Search results for:

It is a battery. A rechargeable battery in German is "Akkumulator", often shortened to "Akku". "Batterie" (German spelling of battery) would usually refer to a non-rechargeable battery...

Potentially stupid question: is the flash storage on the Bangle memory-mapped in one flat chunk? Is it somehow possible
to then get the physical address of, say, a Storage file that could then be directly accessed via compiled C?

I know, this sounds like a recipe for bricking a watch, if I break it I promise to keep both pieces without complaining...

I implemented a simple keyboard in the 'Bordle' app that mimcs the functionality of the 'Scribble' app, but uses Layout library buttons to simplify the code. Arguably it is a very simple keyboard since it mostly supports upper case letters, but perhaps it could be expanded upon...

Aha, thanks. I only every tried executing it from RAM, never tried to upload to storage and run from there. This answers my question and solves my problem.
Thanks.

I have uploaded an app called 'Bee' that sometimes has to process a large dictionary file. I noticed that executing a JS function that does not return within ~1 or 2 seconds causes the interpreter to interrupt it, logging "Execution Interrupted" to the console window (when using the IDE). Calling the same function interactively via the REPL in the IDE does not trigger this behavior. For a simple example, run the following snippet by uploading to RAM in the IDE:

function crunch() {
  for (i=0; i<10000; ++i) {
    var x = Math.sin(Math.random()*Math.PI);
    if (i%50 == 0) console.log(i);
  }
}
crunch();

Execution gets interrupted before the function finishes; calling 'crunch()' interactively from the console window however lets it finish.
I can see the purpose of this behavior to avoid lock ups in an event-loop driven system like espruino, but is there a way to (temporarily) disable it? I played around with E.kickWatchdog(), but that does not quite seem to do the trick.
Thanks.

Very cool idea, seems to work quite well!

I think this might be related to the problem reported in this thread: My GPS Run App called Race
. My hunch is that JS function names above a certain length are causing issues...

Ok, I take that back. The error above occurs with firmware version 2v12.114, but NOT with 2v12.105. I wonder if the latest firmware version limits the length of function names? This issue might be related to the conversation Potential issue with v2.12.114 where similar problems are reported?

This all looks very promising. Unfortunately I am encountering the same issue as Uname - once it acquires GPS lock the screen goes (mostly blank), and the console log shows:

in function called from system
Uncaught ReferenceError: "writeTCXfileTrackpoint" is not defined
 at line 72 col 1506 in race.app.js

Is it possible that we are missing a js module that is not being installed with the app in your repository?

Something like this?

E.showMessage("bar", "foo");
setInterval(()=>{load();}, 3000);

Hi,
This is more of a 'policy' question (mostly for Gordon, but perhaps others have good insights as well).
I wrote an app for Bangle2 that can read and display data (via BLE) from an OpenWind sailing wind instrument: http://www.openwind.de. It displays true and apparent wind speeds and angles (and optionally speed over ground) - see attached screen shot.
I use this for sailing, but I suspect that I am the only person at the intersection of Bangle owners and users of this particular model of wind instrument. Is it considered cluttering the app loader to post a pull request for a (most likely) single-user app or should we take the "the more the better approach" to publishing apps?
Thanks,
Marko

I think that is expected - compacting performs an internal cleanup of the filesystem, it should not change any of the actual files.

I have not noticed any issues with connecting to the appstore with the new bootloader.
In fact I now believe that the problem I had was not related to the new bootloader at all, but rather that the problem lies with the boot code in the health tracker app.

Sorry to reply to my own post, but the more I stare at this, the more I suspect this might not be the bootloader's fault after all. It seems that boot.js for "Health tracking" is flawed. Perhaps someone who is more JS knowledgeable can have a look...

The new version of the bootloader (0.44) seems to generate a broken version of .boot0 for the "Health Tracking" app.
Executing '.boot0' directly yields the following console error message:

Uncaught Error: Function not found!
at line 20 col 43 in .boot0
require("Storage").write(b,f(a),e,17988)}})

Obviously line 20 is specific to my setup, but this is the last line of the 'health tracking' boot code entry. It seems to be missing the final "();", maybe?
The symptom in my case was that the 'ptlauncher' section that came after the 'health tracking' section was not executed (hence ptlauncher no longer worked).
It is possible that this problem existed before bootloader 0.44, but that I had 'health tracking' installed AFTER ptlanucher, in which case the problem would have been hidden...

Thanks Gordon, that helps a lot. I use Linux and MacOS interchangeably and was able to get it to work on Linux...

Perhaps this is covered somewhere already (I briefly searched the forum, but did not find anything):
Is there a way to upload files to Bangle's storage via the command line? I am not entirely sure how the bluetooth connection would be handled in such a case...

In fact, you can try out the current (incomplete) version at https://berkenbu.github.io/BangleApps/
I haven't changed the meta data yet, you will have to select "All apps" in the top left instead of "Bangle.js 2", otherwise it won't install on Bangle2

I started working on moving the interface over to the new Layout library (so it would work on both Bangles), but my effort was based on an older code base. I would have to merge it with the cadence sensor work. I also added a added an optional device selection dialog, in case there is more than one sensor in the vicinity.
Hopefully I can find some time soon to put some more work into it...

Perhaps one further clarification: the app currently only reads the 'speed' portion of the data sent by the sensor, I wonder if that is somehow related to your problem... Is there a way to set the sensor to 'speed only' mode?

Hi,
I am the author of the app. Unfortunately I am spread a little thin at the moment and will not be able to debug this in detail in the next few days (but will hopefully get to it eventually).
I have tested the app with two different sensors, a Wahoo brand sensor and a no-name sensor from a China direct import website (I don't have access to the latter right now, so I can't check on the exact name or make). Both worked without any special tweaks, so I am not sure what is going on here.
That being said, I will have to work on the app anyways to adapt it to the new Bangle 2 layout. I will also try to source a Bk805 sensor to see if I can test it out...
Marko

Looks like you have an extra comma "," after the last closing curly bracket in your apps.json file and you are missing a final closing square bracket "]".

Great, that sounds very useful.
How do you generate the agps data to be uploaded - I briefly looked at your HTML file, is this something you process server-side every day? Looks like the binary files in http://www.espruino.com/agps are all time-stamped sometime yesterday.
Marko

Hi,
Is the graphics frame buffer on the Bangle memory-mapped and accessible from within JS? Or is data transferred in chunks to the LCD controller via SPI?
Thanks,
Marko

Thanks for adding the compiler flags, it works beautifully. I put in a pull request for the app to be added to the app loader.