Does not set all keys at once

[ghost]

When using -an FFFFFF, some keys will stay the previous color; running the command multiple times will solve the problem, and it always happens in a VERY predictable way.

Using US layout keyboard on Arch.

[MatMoul]

-an is breaked and seem not possible with the current code... At this time, I remove the no commit feature, but the code stay here for futur work. I just have mapped -an to -a, -gn to -g and -kn to -k and removed commented text in -h.

[ghost]

It happens regardless of an or a. Any logs I can get you that will help?

On Aug 9, 2016 2:44 PM, "matmoul" notifications@github.com wrote:

-an is breaked and seem not possible with the current code... At this time, I remove the no commit feature, but the code stay here for futur work. I just have mapped -an to -a, -gn to -g and -kn to -k and removed commented text in -h.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/MatMoul/g810-led/issues/4#issuecomment-238668557, or mute the thread https://github.com/notifications/unsubscribe-auth/AO_naoircrHkPEKhToEFjKdwM3yILorVks5qeNizgaJpZM4Jfop9 .

[MatMoul]

It's very hard, I can not reproduce. how is it going? it's always the same keys that problem? or more random?

[ghost]

It's a general pattern, but I was mistaken that it's perfectly predictable. About 75% of the keys will set. Run it again, 75% if the remaining keys set. If I run the command repeatedly, it will eventually set them all. One time took around 20 tries though.

On Aug 9, 2016 2:55 PM, "matmoul" notifications@github.com wrote:

It's very hard, I can not reproduce. how is it going? it's always the same keys that problem? or more random?

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/MatMoul/g810-led/issues/4#issuecomment-238671323, or mute the thread https://github.com/notifications/unsubscribe-auth/AO_nagpBAMwNSq10i8dzuaGnCnY6rrQYks5qeNslgaJpZM4Jfop9 .

[MatMoul]

With reference to the issue #2, what's your CPU?

[ghost]

Off hand, it's some mobile version of the i5. I'm not sure the exact clock speed.

On Aug 9, 2016 3:16 PM, "matmoul" notifications@github.com wrote:

With reference to the issue #2 https://github.com/MatMoul/g810-led/issues/2, what's your CPU?

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/MatMoul/g810-led/issues/4#issuecomment-238677136, or mute the thread https://github.com/notifications/unsubscribe-auth/AO_navPflSYBrB2IwH9rZ9BnfssqICrUks5qeOA1gaJpZM4Jfop9 .

[MatMoul]

I need test on a best PC as my old core 2 duo... Can you test by removing the -O2 in makefile at line 2 : CFLAGS=-Wall -std=gnu++11

[ghost]

Doesn't seem to have helped at all, unfortunately. Same rough patterns. Any commands I should run?

[MatMoul]

Thank you for your tests.

I will test the next week on my work PC (a xeon cpu and a new G810 wait me)

[andrey-str]

Confirm on my laptop(CPU is i7-6820HQ)

[johandc]

I can confirm this bug on i7-6600U, Linux 4.4.0-31-generic #50-Ubuntu x86_64. I have added the following to main.cpp that fixes the issue (when setting profiles):

diff --git a/src/main.cpp b/src/main.cpp
index 04d69a8..ac5eef9 100644
--- a/src/main.cpp
+++ b/src/main.cpp
@@ -5,6 +5,7 @@
 #include <algorithm>
 #include <fstream>
 #include <string>
+#include <unistd.h>
 #include "classes/Keyboard.h"

 using namespace std;
@@ -139,7 +140,7 @@ int loadProfile(string profileFile) {

     while (!file.eof()) {
       getline(file, line);
-      
+      usleep(100);
       if (line.substr(0, 3) == "var") {
         line = line.substr(4);
         ind = line.find(" ");

[MatMoul]

I just tested on my workstation and I also have the problem. I fixed the code with "usleep" in Keyboard.cpp I also increased the variable "maxKeyValueCount" to 12. Now, it works on my Xeon CPU. Without negative feedback from you, I will close the tiquet

[MatMoul]

following return on #2, I close the ticket

[MatMoul]

And thank you to johandc for the usleep

[MatMoul]

I just reactivated no commit feature : -an -gn -kn -c All work very well !