[PATCH][BUG]: eww doesnt trim whitespace from address

[PATCH][BUG]: eww doesnt trim whitespace from address

[Vibhav Pant]

[-- Attachment #1: Type: text/plain, Size: 1158 bytes --]

As of now, eww does not rim the user specified address, this often
leads to problems especially if the address is from the clipboard (eww
reports that the service isnt known). Most browsers automatically do
this, so I have included a patch to solve this problem.

This patch simply setq's url to a "trimmed" version of it.
The change is pretty non trivial, so I think it can be included in this release.

diff --git a/lisp/net/eww.el b/lisp/net/eww.el
index e8eb09c..aa3e245 100644
--- a/lisp/net/eww.el
+++ b/lisp/net/eww.el
@@ -153,6 +153,10 @@ See also `eww-form-checkbox-selected-symbol'."
 If the input doesn't look like an URL or a domain name, the
 word(s) will be searched for via `eww-search-prefix'."
   (interactive "sEnter URL or keywords: ")
+  (setq url (replace-regexp-in-string (rx (or (: bos (* (any " \t\n")))
+                                             (: (* (any " \t\n")) eos)))
+                                     ""
+                                     url))
   (cond ((string-match-p "\\`file://" url))
         ((string-match-p "\\`ftp://" url)
          (user-error "FTP is not supported."))

-- 
Vibhav Pant
vibhavp@gmail.com

[-- Attachment #2: ewwtrim.patch --]
[-- Type: text/x-patch, Size: 717 bytes --]

diff --git a/lisp/net/eww.el b/lisp/net/eww.el
index e8eb09c..aa3e245 100644
--- a/lisp/net/eww.el
+++ b/lisp/net/eww.el
@@ -153,6 +153,10 @@ See also `eww-form-checkbox-selected-symbol'."
 If the input doesn't look like an URL or a domain name, the
 word(s) will be searched for via `eww-search-prefix'."
   (interactive "sEnter URL or keywords: ")
+  (setq url (replace-regexp-in-string (rx (or (: bos (* (any " \t\n")))
+                                             (: (* (any " \t\n")) eos)))
+                                     ""
+                                     url))
   (cond ((string-match-p "\\`file://" url))
         ((string-match-p "\\`ftp://" url)
          (user-error "FTP is not supported."))

Re: [PATCH][BUG]: eww doesnt trim whitespace from address

[Ted Zlatanov]

On Fri, 21 Mar 2014 14:57:37 +0530 Vibhav Pant <vibhavp@gmail.com> wrote: 

VP> As of now, eww does not rim the user specified address, this often
VP> leads to problems especially if the address is from the clipboard (eww
VP> reports that the service isnt known). Most browsers automatically do
VP> this, so I have included a patch to solve this problem.

VP> This patch simply setq's url to a "trimmed" version of it.
VP> The change is pretty non trivial, so I think it can be included in this release.

I think the fairly recent function `string-trim' would be better than
the proposed regex approach.

I also don't know if there are any exceptions to this--are there cases
where you want to preserve the trailing space on a URL and/or encode
them to "%20"?

Thanks
Ted

Re: [PATCH][BUG]: eww doesnt trim whitespace from address

[Yuri Khan]

On Mon, Nov 3, 2014 at 7:46 PM, Ted Zlatanov <tzz@lifelogs.com> wrote:

> I also don't know if there are any exceptions to this--are there cases
> where you want to preserve the trailing space on a URL and/or encode
> them to "%20"?

For what it’s worth, the HTML5 specification now defines all
URL-valued attributes as “valid non-empty URL potentially surrounded
by spaces” and prescribes stripping out the surrounding spaces.

<a href=" &#x09;http://google.com/   &#x0a;">This is valid and leads
to Google</a>

eww could treat the user-specified URL similarly. Spaces at the end of
URL are rare (and at the beginning pretty much inexistent), and if the
user wants an explicit space at the end of URL, s?he can type an
explicit %20, + or other context-specific escape.

Re: [PATCH][BUG]: eww doesnt trim whitespace from address

[Ted Zlatanov]

On Mon, 3 Nov 2014 21:49:23 +0700 Yuri Khan <yuri.v.khan@gmail.com> wrote: 

YK> On Mon, Nov 3, 2014 at 7:46 PM, Ted Zlatanov <tzz@lifelogs.com> wrote:
>> I also don't know if there are any exceptions to this--are there cases
>> where you want to preserve the trailing space on a URL and/or encode
>> them to "%20"?

YK> For what it’s worth, the HTML5 specification now defines all
YK> URL-valued attributes as “valid non-empty URL potentially surrounded
YK> by spaces” and prescribes stripping out the surrounding spaces.

YK> <a href=" &#x09;http://google.com/   &#x0a;">This is valid and leads
YK> to Google</a>

YK> eww could treat the user-specified URL similarly. Spaces at the end of
YK> URL are rare (and at the beginning pretty much inexistent), and if the
YK> user wants an explicit space at the end of URL, s?he can type an
YK> explicit %20, + or other context-specific escape.

All right, I added a simple `string-trim' call to `eww'.

Thanks
Ted