Re: follow-link in grep buffer

[Stefan Monnier]

>>> If we have specific problems in certain modes, let's fix those
>>> modes (e.g. in grep that you have to click on the file:line
>>> part of a line to jump).
>> 
>> AFAICT the specific problem at hand is that a mous-1 click used to
>> get focus should not follow a link.  No matter what kind of link, no
>> matter the major mode.

> That's not the only problem.  A mouse-1 click not used to get focus
> should follow a link.  If it doesn't, that is surprising.  Now
> click-to-focus is strictly a windowmanager and frame thingy.  While
> one also uses the mouse for the purpose of changing the _window_
> instead of the frame, one would be terribly surprised if this had not
> the side effect of setting point, too.  Basically, one sets point into
> a different window since that is the convenient way to change focus.

> It is completely backwards if this side effect of a window change
> leads to a link working or not working, namely if we _do_ set point,
> but _don't_ follow the link.  With a true click-to-focus, we should
> not even have the window-point change, merely the selected window.

AFAICT, the problem at hand is about frame-focus-changes, not about
window-focus-changes (Nick's message said "I might just want to select that
window to resize it" suggesting pretty clearly that he's talking about
frames rather than windows).

>> I think x-mouse-click-focus-ignore-position is pretty much the kind
>> of thing we want, although I'd have expected that the window-manager
>> would take care of those things (it should eat the "click-to-focus"
>> and not pass it on to the application).

> You are confusing frame and window.  Changing the selected window is
> also necessary at times.

No, I'm not confusing them, I was talking about frames (and never mentioned
the word "window"), tho I didn't make it explicit since I thought the
context was sufficiently clear.

BTW, in a click-to-focus environment (I'm still talking only about frames's
focus), I think it'd indeed also make sense (as Nick's message mentioned) to
ignore a click on a sensitive mode-line element if it was used to get focus.

Basically, any non-trivially-safe (point movement and window selection is
safe, most of the rest isn't) operation should not be triggered by
a click-to-focus.

Maybe the mouse-face should also be inactive on frames that don't
have focus.


        Stefan