Update documentation, mainly with steps to add to wifi

notes.txt

@@ -14,8 +14,9 @@ On "add device"
         557 556 555 554 213 354 262
         557 556 555 554 279 342 256
 
-    looks like it could be an mtu scan, especially since the controllers
-    AP mode issues an MTU of 576
+    This looks like it might be some form of length-encoded data.  The length
+    can be seen without knowing the WPA PSK, so the controller could be using
+    that to configure its wifi credentials.
 
 - connects to TCP/8189
 
@@ -26,6 +27,32 @@ On "add device"
     > 38d90fbd1083
     < 3801fc80ff0200320001ff00000300ff83
 
+The steps to successfully connect the LED controller to your wifi are
+both a somewhat fragile process and not fully spelt out in the existing
+documentation.  I have recorded a process that has worked for me below.
+Also be aware that the WIFI details are not permanently stored in the
+controller, so you will need to repeat the process each time you plug it
+in.
+
+Process to successfully "add device" to your Wifi:
+
+- Disconnect power to your wifi controller for at least 2 minutes
+- If the LEDShop app is running, kill it off
+- Copy your wifi PSK into your clipboard
+- Connect your phone to your wifi
+- Start the LEDShop app and let it complete its "scanning for LAN devices"
+- Select "Add device"
+- Paste your wifi PSK into the "password of wifi" box
+- Connect power to your wifi controller
+- Pause no more than 5 seconds
+- Click OK on the "password of wifi" box
+- wait for no more than 2 minutes (If you can check for new devices on your
+  WIFI, do that and watch for the LED controller to appear)
+- Hit the back button on the LEDShop app
+- "Pull down" the LEDShop list of devices, thus requesting it to do a scan
+  for network devices
+
+
 ---
 Some people on github have some hints:
 https://github.com/Spled/spled.github.io/issues/1

test1.py

@@ -148,7 +148,7 @@ def subc_testpreview(sock, args):
     for i in range(100):
         a = test_frame()
         # TODO - split array up into bits that are MSS rounded down to nearest
-        # 15 byte boundary and try to solve the MTU issue
+        # 15 byte boundary and hope to solve the MTU issue
         txn_sync_expect(sock, a, b'\x31')
 
 

x11-to-led.c

@@ -14,6 +14,7 @@ Currently a proof-of-concept toy
     - Handle serpentine LED layout
     - Allow arbitrary source-image to LED mappings
     - Map screen pixel colors to LED pixels RGB triplets
+    - use XSHMGetImage for efficiency?
 
 - need config options (commandline or config file)
     - hostname / port