[openbox] [ANNOUNCE] Openbox-3.0-beta6 released

[Mike]

On Friday, September 26, 2003 at 01:59PM, Ben Jansens wrote:
>On Fri, Sep 26, 2003 at 12:52:12PM -0500, Mike wrote:
>> Ah I see you included the old transient focus behavior stuffs. I've been 
>> trying to get focus to work more like old openbox2/blackbox, and that 
>> helps. :)
>
>Yep that was simply an oversight on my part. I think that is the correct
>behavior.

Well I'm not sure if there's any docs that define this type of thing, so correct or not, I like it. :)

>
>> I have a question about focus_fallback() though.
>> 
>>    if (!config_focus_last && config_focus_follow)
>>        if (focus_under_pointer())
>>            return;
>
>If you are using sloppy focus AND you are NOT using focusLast, then it will
>try to focus under the pointer. If it fails, it will continue as if
>focusLast was on.
>
>> That doesn't seem right to me, maybe I'm just confused or not understanding 
>> the bigger picture.
>> 
>> Should it look more like this?
>> 
>>    if (!config_focus_last || (config_focus_follow && focus_under_pointer())
>>        return;
>
>With this, if you turn off focusLast you will end up focusing nothing. If
>you meant to make it not focus anything if the pointer was not on a window,
>that is not behavior I wish to create. Focus should alawys be somewhere
>usuable if possible.

Ah yeah I see it, my bad.

>> I modify this bit anyway so transients are focused before 
>> focus_under_pointer(). Again it goes back to stuff like Mozilla's find in 
>> page dialog. Although I'm not sure if this is how openbox2 worked or not, 
>> but I like it anyway. :)
>
>Hm, that may be a good idea.

Really the only thing I use this for is if you do a find in page, nothing comes up, you can hit escape and the find dialog stays focused. Otherwise Mozilla's main window is focused (if that's where the pointer is).

I could probably get used to the way it is right now, but it just feels strange.

>
>> Before 3.0 final comes out, it'd be nice to have configurable window 
>> placement. There were also a few posts about popups for window/workspace 
>> cycling. Making that configurable is pretty simple, maybe something to 
>> consider.
>
>Popups can be disabled again in CVS already. And I'm still mulling over
>window placement :)

I know the stuff in my patch isn't exactly beautiful, but it's simple and works.

If there were numerous window placement methods, assigning a focus function when the config is read might work ok. place_transient() || place_dialog() || place->type() || place_random(). But that's not the case. Since there's only 2 and it doesn't sound like there's much interest in focus follows mouse anymore, I dunno.

Thanks,

Mike