacs 24.3.1 and gnutls 3.1.25
Can anyone suggest routes to debugging this?
please reply directly as well as to the group.
thanks in advance
Peter
-- 
Peter Rayner
room 343 
School of Earth Sciences, University of Melbourne, 3010, Vic, Australia
tel: work: +61 (0)3 8344 9708; fax: +61 (0)3 8344 7761 
mobile +61 402 752 379, skype: petermorag 
mail-to: prayner@unimelb.edu.au
google scholar profile <http://scholar.google.com.au/citations?user=H3up71wAAAAJ&hl=en>

Re: problem with https and w3-fetch ... sometimes

[Emanuel Berg]

prayner <prayner@unimelb.edu.au> writes:

> Quite recently my access to google searches using
> emacs/w3 has broken.

What is w3? Is it some progenitor of Emacs-w3m?

If you'd consider using w3m, or w3m code perhaps still
works for you, I have written a lightning-fast system
for all kinds of searches - just bind it to a global
keystroke, and search (not just with Goggle) from
anywhere in Emacs - also, it suggests the region text
as default, and has other cool stuff as well.

http://user.it.uu.se/~embe8573/conf/emacs-init/w3m/search.el

-- 
underground experts united

Re: problem with https and w3-fetch ... sometimes

[Stefan Monnier]

> I think I've chased this down to tls.el so I suspect something has
> changed in gnutls-cli.
> This is with emacs 24.3.1 and gnutls 3.1.25
> Can anyone suggest routes to debugging this?

I'd be happy to take patches that make W3 use the open-network-stream
function's builtin TLS support (via libgnutls) instead of going through
gnutls-cli.


        Stefan

Re: problem with https and w3-fetch ... sometimes

[Bob Proulx]

Emanuel Berg wrote:
> prayner <prayner@unimelb.edu.au> writes:
> 
> > Quite recently my access to google searches using
> > emacs/w3 has broken.
> 
> What is w3? Is it some progenitor of Emacs-w3m?

I previously ran into mention of Emacs W3 and therefore had researched
these references already.

  http://www.emacswiki.org/emacs/w3

  https://en.wikipedia.org/wiki/Emacs/W3

> If you'd consider using w3m, or w3m code perhaps still
> works for you, I have written a lightning-fast system
> for all kinds of searches - just bind it to a global
> keystroke, and search (not just with Goggle) from
> anywhere in Emacs - also, it suggests the region text
> as default, and has other cool stuff as well.
> 
> http://user.it.uu.se/~embe8573/conf/emacs-init/w3m/search.el

Because the original poster mentions emacspeak I presume that is the
reason for using W3.  W3 is the first listed application in the web
task list.  I presume they might be visually impaired and have been
using it this way for quite some time and will be accustomed to the
way emacs W3 works.  But I think that emacspeak will work well with
w3m too because w3m is listed in the list of web applications for
emacspeak too.

  http://www.emacswiki.org/EmacSpeak

  http://emacspeak.sourceforge.net/

Bob

Re: problem with https and w3-fetch ... sometimes

[Bob Proulx]

Hello Peter,

prayner wrote:
> please reply directly as well as to the group.

Unfortunately several of us forgot to do this and there was some
discussion about your problem on the mailing list that did not include
you.  Although unfortunately none of us who responded knew of your
problem or how to address it directly.  So that probably doesn't
matter anyway.  Sorry.  You can follow the complete set of responses
in this next link to the mailing list archives.  This next link goes
to your message in the archive and the responses follow it there.

  http://lists.gnu.org/archive/html/help-gnu-emacs/2014-08/msg00108.html

However Emanuel Berg wrote this following that sounds hopeful to me:

    If you'd consider using w3m, or w3m code perhaps still
    works for you, I have written a lightning-fast system
    for all kinds of searches - just bind it to a global
    keystroke, and search (not just with Goggle) from
    anywhere in Emacs - also, it suggests the region text
    as default, and has other cool stuff as well.
    http://user.it.uu.se/~embe8573/conf/emacs-init/w3m/search.el

The link at the end of his response links directly to an emacs lisp
file with the code for w3m that he is suggesting you try.

Also Stefan Monnier wrote this following sentence:

    I'd be happy to take patches that make W3 use the open-network-stream
    function's builtin TLS support (via libgnutls) instead of going
    through gnutls-cli.

That sounds hopeful to me.  Someone familiar with emacs lisp and emacs
w3 could recode it to avoid the tls network problem.

> Can anyone suggest routes to debugging this?

I fear that few people are using Emacs W3 these days and perhaps no
one on this mailing list is using it.  You may need to find an emacs
w3 expert elsewhere.  I tried your example link with emacs w3m and the
full page was rendered okay.  I think someone familiar with Emacs W3
code would need to update W3 for whatever changes happened at Google.

I think that if w3m works for you then w3m would be much better
supported these days than W3.  Since w3m is listed as a supported
application in the emacspeak task list I would try it next.  Either
way please write us on the mailing list again and give us an update on
things.

Bob

Re: problem with https and w3-fetch ... sometimes

[Emanuel Berg]

(I'm sending this to the original posters mail as well,
and I did so with my previous reply as well. Doesn't
that show up? It is in the To: header, perhaps you
looked in the Cc: header?)

Bob Proulx <bob@proulx.com> writes:

> Because the original poster mentions emacspeak I
> presume that is the reason for using W3.

Oh, I didn't even react to that, which I should have,
because I even experimented with emacspeak a while back
and wrote about it on this very list.

This is what I wrote then - note that this was from my
situation, which (at that time) was very bad compared
to a fully healthy person, but still of course cannot
be compared to a severely impaired or blind person. The
reason I didn't like plain emacspeak was that it was
too verbose, so I thought I'd setup functions to have
it read certain long paragraphs for me - selectively,
not everything and always. Oh, and let me say I am
very, very confident in Emacs being the right tool for
lots of people with special situations, because of
Emacs configurability and the ease with which new
modules can be put to work. In my case, I was able to
do that myself, but I'm a programmer... You know the
other day when I said that programmers shouldn't invent
"new" programming languages and so on, they should put
their talents to help humanity? This sounds pompous to
be sure but actually it is very down-to-earth, and here
is a fine example (though I don't know any details).
This was a long excursion, this is what I wrote:

... it [emacspeak] was based on espeak, which is
available in the Debian repos as a stand-alone-tool,
and so I wrote

es () {
   espeak -s 130 -k 20 -v en -f $1 -w `basename $1 .txt`.wav
}

to create sound files from text, and it should be a
small task to setup defuns to interact with it.

... I don't know if it can replace reading. It will
certainly not be as enjoyable. But what do you do? All
tricks to compensate for the loss of information intake
should be explored.

-- 
underground experts united