goto-address-at-point doesn't get the right url when there is a trailing ')'

goto-address-at-point doesn't get the right url when there is a trailing ')'

[Amos Bird]

Deer emacs devs,

When doing  (goto-address-at-point) at parenthesis surrounded urls like this

(https://github.com/nikomatsakis/rayon/issues/231) in any prog mode (c, java, python ...)

it will grab the trailing ')' thus ends up with opening a wrong url.

it works fine in text-mode.

Could you tell me a workaround for this?

Thank you!

Amos

Re: goto-address-at-point doesn't get the right url when there is a trailing ')'

[Yuri Khan]

On Wed, Jun 14, 2017 at 5:11 PM, Amos Bird <amosbird@gmail.com> wrote:

> When doing  (goto-address-at-point) at parenthesis surrounded urls like this
> (https://github.com/nikomatsakis/rayon/issues/231) in any prog mode (c, java, python ...)
>
> it will grab the trailing ')' thus ends up with opening a wrong url.

> Could you tell me a workaround for this?

Don’t do that, then?

A closing parenthesis is a valid URL character. Any heuristic that
fixes your problem will break legal, standard-conforming URLs, such as
this:

    https://en.wikipedia.org/wiki/Dunnet_(video_game)

If you want to enclose a URL in parentheses, put a pair of angle
brackets <> around it first:

    (<https://github.com/nikomatsakis/rayon/issues/231>)

Angle brackets are not valid URL characters and are expicitly
recommended as delimiters around URLs by RFC 2396 § 2.4.3 and RFC 3986
Appendix C. Also, it is a Markdown construct for URLs displayed as-is.

Re: goto-address-at-point doesn't get the right url when there is a trailing ')'

[Amos Bird]

Thanks! will do.

Yuri Khan <yuri.v.khan@gmail.com> writes:

> On Wed, Jun 14, 2017 at 5:11 PM, Amos Bird <amosbird@gmail.com> wrote:
>
>> When doing  (goto-address-at-point) at parenthesis surrounded urls like this
>> (https://github.com/nikomatsakis/rayon/issues/231) in any prog mode (c, java, python ...)
>>
>> it will grab the trailing ')' thus ends up with opening a wrong url.
>
>> Could you tell me a workaround for this?
>
> Don’t do that, then?
>
> A closing parenthesis is a valid URL character. Any heuristic that
> fixes your problem will break legal, standard-conforming URLs, such as
> this:
>
>     https://en.wikipedia.org/wiki/Dunnet_(video_game)
>
> If you want to enclose a URL in parentheses, put a pair of angle
> brackets <> around it first:
>
>     (<https://github.com/nikomatsakis/rayon/issues/231>)
>
> Angle brackets are not valid URL characters and are expicitly
> recommended as delimiters around URLs by RFC 2396 § 2.4.3 and RFC 3986
> Appendix C. Also, it is a Markdown construct for URLs displayed as-is.

Re: goto-address-at-point doesn't get the right url when there is a trailing ')'

[Philipp Stephani]

[-- Attachment #1: Type: text/plain, Size: 866 bytes --]

Yuri Khan <yuri.v.khan@gmail.com> schrieb am Mi., 14. Juni 2017 um
17:03 Uhr:

> On Wed, Jun 14, 2017 at 5:11 PM, Amos Bird <amosbird@gmail.com> wrote:
>
> > When doing  (goto-address-at-point) at parenthesis surrounded urls like
> this
> > (https://github.com/nikomatsakis/rayon/issues/231) in any prog mode (c,
> java, python ...)
> >
> > it will grab the trailing ')' thus ends up with opening a wrong url.
>
> > Could you tell me a workaround for this?
>
> Don’t do that, then?
>
> A closing parenthesis is a valid URL character. Any heuristic that
> fixes your problem will break legal, standard-conforming URLs, such as
> this:
>
>     https://en.wikipedia.org/wiki/Dunnet_(video_game)
>

The URL heuristic could match only paired parentheses. At least that's what
my mail app appears to be doing, as it correctly matches both URLs.

[-- Attachment #2: Type: text/html, Size: 1447 bytes --]

Re: goto-address-at-point doesn't get the right url when there is a trailing ')'

[Amos Bird]

hmm, can you try doing it at a prog mode? say cc-mode or java-mode.

Philipp Stephani <p.stephani2@gmail.com> writes:

> Yuri Khan <yuri.v.khan@gmail.com> schrieb am Mi., 14. Juni 2017 um
> 17:03 Uhr:
>
>> On Wed, Jun 14, 2017 at 5:11 PM, Amos Bird <amosbird@gmail.com> wrote:
>>
>> > When doing  (goto-address-at-point) at parenthesis surrounded urls like
>> this
>> > (https://github.com/nikomatsakis/rayon/issues/231) in any prog mode (c,
>> java, python ...)
>> >
>> > it will grab the trailing ')' thus ends up with opening a wrong url.
>>
>> > Could you tell me a workaround for this?
>>
>> Don’t do that, then?
>>
>> A closing parenthesis is a valid URL character. Any heuristic that
>> fixes your problem will break legal, standard-conforming URLs, such as
>> this:
>>
>>     https://en.wikipedia.org/wiki/Dunnet_(video_game)
>>
>
> The URL heuristic could match only paired parentheses. At least that's what
> my mail app appears to be doing, as it correctly matches both URLs.

Re: goto-address-at-point doesn't get the right url when there is a trailing ')'

[Philipp Stephani]

[-- Attachment #1.1: Type: text/plain, Size: 1269 bytes --]

The attached patch appears to do the right thing for me (but I haven't
tested it extensively).

Amos Bird <amosbird@gmail.com> schrieb am Do., 15. Juni 2017 um 11:04 Uhr:

>
> hmm, can you try doing it at a prog mode? say cc-mode or java-mode.
>
> Philipp Stephani <p.stephani2@gmail.com> writes:
>
> > Yuri Khan <yuri.v.khan@gmail.com> schrieb am Mi., 14. Juni 2017 um
> > 17:03 Uhr:
> >
> >> On Wed, Jun 14, 2017 at 5:11 PM, Amos Bird <amosbird@gmail.com> wrote:
> >>
> >> > When doing  (goto-address-at-point) at parenthesis surrounded urls
> like
> >> this
> >> > (https://github.com/nikomatsakis/rayon/issues/231) in any prog mode
> (c,
> >> java, python ...)
> >> >
> >> > it will grab the trailing ')' thus ends up with opening a wrong url.
> >>
> >> > Could you tell me a workaround for this?
> >>
> >> Don’t do that, then?
> >>
> >> A closing parenthesis is a valid URL character. Any heuristic that
> >> fixes your problem will break legal, standard-conforming URLs, such as
> >> this:
> >>
> >>     https://en.wikipedia.org/wiki/Dunnet_(video_game)
> >>
> >
> > The URL heuristic could match only paired parentheses. At least that's
> what
> > my mail app appears to be doing, as it correctly matches both URLs.
>
>

[-- Attachment #1.2: Type: text/html, Size: 2115 bytes --]

[-- Attachment #2: 0001-Correctly-detect-URLs-surrounded-by-parentheses-in-com.txt --]
[-- Type: text/plain, Size: 1178 bytes --]

From 9c780b6ef06cacad989ea6175944af95c3973f14 Mon Sep 17 00:00:00 2001
From: Philipp Stephani <phst@google.com>
Date: Thu, 15 Jun 2017 11:49:56 +0200
Subject: [PATCH] Correctly detect URLs surrounded by parentheses in comments

* lisp/thingatpt.el (thing-at-point--bounds-of-well-formed-url):
Make parentheses match work inside comments.
---
 lisp/thingatpt.el | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/lisp/thingatpt.el b/lisp/thingatpt.el
index 25e01df70e..7c3d73e52b 100644
--- a/lisp/thingatpt.el
+++ b/lisp/thingatpt.el
@@ -380,7 +380,9 @@ thing-at-point--bounds-of-well-formed-url
 	     (save-restriction
 	       (narrow-to-region (1- url-beg) (min end (point-max)))
 	       (setq paren-end (ignore-errors
-				 (scan-lists (1- url-beg) 1 0))))
+                                 ;; Make the scan work inside comments.
+                                 (let ((parse-sexp-ignore-comments nil))
+                                   (scan-lists (1- url-beg) 1 0)))))
 	     (not (blink-matching-check-mismatch (1- url-beg) paren-end))
 	     (setq end (1- paren-end)))
 	;; Ensure PT is actually within BOUNDARY. Check the following
-- 
2.13.1

Re: goto-address-at-point doesn't get the right url when there is a trailing ')'

[Amos Bird]

Thanks Philipp! I did some tests and it works quite well.

regards,
Amos

Philipp Stephani <p.stephani2@gmail.com> writes:

> The attached patch appears to do the right thing for me (but I haven't
> tested it extensively).
>
> Amos Bird <amosbird@gmail.com> schrieb am Do., 15. Juni 2017 um 11:04 Uhr:
>
>>
>> hmm, can you try doing it at a prog mode? say cc-mode or java-mode.
>>
>> Philipp Stephani <p.stephani2@gmail.com> writes:
>>
>> > Yuri Khan <yuri.v.khan@gmail.com> schrieb am Mi., 14. Juni 2017 um
>> > 17:03 Uhr:
>> >
>> >> On Wed, Jun 14, 2017 at 5:11 PM, Amos Bird <amosbird@gmail.com> wrote:
>> >>
>> >> > When doing  (goto-address-at-point) at parenthesis surrounded urls
>> like
>> >> this
>> >> > (https://github.com/nikomatsakis/rayon/issues/231) in any prog mode
>> (c,
>> >> java, python ...)
>> >> >
>> >> > it will grab the trailing ')' thus ends up with opening a wrong url.
>> >>
>> >> > Could you tell me a workaround for this?
>> >>
>> >> Don’t do that, then?
>> >>
>> >> A closing parenthesis is a valid URL character. Any heuristic that
>> >> fixes your problem will break legal, standard-conforming URLs, such as
>> >> this:
>> >>
>> >>     https://en.wikipedia.org/wiki/Dunnet_(video_game)
>> >>
>> >
>> > The URL heuristic could match only paired parentheses. At least that's
>> what
>> > my mail app appears to be doing, as it correctly matches both URLs.
>>
>>
> From 9c780b6ef06cacad989ea6175944af95c3973f14 Mon Sep 17 00:00:00 2001
> From: Philipp Stephani <phst@google.com>
> Date: Thu, 15 Jun 2017 11:49:56 +0200
> Subject: [PATCH] Correctly detect URLs surrounded by parentheses in comments
>
> * lisp/thingatpt.el (thing-at-point--bounds-of-well-formed-url):
> Make parentheses match work inside comments.
> ---
>  lisp/thingatpt.el | 4 +++-
>  1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/lisp/thingatpt.el b/lisp/thingatpt.el
> index 25e01df70e..7c3d73e52b 100644
> --- a/lisp/thingatpt.el
> +++ b/lisp/thingatpt.el
> @@ -380,7 +380,9 @@ thing-at-point--bounds-of-well-formed-url
>  	     (save-restriction
>  	       (narrow-to-region (1- url-beg) (min end (point-max)))
>  	       (setq paren-end (ignore-errors
> -				 (scan-lists (1- url-beg) 1 0))))
> +                                 ;; Make the scan work inside comments.
> +                                 (let ((parse-sexp-ignore-comments nil))
> +                                   (scan-lists (1- url-beg) 1 0)))))
>  	     (not (blink-matching-check-mismatch (1- url-beg) paren-end))
>  	     (setq end (1- paren-end)))
>  	;; Ensure PT is actually within BOUNDARY. Check the following

Re: goto-address-at-point doesn't get the right url when there is a trailing ')'

[Philipp Stephani]

[-- Attachment #1: Type: text/plain, Size: 3084 bytes --]

OK, I've pushed this as 3b6e01cccf89ba0f3485751125f43463bc429345 to master
and added a unit test.

Amos Bird <amosbird@gmail.com> schrieb am Do., 15. Juni 2017 um 15:49 Uhr:

>
> Thanks Philipp! I did some tests and it works quite well.
>
> regards,
> Amos
>
> Philipp Stephani <p.stephani2@gmail.com> writes:
>
> > The attached patch appears to do the right thing for me (but I haven't
> > tested it extensively).
> >
> > Amos Bird <amosbird@gmail.com> schrieb am Do., 15. Juni 2017 um 11:04
> Uhr:
> >
> >>
> >> hmm, can you try doing it at a prog mode? say cc-mode or java-mode.
> >>
> >> Philipp Stephani <p.stephani2@gmail.com> writes:
> >>
> >> > Yuri Khan <yuri.v.khan@gmail.com> schrieb am Mi., 14. Juni 2017 um
> >> > 17:03 Uhr:
> >> >
> >> >> On Wed, Jun 14, 2017 at 5:11 PM, Amos Bird <amosbird@gmail.com>
> wrote:
> >> >>
> >> >> > When doing  (goto-address-at-point) at parenthesis surrounded urls
> >> like
> >> >> this
> >> >> > (https://github.com/nikomatsakis/rayon/issues/231) in any prog
> mode
> >> (c,
> >> >> java, python ...)
> >> >> >
> >> >> > it will grab the trailing ')' thus ends up with opening a wrong
> url.
> >> >>
> >> >> > Could you tell me a workaround for this?
> >> >>
> >> >> Don’t do that, then?
> >> >>
> >> >> A closing parenthesis is a valid URL character. Any heuristic that
> >> >> fixes your problem will break legal, standard-conforming URLs, such
> as
> >> >> this:
> >> >>
> >> >>     https://en.wikipedia.org/wiki/Dunnet_(video_game)
> >> >>
> >> >
> >> > The URL heuristic could match only paired parentheses. At least that's
> >> what
> >> > my mail app appears to be doing, as it correctly matches both URLs.
> >>
> >>
> > From 9c780b6ef06cacad989ea6175944af95c3973f14 Mon Sep 17 00:00:00 2001
> > From: Philipp Stephani <phst@google.com>
> > Date: Thu, 15 Jun 2017 11:49:56 +0200
> > Subject: [PATCH] Correctly detect URLs surrounded by parentheses in
> comments
> >
> > * lisp/thingatpt.el (thing-at-point--bounds-of-well-formed-url):
> > Make parentheses match work inside comments.
> > ---
> >  lisp/thingatpt.el | 4 +++-
> >  1 file changed, 3 insertions(+), 1 deletion(-)
> >
> > diff --git a/lisp/thingatpt.el b/lisp/thingatpt.el
> > index 25e01df70e..7c3d73e52b 100644
> > --- a/lisp/thingatpt.el
> > +++ b/lisp/thingatpt.el
> > @@ -380,7 +380,9 @@ thing-at-point--bounds-of-well-formed-url
> >            (save-restriction
> >              (narrow-to-region (1- url-beg) (min end (point-max)))
> >              (setq paren-end (ignore-errors
> > -                              (scan-lists (1- url-beg) 1 0))))
> > +                                 ;; Make the scan work inside comments.
> > +                                 (let ((parse-sexp-ignore-comments nil))
> > +                                   (scan-lists (1- url-beg) 1 0)))))
> >            (not (blink-matching-check-mismatch (1- url-beg) paren-end))
> >            (setq end (1- paren-end)))
> >       ;; Ensure PT is actually within BOUNDARY. Check the following
>
>

[-- Attachment #2: Type: text/html, Size: 4605 bytes --]