commit 4e0c5aa5ed3d791404fd134e8c6854fdcfc67693 Author: Janusz Krzysztofik Date: Tue Dec 7 15:30:15 2010 +0100 Heyu 2.9.2 Signed-off-by: Janusz Krzysztofik commit 5141a74d1b6b4d0e81e338b70a39cc2e69835406 Author: Janusz Krzysztofik Date: Tue Nov 30 15:02:17 2010 +0100 Re: Oregon Scientific THGR328N An attempt to fix the Heyu handling of the Oregon Scientific THGR328N sensor model (supported at an Alpha level so far), based on test data provided by John O'Hare. Reported-by: John O'Hare Tested-by: John O'Hare Signed-off-by: Janusz Krzysztofik commit 6352911c2026b47a46aea7ac98376accc8a8612f Author: Janusz Krzysztofik Date: Mon Nov 22 13:05:56 2010 +0100 Fix selected Heyu commands waiting forever on Darwin It occured that on a Darwin based system, the alarm() system call, used in the Heyu code for timed reads from the spool file, among others, may interact badly with nanosleep() system calls, intended for limiting system load while waiting for an ALARM signal. This resulted in selected Heyu commands waiting forever for a missed ALARM on affected platforms. OTOH, there were no observations of excesive system load if theese nanosleep()'s were omitted on a Darwin. Don't include nanosleep() based millisleep() helper function calls into the procedure of reading from the spool file if building for a Darwin system. Created against heyu-2.9.1. Reported by: Henry Butz Signed-off-by: Janusz Krzysztofik commit 94743f92f368d012752bd6b68586c4e2d467a107 Author: Janusz Krzysztofik Date: Fri Nov 19 20:06:29 2010 +0100 X10_Timer should return a number rather than name --- In heyu_users@yahoogroups.com, itaide@... wrote: > > When a timer expires a script can be launched and the environment > populated with amongst other things X10_Timer. > The way I read the 'man x10scripts' leads me to believe that this > should contain a numeric value of the timer that expired. > > > 'X10_Timer' (= timer number 1-16 which launched script) > > However, it contains the string 'timer' followed by the number of the > timer. Replace those default "timerNN" names with just "NN", for consistency with what settimer accepts. Created and tested against heyu-2.9.1. Reported-by: John O'Hare Signed-off-by: Janusz Krzysztofik commit 148746d09a3e73ea66501baea9a8dfb1eec56535 Author: Janusz Krzysztofik Date: Fri Nov 19 15:06:37 2010 +0100 Fix Oregon vflags handling I found the oregon tmin:/tmax: state flags not working as expected. After examining the related code, I found the reason was vflags being destroyed by a subsequent signal carrying a different parameter value, humidity in my case. The same applies to all oregon sensor models that report more than one paramter. The patch below corrects the problem for me for both real and emulated oregon sensor modules, hopefully not breaking those not affected. Created and tested against heyu-2.9.1. Signed-off-by: Janusz Krzysztofik commit 53198dc74667f4dffe47c34f17e865e9b109fe15 Author: Janusz Krzysztofik Date: Fri Nov 19 14:29:55 2010 +0100 Fix tmax:/nottmax: state flags not recognized Both tmax:Hu and nottmax:Hu state flags were not recognized correctly due to a trivial copy-and-paste bug. The patch below corrects the problem for me. Created and tested against heyu-2.9.1. Signed-off-by: Janusz Krzysztofik commit 2338c36aca8826c5b151f915b2d5cdc6ba4174aa Author: Janusz Krzysztofik Date: Fri Nov 19 14:11:06 2010 +0100 Fix counters incremented/decremented twice Tuesday 12 October 2010 13:44:28 John O'Hare wrote: > > I think there may be a couple of bugs around the counter & inccount > commands: The problem was a counter being incremented/decremented a second time when reported to the log file. This bug affected command line usage only, internal commands were still working as expected. The patch below corrects the problem for me. Reported-by: John O'Hare Signed-off-by: Janusz Krzysztofik commit 8d67d17b2605d801812c1ffc7760d9a6820fd67d Author: Janusz Krzysztofik Date: Fri Nov 19 02:35:17 2010 +0100 Avoid launching a script if wrong script name specified If a user happens to put a "LAUNCHER name ..." clause in his x10config without providing a corresponding "SCRIPT -l name ..." definition, a first script, in the x10config file order, is launched on every signal that matches launch conditions of the inconsistent LAUNCHER. This patch fixes the problem by preventing heyu from starting up if such error in the x10config file is detected. Created and tested against heyu-2.9.1. Signed-off-by: Janusz Krzysztofik commit 3b3f53f07cb9f23d791b083265c6b746a5e21433 Author: Janusz Krzysztofik Date: Fri Nov 19 02:11:13 2010 +0100 Heyu counter state command fix This patch corrects a minor problem of not being able to examine a counter value with the 'heyu counter NN' state command for NN greater than 16. Created and tested against heyu-2.9.1. Signed-off-by: Janusz Krzysztofik