Re: follow-link in grep buffer

[Reiner Steib]

On Fri, Feb 25 2005, David Kastrup wrote:

> I would recommend not making any behavior by default dependent on the
> single-click length, nor on the focus situation: both approaches are
> completely obfuscate and confusing.
>
> So the change to 21.4 behavior would be the following:
>
> a) a double-click on a location that has no double-click binding, but
>    a local mouse-2 binding, will execute the mouse-2 binding.  This
>    will make double-clicks, where not overriden, follow links without
>    further code changes compared to 21.4
> b) setting an explicit link property will _additionally_ remap mouse-1
>    to mouse-2.  I would not by default make this dependent on anything
>    else.

I agree with David's arguments and his suggestions.

I also find the current behavior confusing in some modes.  After
several week, I still sometimes follow the link when I intended to
select the window or set the point.

Some examples: In Gnus Group buffer, I often want to put the cursor on
a certain group and then I intend to do `C-u a' (post to group under
point), `10 RET' (show 10 article in that group), M-g (Check for new
news in the current topic.).  In PCL-CVS, I often want to put the
point on a file name and then do `=' (cvs-mode-diff), `l'
(cvs-mode-log) ...

Because I followed the discussions on emacs-devel, I remembered why
mouse-1 behaves like this, but I think for ordinary users it is really
surprising.

In modes like info, mouse-1 to follow a link is a good default.

> So grep should not set the link property in my opinion in this scheme,
> and gnus should set it sparingly: on MIME buttons it generates itself,
> but not on header lines or things looking like a link in the article
> text.

It should not be set (by default) in the group and summary buffer.  As
for the MIME buttons and other links in the article buffer: AFAICS,
these are the same kind of all "buttons", so it might require some
work.

[...]
> Telling them "use mouse-2" has always been sort of embarrassing.
> Telling them "use mouse-1, but don't press it longer than 200
> milliseconds if you want to follow the link, and it won't work if you
> have not the focus" will kill the "Emacs is usable for common human"
> proposition dead.

ACK.  In Gnus group buffer, I often found myself clicking too short so
that the key presses intended for the group buffer were executed in
the summary buffer and thus triggered completely different actions.

> Telling people "double click to follow some possible cross
> connection" will make them feel at home.

Bye, Reiner.
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo---  |  PGP key available  |  http://rsteib.home.pages.de/