Re: [PATCH] consequence: hook

[Thorsten Wißmann <edu _at_ thorsten _minus_ wissmann _dot_ de>]

Hi,

On Mon, Dec 24, 2012 at 09:14:35PM -0800, Tyler Thomas Hart wrote:
> On Mon, Dec 24, 2012 at 04:18:18PM +0100, Thorsten Wi?mann wrote:
> > The main thing is, that the rule should not directly emit the specified
> > hook but a hook called "rule" with a customizable parameter.
> 
> Agreed. I included a colon eg. 'rule:', feel free to change it back as
> you see fit.

I changed it back. Each hook (like a command) is an array of strings,
e.g:

    fullscreen  on  0x1000009
    rule        myfancyrulename
    rule        a_wild_firefox_appears

so one can listen to it e.g. using: herbstclient -i fullscreen. Waiting
for the next firefox: herbstclient --wait rule a_wild_firefox_appears
But is *not* just one line, so there must not be an extra separator,
especially no colon.

> > So the first possible solution is:
> > 
> >   char* hook_str[] = { "rule", cons->value.str };
> >   hook_emit(LENGTH(hook_str), hook_str);
> > 
> > And the second possibility is:
> > 
> >   hook_emit_list("rule", cons->value.str, NULL);
> 
> I went with the first choice. As with my suggestion on irc, avoiding
> the call to g_new and the va loop in emit_list() seems more efficient.
> 
> I've attached the amended patch. If you want to make any other
> changes, feel free to make them.

OK. But efficiency of the code at this point is not the bottleneck when doing
X-communication.

Your patch now is upstream as:

40fb7a3 Add consequence: hook

Regards,
Thorsten