Re: [RP] How to put windows into groups by application type.

[tomm smith]

yeah I was getting the same issues with multiple runs(continuous runs rather).

I attempted to fiddle with "e->height <= 10 && e->width <= 10" it
might be a valid solution, but it seems liferea and a few others ran
the hook < then 1 times. While I was fiddling and surfing source files
to obtain info, I had a thought of calling the hook from when the
window name is inserted into the window list. I ran on this epiphany
and tried adding the hook call to insert_into_list() in window.c well
sure enough all the tests I have done thus far have returned exactly
one execution of the hook per new window.

I have attached the diff of the changes, I want to test it more but I
feel pretty confident with this solution.

(P.S. If the diffs I am submitting are not quite how you like patches
or I am generating them wrong please do tell me, I am pretty new to
patch writing for C so sketchy on my methods.)

On 3/15/10, Eric A <address@hidden> wrote:
> On Sun, Mar 14, 2010 at 5:09 AM, tomm smith <address@hidden> wrote:
>> I believe I have created a patch for this desired hook. Only had to
>> change a few lines, basically define the hook and add a hook call to
>> new_window(). I am a newborn in the C world so if any of this is
>> hackery that does not comply with GNU coding standards I do apologize.
>> I have attached the patch to this email. I tested out the code very
>> shortly I plan to test it more, but initial test showed everything was
>> working smoothly.
>>
>
> Thanks, I think this is what I'm looking for! I tried it out, but ran
> into a problem. It seems that for most programs,
> new_window(XCreateWindowEvent *e) is called multiple times. For
> example, starting midori issues 2 events and firefox 3 events. In
> fact, the only program that I've tested that calls it just once is
> urxvt.
>
> I tried filtering out the multiple events by looking at the window
> properties. It seems that all the additional events have window
> height/widths (e->height and e->width) of 1, 2, or 10. So by filtering
> all windows with height/widths of <=10, I was able to produce only one
> event per "real window".
>
> I'm not sure whether this would work 100% of the time though. Do you
> have any better ideas?
>
> Thanks,
> Eric
>

rat-crtHook.diff
Description: Text Data