[Date Prev][Date Next][Thread Prev][][Date Index][Thread Index]

[BUG][REGRESSION] Re: w3m-input-url



There seems to have been a regression of a bug that was documentated,
patched, and committed afew months ago in email list messages #12656,
#12659, and #12662.

The original problem was described as:


#+BEGIN_QUOTE
The attached patch corrects a nuisance I've had with the package since
I started using it. I hope it or an equivalent can be accepted.

The issue occurs when using command 'g', 'G', or 'U' to open a new url
or perform a search. When one enters something of the form xxx.xxx
without a protocol prefix xxxx://, the package interprets the input as
a search request.

My expectation was that inputing:

  https://google.com    go there, and it does
  google.com            go there, and it doesn't
  google                peform a google search on the word

The patch also interprets any input containing a space as a search
request, so the following will perform searches:

  what is google.com
   what.is.google.com
#+END_QUOTE

The original resolution was...

On 2017-05-30 04:34, Boruch Baum wrote:
> On 2017-05-30 17:23, Katsumi Yamaoka wrote:
> > In [emacs-w3m : No.12656]
> > . . .
> > I feel like a better place to modify would be `w3m-canonicalize-url'.
> > How about the following setting and the patch?
>
> If it works, I'm happy!
>
> >
> > (setq w3m-enable-google-feeling-lucky nil)
> >
>
> > --- w3m.el~	2017-05-29 02:40:19.000000000 +0000
> > +++ w3m.el	2017-05-30 08:21:42.041291600 +0000
> > @@ -4629,7 +4629,8 @@
> >  	replaced)
> >         ((match-beginning 1)
> >  	url)
> > -       (feeling-lucky
> > +       ((or feeling-lucky
> > +	    (string-match " " url))
> >  	(concat "\
> >  http://www.google.com/search?btnI=I%%27m+Feeling+Lucky&ie=UTF-8&oe=UTF-8&q=";
> >  		(w3m-url-encode-string url nil t)))
>

Is the regression a consequence of the 'feeling searchy' thread?

-- 
hkp://keys.gnupg.net
CA45 09B5 5351 7C11 A9D1  7286 0036 9E45 1595 8BC0