June 2023 - OvmsDev - Open Vehicles

Internet outage
by Mark Webb-Johnson 21 Jun '23

21 Jun '23

We had a nasty internet outage (fibre line break) this afternoon that caused substantial downtime for api.openvehicles.com. Apologies for this. All is back now, and we will be trying to make some changes in the near future, to make it easier to fall over to the backup link. Regards, Mark

1 0

Nominal battery pack voltage
by Solterra 21 Jun '23

21 Jun '23

Howdy, Is there a standard metric for nominal battery pack voltage? I'm looking to use it to convert v.b.cac to kWh in a rewrite of the ABRP plug-in. Thank you

1 0

Recommen
by Solterra 20 Jun '23

20 Jun '23

Hello, I have a couple of questions on the CAN poll TX framework. Is there a recommended method for adding a one-time CAN poll to the queue with the response to be processed by IncomingPollReply? What is the recommended method for adding or modifying obdii_polls after initialization? Thank you

3 5

Using the override keyword
by Michael Geddes 17 Jun '23

17 Jun '23

Hi all, So at my work, the main language I use is Delphi (so Object Pascal), and one of the things about Delphi is that you are forced to always be explicit about overriding virtual methods. (Missing out on 'override' directive indicates that you are Hiding (a warning) rather than overriding the virtual method). Since C++ now allows use of the override keyword now to explicitly say a method is overriding a virtual function, I have been adding that in all the code I use (or at least trying to). Firstly, it is a visual indicator that the method is, in fact, overriding a virtual function (which is not always obvious), but it also means that if a virtual method changes its signature in any way, you will get an error if an 'override' marked method no longer matches! I'm not sure what I'm after except just to bring it to people's attention in case they might have missed it, and find out how people view this. My latest pull-request has changed the signature of OvmsVehicle methods and I have added 'override' to all the places where this is overridden in derived Vehicle classes. //.ichael

1 0

Error: can: can1 watchdog inactivity timeout - resetting bus
by Solterra 12 Jun '23

12 Jun '23

Hello, I'm getting this error. However, I don't think I'm using can1... My vehicle uses can2. I don't think I've initialized can1. Any clues to help me debug? Do I need to do something to actively disable can1 during vehicle initialization? Thank you

1 1

Subaru Solterra
by Solterra 11 Jun '23

11 Jun '23

Any interest in adding support for Subaru Solterra?

4 9

CI ?
by Ludovic LANGE 08 Jun '23

08 Jun '23

Dear list, I've been playing a little with GitHub's automatic builds / CI for a few months, and I'm thinking about having it in master. The advantages I can think of would be: * Ability to "catch" early on any incompatibility we could introduce in a PR. Still in a transition period between ESP-IDFv3.3 - the official builds, and ESP-IDFv5.x (the future, I hope, once all the issues are ironed out) ; it can add some value in trying to build the firmware for different combinations, and thus with different compiler versions / libraries versions. * It could also be a backup build system for the nightly / releases firmwares. Hosting, compiling there is free of charge for public GitHub projects Possible concerns: * Very tied to GitHub's hosting and API - vendor lock-in However a) it's completely isolated in its own directory, and b) some of GitHub's competitors have a similar mechanism (GitLab, ...) In my tests, I'm using it to: * Ensure that the documentation always compile without any warning. Sometime, we have very small issues that can impact the visual rendering of the HTML, and this step can trigger when the compiler emits a warning (or error) * Ensure that the changes I'm doing are building without error for a "matrix" of different environments: o target: ESP32 (but could also be other processor of the Espressif portfolio, as supported by the build chains: -S2, -S3, -C2, -C3, -C6, -H2 ...) o ESP-IDF version : v3.3.4 (our current), v4.4 branch, v5.0, v5.0, v5.0.1, v5.0.2, v5.0 branch (could also use v5.1 branch and latest, but compile errors from time to time) o Mongoose version: 7.9, 7.10, 6.11 (our current) If you want to have a look, please check: * Some of the automated builds here https://github.com/llange/Open-Vehicle-Monitoring-System-3/actions * A set of corresponding build recipes here https://github.com/llange/Open-Vehicle-Monitoring-System-3/tree/build-maste… * The output of a random build: o The log of commands: https://github.com/llange/Open-Vehicle-Monitoring-System-3/actions/runs/519… o The build artefacts: https://github.com/llange/Open-Vehicle-Monitoring-System-3/suites/134194909… (NB: Do not flash ! Just an example :-)) I'd be happy to have some feedback about the interest (or not) to have this in master. For that the next plans would be: * Step1 : a PR (to be written) including all these config files in <root>/.github/workflows * Step2 : a configuration of our GitHub repo to : o activate the CI actions o After some time, mark some of these actions as "required" in PRs (Can be overriden of course, but will show the creator of the PR how to improve the PR and have it work across all the combinations of environments we aim to support) Thanks for your time. Regards,

2 1

Proposed changes to IncomingPollReply and IncomingPollError - before making a pull request
by Michael Geddes 02 Jun '23

02 Jun '23

Hi all, So I have some changes coming that depend heavily on this function signature change. What it is doing is taking the members of OvmsVehicle that are specific to the ISOTP (and VWTP) protocols and making the parameters to both IncomingPollReply and IncomingPollError. This way of doing it makes all the bits explicit parameters, which means a bunch more parameters passed on the stack. *Original*: - virtual void IncomingPollReply(canbus* bus, uint16_t type, uint16_t pid, const uint8_t* data, uint8_t length, uint16_t mlremain); - virtual void IncomingPollError(canbus* bus, uint16_t type, uint16_t pid, uint16_t code); *New:* virtual void IncomingPollReply(canbus* bus, uint32_t moduleidsent, uint32_t moduleid, uint16_t type, uint16_t pid, const uint8_t* data, uint16_t mloffset, uint8_t length, uint16_t mlremain, uint16_t mlframe, const OvmsPoller::poll_pid_t &pollentry); virtual void IncomingPollError(canbus* bus, uint32_t moduleidsent, uint32_t moduleid, uint16_t type, uint16_t pid, uint16_t code, const OvmsPoller::poll_pid_t &pollentry); The other way of achieving the same result would be to create a struct that contains these elements, and then to pass that as a const reference. Thoughts? If I need to make modifications to the signatures, I'd prefer to do it before I make the pull request. //.ichael

2 2