FreeEMS-Tuner Development Diary - Comments

[Fred]

Make it work, Make it right, then make it fast. :-)

[Fred]

Sounds like you are working on the make it right part. I approve, sounds good. You have to love code that gets simpler and does the same thing better. Keep up the good work :-)

Fred.

[sry_not4sale]

Cheers mate

Since I am making some big changes, it makes sense to tidy things up as I see them too Adding lots more comments as I go through stuff.

[Fred]

I look forward to testing the first snapshot against the firmware code :-)

Partially functioning stuff will be happily accepted!

Fred.

[sry_not4sale]

I've only just got it to bring up the gui again, so give me a couple more days Will definately supply semi-functional stuff for people to look over.

[Fred]

Screw looking it over, I wanna run it and use it. I should hook you up with the latest datalog packets.

Code: Select all

2008-12-19 12:04:40,257 DEBUG    comms.Serial - Packet received by Serial connection: ['0xAA', '0x10', '0x01', '0x2D', '0x00', '0x60', '0x73', '0x43', '0xFF', '0xFF', '0xFA', '0x00', '0xC0', '0x00', '0x65', '0x90', '0x29', '0xFE', '0x5F', '0xB4', '0x50', '0x41', '0x62', '0xE8', '0x1E', '0xA5', '0x04', '0xCB', '0x00', '0x00', '0x00', '0x00', '0x28', '0x6A', '0x00', '0x00', '0x00', '0x00', '0x65', '0x90', '0xC0', '0x00', '0x5F', '0xC8', '0x61', '0x80', '0x3C', '0xA1', '0x64', '0xF2', '0x01', '0xF4', '0x00', '0x00', '0x00', '0x00', '0x64', '0xF2', '0x66', '0xE6', '0x00', '0x00', '0x00', '0x00', '0x00', '0x00', '0x00', '0x00', '0x00', '0x00', '0x01', '0xDE', '0x00', '0x00', '0x03', '0xFF', '0x03', '0xFF', '0x03', '0xFF', '0x03', '0xFF', '0x03', '0xFF', '0x03', '0xFF', '0x01', '0x17', '0x01', '0x1E', '0x01', '0x16', '0x01', '0x1C', '0x01', '0x27', '0x01', '0x1F', '0x01', '0x1D', '0x01', '0x19', '0x97', '0xCC']

Consists of :

start byte
header byte
0x012D payload ID
0x0060 payload length
0x60 bytes of variables = 48 16 bit numbers (core vars, derived vars, ADCs)
checksum byte
stop byte

Different eg packet payload only :

Code: Select all

0000:  65 02 64 ab 9d 3a 92 54  2e 54 1a e3 17 c3 72 77   e.d..:.T.T....rw
0010:  5e c7 1b 90 03 72 00 00  00 00 00 00 00 00 00 00   ^....r..........
0020:  2e 54 c0 00 6b f2 2c 79  3b e1 2e 9f 0e 02 00 00   .T..k.,y;.......
0030:  00 00 2e 9f 3c a1 00 00  00 00 00 00 00 00 00 00   ....<...........
0040:  03 62 03 69 02 41 02 92  00 fe 01 c1 02 58 01 f4   .b.i.A.......X..
0050:  00 f6 00 fe 00 f7 00 e7  00 f7 00 fa 00 f9 00 f4   ................

The two var blocks end in zeros because they are unused at this time.

I will probably add some time stamping and sequencing to that before too long so expect to grab a few new vars out from in front of the big block.

Fred.

[Fred]

Is there any monitoring facility for the length of these low priority queues? I'm worried they may back up and we won't know how long they are :-)

Keep up the good work.

Fred.

[sry_not4sale]

Umm monitoring would probably be watching the log file... I have found it difficult adding things to the controller loop as I have to hack around the normal method of adding an action, cos that would just get appended to the queue. I might add some private methods specifically for doing things in the controller thread that run straight away... and are very fast so they don't slow it down.

I am wondering if maybe there was some python package to do all this already - it seems like I am writing my own python kernel...! The run loop is only around 25 lines, and doesn't run many (or any large) methods so I am confident it should be fairly quick and bug free - but it still has the kernel feeling lol

[Fred]

If the log file is low priority, watching it to see how far behind the 8 ball IT is won't help much.

I'm just being paranoid about speed. Hopefully it can parse a full stream at or less than 50% cpu usage, if that's the case, no worries.

Right now I send a packet and get a response and don't see the response for minutes because of the back log of datalog packets.

I've just hacked in a delay between packets and datalog on/off back end to the orange button so it's not an issue right now.

Excuse my paranoia :-)

Fred.

[Fred]

As per our conversation last night/yesterday, I make it about 40 hours to go :-) xmas is coming late this year LOL, and I'll probably be away from the test rig at the time, but I'll fire it up and check it out anyway. You can go back on your comment if you want, it's summer there, enjoy it! I know I would be!

Fred.