Internal/external browsers

Internal/external browsers

[Lars Magne Ingebrigtsen]

When you hit RET on a link in shr/eww, it uses the browse-url browser to
follow the link.  If you hit `C-u RET', it uses the
`shr-external-browser' browser.

It's been that way for some months now, and I must say I really like
that.  I have eww set as the default browser, and quite a few web pages
display well enough in eww, and it's much more pleasant to read those
web pages inside Emacs.

And when I know that the link points to a page that eww is unlikely to
cope with, I just hit `C-u RET' and get a Firefox browser window popping
up.

So now I want all URL links in Emacs to work that way, like in erc mode,
etc.

Does this make sense to people?  If so, I think browse-url itself should
be extended with a browse-url-external-browser variable, and then the
`browse-url' function itself should use these two variables, instead of
shr implementing this stuff itself...

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

Re: Internal/external browsers

[Ted Zlatanov]

On Sat, 14 Dec 2013 18:57:56 +0100 Lars Magne Ingebrigtsen <larsi@gnus.org> wrote: 

LMI> When you hit RET on a link in shr/eww, it uses the browse-url browser to
LMI> follow the link.  If you hit `C-u RET', it uses the
LMI> `shr-external-browser' browser.

LMI> It's been that way for some months now, and I must say I really like
LMI> that.  I have eww set as the default browser, and quite a few web pages
LMI> display well enough in eww, and it's much more pleasant to read those
LMI> web pages inside Emacs.

LMI> And when I know that the link points to a page that eww is unlikely to
LMI> cope with, I just hit `C-u RET' and get a Firefox browser window popping
LMI> up.

LMI> So now I want all URL links in Emacs to work that way, like in erc mode,
LMI> etc.

LMI> Does this make sense to people?  If so, I think browse-url itself should
LMI> be extended with a browse-url-external-browser variable, and then the
LMI> `browse-url' function itself should use these two variables, instead of
LMI> shr implementing this stuff itself...

To me, that makes good sense everywhere in Emacs.  Except I'd like `C-u
RET' to show a list of browsers, if I have more than one I like to use.

Ted

Re: Internal/external browsers

[Lars Magne Ingebrigtsen]

Ted Zlatanov <tzz@lifelogs.com> writes:

> To me, that makes good sense everywhere in Emacs.  Except I'd like `C-u
> RET' to show a list of browsers, if I have more than one I like to use.

Yeah, that might make sense.  But is that a common use case?  I only use
Firefox for "real browser" stuff, and I don't really see why I'd switch
between Firefox and (say) Chrome on a regular basis...

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

Re: Internal/external browsers

[Josh]

On Sat, Dec 14, 2013 at 10:12 AM, Lars Magne Ingebrigtsen
<larsi@gnus.org> wrote:
> Ted Zlatanov <tzz@lifelogs.com> writes:
>
>> To me, that makes good sense everywhere in Emacs.  Except I'd like `C-u
>> RET' to show a list of browsers, if I have more than one I like to use.
>
> Yeah, that might make sense.  But is that a common use case?  I only use
> Firefox for "real browser" stuff, and I don't really see why I'd switch
> between Firefox and (say) Chrome on a regular basis...

I can see why people might want to, for example if the link target
led to a page the user knew to have features better supported by one
browser or another, or if she were a web developer who wanted to
check the linked page's behavior in a particular browser, or even to
launch the usual browser with different command-line switches.

Perhaps C-u RET could immediately open the link in the user's
preferred "alternate" browser and C-u C-u RET could prompt the user
for the complete browser invocation à la `M-x compile', with the
command history being seeded with invocations of one or more
standard external browsers.

Re: Internal/external browsers

[joakim]

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

> Ted Zlatanov <tzz@lifelogs.com> writes:
>
>> To me, that makes good sense everywhere in Emacs.  Except I'd like `C-u
>> RET' to show a list of browsers, if I have more than one I like to use.
>
> Yeah, that might make sense.  But is that a common use case?  I only use
> Firefox for "real browser" stuff, and I don't really see why I'd switch
> between Firefox and (say) Chrome on a regular basis...

I normally use a bunch of browsers for various reasons. But patching the
upcoming function for my use-case will be good enough for me if thats
not deemed common.
-- 
Joakim Verona

Re: Internal/external browsers

[Ted Zlatanov]

On Sat, 14 Dec 2013 19:12:23 +0100 Lars Magne Ingebrigtsen <larsi@gnus.org> wrote: 

LMI> Ted Zlatanov <tzz@lifelogs.com> writes:
>> To me, that makes good sense everywhere in Emacs.  Except I'd like `C-u
>> RET' to show a list of browsers, if I have more than one I like to use.

LMI> Yeah, that might make sense.  But is that a common use case?  I only use
LMI> Firefox for "real browser" stuff, and I don't really see why I'd switch
LMI> between Firefox and (say) Chrome on a regular basis...

For web development it's nice, but not a common case for sure.

Ted

Re: Internal/external browsers

[Mathias Dahl]

[-- Attachment #1: Type: text/plain, Size: 606 bytes --]

>
> >> To me, that makes good sense everywhere in Emacs.  Except I'd like `C-u
> >> RET' to show a list of browsers, if I have more than one I like to use.
>
> LMI> Yeah, that might make sense.  But is that a common use case?  I only
> use
> LMI> Firefox for "real browser" stuff, and I don't really see why I'd
> switch
> LMI> between Firefox and (say) Chrome on a regular basis...
>
> For web development it's nice, but not a common case for sure.


I agree. Maybe there should be a setting that web developers can turn on to
get that behaviour and let the simple behaviour be the default one?

/Mathias

[-- Attachment #2: Type: text/html, Size: 963 bytes --]

Re: Internal/external browsers

[Ted Zlatanov]

On Sun, 22 Dec 2013 14:50:55 +0100 Mathias Dahl <mathias.dahl@gmail.com> wrote: 

>> >> To me, that makes good sense everywhere in Emacs.  Except I'd like `C-u
>> >> RET' to show a list of browsers, if I have more than one I like to use.
>> 
LMI> Yeah, that might make sense.  But is that a common use case?  I only use
LMI> Firefox for "real browser" stuff, and I don't really see why I'd switch
LMI> between Firefox and (say) Chrome on a regular basis...
>> 
>> For web development it's nice, but not a common case for sure.

MD> I agree. Maybe there should be a setting that web developers can turn on to
MD> get that behaviour and let the simple behaviour be the default one?

I don't see what's wrong with `C-u RET' to always do it?  It seems like
a pretty common pattern in Emacs to have the universal-prefix used thus.

Ted