I used to use the usbremote plugin, now I switched to lirc, because I wanted to play with irexec. Since then vdr didn´t react to my remote control as I was used to any more. I had sometimes double reactions on one keypress as other people have described and the speed of the reactions was weird. After investigating I found following solution, which makes me 100% happy. Maybe other people like it too?
Btw. the count´s provided by lirc are not reliably incoming to the lirc part of vdr, meaning if I press a key for a long time instead of going up continously, they start from time to time at zero, sometimes even all are zero. That is not catched by this patch.
Some more background information: I compared the remote plugin for the FF cards with the usbremote plugin for the igorplug (which was taken from the remote plugin) and the lirc part of vdr. My remote (OFA URC 7541) with RC5 codes on each key toogles the RC5 code on each key press. (I suppose all RC5 remotes do(?)). The firmware of the driver for the FF cards changes that to key down/key up (AFAIR). The usbremote plugin receives the codes as they are sent. Lirc changes the toogled codes into counts starting from zero. With the igorplug the counts sometimes start at zero again, with a serial receiver it does always count up. The lirc part of vdr however has a problem here. I observed that sometimes all counts are zero with the igorplug, with the serial driver this is much better, but happens sometimes as well.The remote plugin and the usbremote plugin work very well, the lirc-vdr didn´t for me. I don´t know how the RCU processes the RC5 code, and hence don´t understand the rcu code. The rcu code was taken for lirc in vdr (some adaptations were made in time). I put a lot of printf´s to see how the code logic operates and what actually happens on key presses. The improved code, which is basically just taken from the remote plugin, works perfectly for me.
Jörg
--- lirc.c.orig.c 2006-01-27 16:59:47.000000000 +0100 +++ lirc.c.new.c 2006-05-26 22:40:26.000000000 +0200 @@ -13,9 +13,9 @@ #include <netinet/in.h> #include <sys/socket.h>
-#define REPEATLIMIT 20 // ms #define REPEATDELAY 350 // ms -#define KEYPRESSDELAY 150 // ms +#define REPEATFREQ 100 // ms +#define REPEATTIMEOUT 500 // ms #define RECONNECTDELAY 3000 // ms
cLircRemote::cLircRemote(const char *DeviceName) @@ -94,8 +94,8 @@ continue; } if (count == 0) { - if (strcmp(KeyName, LastKeyName) == 0 && FirstTime.Elapsed() < KEYPRESSDELAY) + if (strcmp(KeyName, LastKeyName) == 0 && FirstTime.Elapsed() < REPEATDELAY) continue; // skip keys coming in too fast if (repeat) Put(LastKeyName, false, true); strcpy(LastKeyName, KeyName); @@ -104,8 +104,10 @@ timeout = -1; } else { - if (FirstTime.Elapsed() < REPEATDELAY) - continue; // repeat function kicks in after a short delay + if (LastTime.Elapsed() < REPEATFREQ) + continue; // repeat function kicks in after a short delay (after last key instead of first key) + if (FirstTime.Elapsed() < REPEATDELAY) + continue; // skip keys coming in too fast (for count != 0 as well) repeat = true; timeout = REPEATDELAY; } @@ -113,7 +115,7 @@ Put(KeyName, repeat); } else if (repeat) { // the last one was a repeat, so let's generate a release - if (LastTime.Elapsed() >= REPEATDELAY) { + if (LastTime.Elapsed() >= REPEATTIMEOUT) { Put(LastKeyName, false, true); repeat = false; *LastKeyName = 0;
On Sat, 2006-05-27 at 19:27 +0200, Joerg Riechardt wrote:
I used to use the usbremote plugin, now I switched to lirc, because I wanted to play with irexec. Since then vdr didn´t react to my remote control as I was used to any more. I had sometimes double reactions on one keypress as other people have described and the speed of the reactions was weird. After investigating I found following solution, which makes me 100% happy. Maybe other people like it too?
I like it. The patch you posted had some whitespace/linewrap issues though, which are hopefully fixed by the attached one (no other changes in it).
I demand that Ville Skyttä may or may not have written...
--- vdr-1.4.0/lirc.c~ 2006-01-27 17:59:47.000000000 +0200 +++ vdr-1.4.0/lirc.c 2006-05-27 22:04:19.000000000 +0300 @@ -13,9 +13,9 @@ #include <netinet/in.h> #include <sys/socket.h>
-#define REPEATLIMIT 20 // ms #define REPEATDELAY 350 // ms -#define KEYPRESSDELAY 150 // ms +#define REPEATFREQ 100 // ms
No. You want at least 114 (for RC5).
[snip]
Darren Salt schrieb:
I demand that Ville Skyttä may or may not have written...
--- vdr-1.4.0/lirc.c~ 2006-01-27 17:59:47.000000000 +0200 +++ vdr-1.4.0/lirc.c 2006-05-27 22:04:19.000000000 +0300 @@ -13,9 +13,9 @@ #include <netinet/in.h> #include <sys/socket.h>
-#define REPEATLIMIT 20 // ms #define REPEATDELAY 350 // ms -#define KEYPRESSDELAY 150 // ms +#define REPEATFREQ 100 // ms
No. You want at least 114 (for RC5).
[snip]
Do you mean REPEATFREQ? It works well here with RC5 and 100. Doesn´t it work for you?? Jörg
I demand that Joerg Riechardt may or may not have written...
Darren Salt schrieb:
I demand that Ville Skyttä may or may not have written...
[snip]
-#define REPEATLIMIT 20 // ms #define REPEATDELAY 350 // ms -#define KEYPRESSDELAY 150 // ms +#define REPEATFREQ 100 // ms
No. You want at least 114 (for RC5). [snip]
Do you mean REPEATFREQ? It works well here with RC5 and 100.
(Actually, it's also misnamed: that's a repeat interval.)
The RC5 repeat interval is (IIRC) 113.4ms. Any soft repeat handler needs to have an interval which is no less than this else sometimes you'll be unable to avoid an extra repeat.
Doesn't it work for you??
Not checked; I'm using LIRC via gxine and irexec.
Darren Salt schrieb:
I demand that Joerg Riechardt may or may not have written...
Darren Salt schrieb:
I demand that Ville Skyttä may or may not have written...
[snip]
-#define REPEATLIMIT 20 // ms #define REPEATDELAY 350 // ms -#define KEYPRESSDELAY 150 // ms +#define REPEATFREQ 100 // ms
No. You want at least 114 (for RC5). [snip]
Do you mean REPEATFREQ? It works well here with RC5 and 100.
(Actually, it's also misnamed: that's a repeat interval.)
The RC5 repeat interval is (IIRC) 113.4ms. Any soft repeat handler needs to have an interval which is no less than this else sometimes you'll be unable to avoid an extra repeat.
If you look at how the code works, you will find that that´s not true. And the repeat interval (incoming keys) depends on the lirc driver. With different drivers there are different intervals for the same remote. And I am not writing from theory, i checked it. Jörg
Doesn't it work for you??
Not checked; I'm using LIRC via gxine and irexec.
-
Darren Salt -
Joerg Riechardt -
Ville Skyttä