<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
</head>
<body>
<div class="moz-cite-prefix">Hi Michael,</div>
<div class="moz-cite-prefix"><br>
</div>
<div class="moz-cite-prefix">So if you don't mind, I'll be creating
GitHub "issues" for following those improvements, in order to
describe, plan, document, follow the progress on those ideas.</div>
<div class="moz-cite-prefix"><br>
</div>
<div class="moz-cite-prefix">Regarding WiFi, the idea would be more
like the "mesh" mode ; but with different SSIDs.</div>
<div class="moz-cite-prefix">Say you have a list of different SSIDs
/ password. The device should try to connect to the first one in
the config. If unable to connect, or if the RSSI steadily falls
below a certain minimum, it will try the next in order in the
config. Then when it reaches the end of the list, it cycles back
to the start. Or optionally it should try to connect via modem
connection (but I'd make this a config option also) - and if this
one drops also, back to wifi APs list.<br>
</div>
<div class="moz-cite-prefix">(RSSI level should be a config item
also).</div>
<div class="moz-cite-prefix"><br>
</div>
<div class="moz-cite-prefix">Use-case is the following:</div>
<div class="moz-cite-prefix">I have a vehicle under test ; it has a
OVMS on-board. The test pilot does some different journeys. Some
starting from "home", ending at "shop A". Then from "shop A" to
"shop B". etc... and back to "home".<br>
</div>
<div class="moz-cite-prefix">It's not possible (and not wanted) to
have all these locations set up with the same SSID / Password. So
it could be interesting to have OVMS able to connect to each
different AP.</div>
<div class="moz-cite-prefix">(Automatically - the test pilot is not
expected to deal with OVMS configuration)</div>
<div class="moz-cite-prefix"><br>
</div>
<div class="moz-cite-prefix">Scanning mode seems (I have not fully
tested all the configuration) to be able to list the visible APs ;
but it seems there is no provision to configure passwords for
multiple one ? Let me know if I understood it incorrectly.<br>
</div>
<div class="moz-cite-prefix"><br>
</div>
<div class="moz-cite-prefix">Regards,</div>
<div class="moz-cite-prefix"><br>
</div>
<div class="moz-cite-prefix"><br>
</div>
<div class="moz-cite-prefix">Le 21/10/2022 à 15:43, Michael Balzer a
écrit :<br>
</div>
<blockquote type="cite"
cite="mid:e310b5b1-70b3-c792-4065-e82ee865d228@expeedo.de">
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
Ludovic,<br>
<br>
sounds all reasonable.<br>
<br>
Regarding Wifi, doesn't our client "scanning mode" implement what
you're looking for?<br>
(<a class="moz-txt-link-freetext"
href="https://docs.openvehicles.com/en/latest/userguide/wifi.html#client-access-point-modes"
moz-do-not-send="true">https://docs.openvehicles.com/en/latest/userguide/wifi.html#client-access-point-modes</a>)<br>
<br>
Regards,<br>
Michael<br>
<br>
<br>
<div class="moz-cite-prefix">Am 19.10.22 um 00:44 schrieb Ludovic
LANGE:<br>
</div>
<blockquote type="cite"
cite="mid:9e83c4ad-8974-508b-f69d-1d9f51068e6d@lange.nom.fr">
<meta http-equiv="Content-Type" content="text/html;
charset=UTF-8">
<div class="moz-cite-prefix">Thanks Michael !</div>
<div class="moz-cite-prefix"><br>
</div>
<div class="moz-cite-prefix">I'll try to have a look at the can
play framework - it could surely help me speed up my dashboard
development work (for the moment I'm injecting on a real CAN
Bus but it's not always ideal). But not everything is clear
for me in this concept, given that I don't know the whole
ecosystem very well (not helping is my inability to read C++).<br>
</div>
<div class="moz-cite-prefix">If @Mark is reading, would you mind
exchanging on this topic if you have some time (and memories
of what you had in mind when designing this part) ?</div>
<div class="moz-cite-prefix"><br>
</div>
<div class="moz-cite-prefix"><br>
</div>
<div class="moz-cite-prefix">Here is a random dump of some other
crazy ideas I'd like to share with the list (read: things I
thought would be nice to discuss):</div>
<div class="moz-cite-prefix">
<ul>
<li>Some features I like (seen in the doc of a related
product):<br>
<br>
</li>
<ul>
<li>Wifi: multiple AP/SSIDs defined in order to
auto-connect to the strongest signal (or ability to
"roam" from one site to another without having to
re-configure the Wifi)<br>
<br>
</li>
<li>CanLog: ability to "split" log files according to some
criterion : either the size (in bytes) of the log file ;
or the timespan of the capture<br>
<br>
</li>
<li>CanFormat: support for reading / writing a compressed
file format (e.g. Vector's BLF, etc...)<br>
<br>
</li>
<li>CanLog: ability to sync the log files to an external
server (whenever a network connection is available),
with (optional in my mind) local delete of the file
after transfer.<br>
<br>
</li>
</ul>
<li>DBC-based vehicle :</li>
<ul>
<li>In my DBC experiments, I found I needed to register
new metrics. While it's possible to create a new
"vehicle_" module and have the registering occur there,
it kinds of defeat the dynamic aspect of the DBC
approach. So I was wondering if we could introduce a
dynamic registration of metrics (e.g.: have a config
setup, or a file in the vfs, that lists all the metrics
that we want to register during vehicle module loading)<br>
<br>
</li>
</ul>
<li>Wireguard interface : I was toying with the idea of
having a network-enabled OVMS auto-registering in a
dedicated network ; and being "locally" available, mdns
working, web and ssh reachable... while exposing no local
service on its main interface. I've seen discussion around
a "firewall", and with this approach there is no more need
to have one if no services are exposed.<br>
Don't know if it's feasible, a project like <a
class="moz-txt-link-freetext"
href="https://github.com/smartalock/wireguard-lwip"
moz-do-not-send="true">https://github.com/smartalock/wireguard-lwip</a>
is certainly interesting to look at (and also <a
class="moz-txt-link-freetext"
href="https://github.com/ciniml/WireGuard-ESP32-Arduino"
moz-do-not-send="true">https://github.com/ciniml/WireGuard-ESP32-Arduino</a>).<br>
</li>
</ul>
</div>
<div class="moz-cite-prefix"><br>
</div>
<div class="moz-cite-prefix">Let me know what you think about
any of this<br>
</div>
<div class="moz-cite-prefix"><br>
</div>
<div class="moz-cite-prefix">Regards,</div>
<div class="moz-cite-prefix">Ludovic<br>
</div>
</blockquote>
</blockquote>
<br>
<div id="grammalecte_menu_main_button_shadow_host" style="width:
0px; height: 0px;"></div>
</body>
</html>