standby mode or wiimote sleeping

This would be easy once ticket:15 was implemented. There could be a command line option to specify the timeout, and when it sees the wiimote wants to connect again wminput makes sure it uses the same config file. I don't think the wiimote has an actual "sleep" mode, so this sounds more like a request for an idle timeout feature.

ditto. Basically, this can be implemented identically to #15, or by detecting the power button, which has not yet been reverse engineered.

ok

if you run wminput in daemon mode using a script you can disconnect wiimote by hand but wminput could turn off wiimote when pass idle time, i dont know if this feature is implemented but i havent found this.

Nope, not implemented, but shouldn't be too hard. We'll add it to the queue.

A posible solution patch. When lost wiimote connectivity it trys again.

Added always trying. -a option.

Apply the patch:

69a70
> 
82a84
> char always = 0;
96a99
>       printf("\t-a, --always\t\tWait always. Don't kill when wiimote disconnect.\n");
98a102,110
> void handle_ctrlc(int sig){
>       signal (SIGINT,SIG_IGN);
>       printf ("Exiting. CTRL+C \n");
>       always=0;
>       signal (SIGINT,SIG_DFL);
>       return;
> }
> 
> 
129c141
<               c = getopt_long (argc, argv, "hwc:v", long_options, &option_index);
---
>               c = getopt_long (argc, argv, "hwac:v", long_options, &option_index);
139a152,154
>               case 'a':
>                       always=1;
>                       break;
159a175
> 
214a231,233
>       signal (SIGINT,handle_ctrlc);
>       do {
> 
295a315,317
> 
>       } while (always==1);
> 

So now you can do: wminput -a -w -c in_ptr [your wii mac]

and you can connect alleays ;)

Now there is another problem, the wiimote always connect to the computer, but sometimes I need to connect to the wii, but it's imposible. So there is another patch to implement wait time between turn off and rediscover:

69a70
> 
82a84,85
> char always = 0;
> int wait_time = 0;
96a100,108
>       printf("\t-a, --always <s>\tWait always but wait N seconds before discover.\n\t\t\t\tDon't kill when wiimote disconnect.\n");
> }
> 
> void handle_ctrlc(int sig){
>       signal (SIGINT,SIG_IGN);
>       printf ("Exiting. CTRL+C \n");
>       always=0;
>       signal (SIGINT,SIG_DFL);
>       return;
98a111
> 
129c142
<               c = getopt_long (argc, argv, "hwc:v", long_options, &option_index);
---
>               c = getopt_long (argc, argv, "hwa:c:v", long_options, &option_index);
139a153,158
>               case 'a':
>                       always=1;
>                       if (optarg) {
>                               wait_time = atoi(optarg);
>                       }
>                       break;
159a179
> 
214a235,237
>       signal (SIGINT,handle_ctrlc);
>       do {
> 
295a319,325
>               if ( wait_time ) {
>                       printf ("wait_time=%d\n",wait_time);
>                       sleep ( wait_time );
>               }
> 
>       } while (always==1);
> 

Thanks for your patches - I should have time to take a look at them this weekend.

I'm not sure how this is supposed to work...from the looks of your patch, SIGINT always kills wminput, but SIGTERM does not. I think SIGTERM and SIGINT should always kill the app, and another signal (SIGUSR1, maybe) should be sent on wiimote error, causing main to loop.

OK, now I capture SIGINT (ctrl+c), SIGTERM (kill) and SIGUSR1 to exit.

This is the complete patch:

17a18,24
>  *
>  *  2007-07-24 Pablo Catalina <xkill@ciberhell.ath.cx>
>  *  * Signaling support to trap SIGINT, SIGTERM and SIGUSR1 that exit the program.
>  *
>  *  2007-07-19 Pablo Catalina <xkill@ciberhell.ath.cx>
>  *  * always listen in wiimote standby mode, with wait time support
>  *
69a77
> 
82a91,92
> char always = 0;
> int wait_time = 0;
96a107
>       printf("\t-a, --always <s>\tWait always but wait N seconds before discover.\n\t\t\t\tDon't kill when wiimote disconnect.\n");
98a110,116
> void handle_exit(int sig){
>         signal (sig,handle_exit);
>       printf ("User exit. CTRL+C or KILL \n");
>         always=0;
> }
> 
> 
129c147
<               c = getopt_long (argc, argv, "hwc:v", long_options, &option_index);
---
>               c = getopt_long (argc, argv, "hwa:c:v", long_options, &option_index);
139a158,163
>               case 'a':
>                       always=1;
>                       if (optarg) {
>                               wait_time = atoi(optarg);
>                       }
>                       break;
159a184
> 
189d213
< 
214a239,243
>       signal (SIGINT,handle_exit);
>       signal (SIGTERM,handle_exit);
>       signal (SIGUSR1,handle_exit);
>       do {
> 
295a325,331
>               if ( wait_time ) {
>                       printf ("wait_time=%d\n",wait_time);
>                       sleep ( wait_time );
>               }
> 
>       } while (always==1);
> 

Sorry, the before post is OK, but with spaggueti code ;) I think that is better to remove the lines added to ignore this signals.

I prefer to catch the signals but without ignore it.

So I remove the lines

[...]
sigset_t sigset;
int signum, ret=0;
[...]
                /* wait */
                sigemptyset(&sigset);
                sigaddset(&sigset, SIGTERM);
                sigaddset(&sigset, SIGINT);
                sigprocmask(SIG_BLOCK, &sigset, NULL);
                sigwait(&sigset, &signum);
[...]

Ticket attachments have been enabled for authenticated users, but patches are sufficient - please don't post another full code listing, as it just makes this page impossible to read. I may remove the full listings from your posts.

Add your username or email address to the CC field to receive change notification.

Have you tested this code? Reread my previous post: SIGINT and SIGTERM should exit. Send a SIGUSR1 should not (directly) exit - send SIGUSR1 from cwiid_callback to cause main to loop if always == true. And put semwait back in - polling is bad form, and as you have it implemented, the main loop will continuously reinitialize itself.

wminput main.c

I implemented a daemon patch that (I think) takes care of what you're going for here. It doesn't implement standby, so it seemed to fit in #15 more than here.

This functionality now exists with various options added to wminput.