[Ovmsdev] v2.5.3

Mark Webb-Johnson mark at webb-johnson.net
Wed Sep 18 20:12:12 HKT 2013

	Ockham's Razor: "Entia non sunt multiplicanda praeter necessitatem"
	(or paraphrased and applied to science: "when you have two competing theories that make exactly the same predictions, the simpler one is the better")

I've been living by that saying for most of my professional career (dating back to my work on an occam compiler back in the '80s).

For the past couple of weeks, I've been fighting some data corruption bugs in the ACC/LOGGING code. The symptom was that the net_state got set to something impossible, and the GPRS connection hung until the box restarted (at which point the net_state corruption would happen again). This, coupled with lousy cellular reception at my home, has been driving me crazy.

Last night, I found the cause. The gory details are immortalised in github commit 80e275b, and an example of why Ockham's Razor works so well. Rather than concentrating on the simple (why was net_state getting corrupted), I searched wildly for the complex (buffer overflows, stack overflows, etc). It turned out that I had mis-coded acc_state_enter() for net_state_enter() in the ACC module. Doh!

Anyway, we're back on track. This evening, I setup ACC in my car to do a cool down, then charge to 90%, and it seemed to work just fine - nice logging and the cooldown went well.

All the code is committed. I've got a few little changes to make, and then I'll build binaries for v2.5.3. I should be sending out an eMail later tonight with some documentation on how this works, and some requests for wider testing (amongst Roadster owners for the ACC, and for everyone with Logging).

Regards, Mark.

Begin forwarded message:

> From: Mark Webb-Johnson <mark at openvehicles.com>
> Subject: [Ovmsdev] v2.5.3
> Date: 10 September, 2013 9:25:34 AM HKT
> To: OVMS Developers <ovmsdev at lists.teslaclub.hk>
> Reply-To: OVMS Developers <ovmsdev at lists.teslaclub.hk>
> A few days ago, I committed a bunch of code past v2.5.2, labelled v2.5.3. I've been testing this in my car, but the results are not good. I'm getting crashes, and that is causing the cooldown (when configured in ACC mode) to restart. For the moment, I suggest people don't try past 7a6ecc4 (tag: v2.5.2). I'm working on getting the bugs out, but the test cycles are taking me 24-48 hours each, so this is taking some time. I'll let you know as soon as it is stable enough for wider testing.
> A bit of better news: I finally found a minor typo (CAN ID 0x100 B1=0x8F not B1=0x87) that has been driving me up the wall for the past few weeks. HVAC monitoring during cooldown now seems to be working well, and I'm getting better results for cooldown. At this rate, I may just have cooldown working properly in time to work on heatup for the winter :-(
> Regards, Mark.
> _______________________________________________
> OvmsDev mailing list
> OvmsDev at lists.teslaclub.hk
> http://lists.teslaclub.hk/mailman/listinfo/ovmsdev

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openvehicles.com/pipermail/ovmsdev/attachments/20130918/af7e0171/attachment.htm>

More information about the OvmsDev mailing list