toybox | A collection of small toy programs

localhost:8081 (Authentication server) localhost:8083 (UI) localhost:8101 All these will get treated like a different domain by the browser. So a cookie set by one of them will not be sent along with the request to others. You have 2 options here

1. First option is: If you want to stick with the traditional way of authenticating, let your Auth server handle all the request, let it proxy the UI and Upload service request to the corresponding upstream services. And you can use other means such as firewall - iptables/firewalld or security-groups/ACLs(if you are on cloud) to control access to other 2 apps (UI and Upload)

2. Second Option is: Modernize your stack, have oauth2 server as the auth server which will issue oauth2/jwt token after authenticating the user. Your javascript (from your webpage running on client's browser) should send that token along with all subsequent requests. Your other services (UI and Upload) will have provisions to talk to Auth server to verify the validity of the token. If found invalid, redirect them to the corresponding page.

The answer is that Clair doesn't support scanning Photon images. While this is readily apparent in the error message, it's hard to wrap my head around from a VMware perspective - I couldn't see VMware creating a platform like vSphere without having some workaround for scanning an image that is built on an OS that they themselves created. Unfortunately, this is the case...

CIQ apps of type "watch face" do not have access to sensors in that manner. Instead you need to use the methods available in the Activity and/or ActivityMonitor modules.

If the watch device is newer it will likely support calling this method, which returns an heart rate value that is updated every second:

Use a debugger (like GDB) and/or strace ./a.out to trace the system calls your program makes and figure out what's wrong. It's normal for asm programs to either crash or silently do nothing, because bad args to a system call typically result in -EFAULT, -EINVAL, or -ENOSYS, not raising a SIGSEGV or anything like that. But you won't know which or why unless you write error handling code or use strace.

(The OP reports that after doing this, it was obvious there was a missing movw r7,#291 for the recv call number.)

(after fixing that mistake that debugging / tracing tools make obvious):

using x/40x $pc I saw that its all zeroes with nothing filled in

$pc is the program counter, r15. You want x &to_receive or x $r1 or something to dump the memory in the static buffer you passed to recv.

Yes, the 0x1a return value should be a byte count.

flags=100 ( 0x64) also seems weird, but there's no comment to explain which flags you want or why you're writing them in decimal.

You don't even need recv in the first place; you have a connected TCP socket and probably don't need any flags, so you can simply read() from the socket FD.

If toybox can't do multiple exec, there are alternatives.

In this particular case, you may be able to just use a single stat:

No, the current Connect IQ SDK does not provide (direct) access to the gyroscope. As a matter of fact, the gyroscope is not even mentioned in the official user manual.

Based on the research of some (avid) users, it seems that the gyroscope is internally used in order to improve tracking:

https://forums.garmin.com/forum/on-the-trail/wrist-worn/fenix-5x/155802-what-s-the-main-advantage-of-gyroscope-in-fenix-5x-please-explain

For the Fenix 3/5 there exists a Service Test Boot Mode where you can see gyroscope data but, as indicated above, this is not accessible from the SDK (yet).

https://www.facebook.com/GarminFenix5/photos/a.584053075128378/640524422814576/?type=3&theater

-cfs

I also think that this functionality is not directly provided out of the box through the API.

However, you could write your own ActivityTracker based on the "ActivityTracker" sample in the Connect IQ API Docs. Unfortunately, that essentially means you would have to duplicate the logic of the display you currently see just to add the three fields.

You'd call Toybox.Activity.getActivityInfo() for obtaining the current heart rate at, say, a 1 second intervall (the ActivityTracker example shows you how to implement a 1 second callback), determine in which zone the heart rate is, and increment a counter for that zone. Then you add up the three zones and display that as well.

-cfs

The answer is ... it depends.

Generally, it is simpler and cleaner (and the API is smaller) if the ToyBox API treats all shapes the same. Treating them the same also means that you don't need to change the API to add (say) Triangle objects to the ToyBox API.

However, if there was something about the API requirements that mean that different shapes needed to be "contained" by the ToyBox in fundamentally different ways, then it could make sense for the ToyBox to have different API methods.

The implementation probably should reflect the API, but it doesn't absolutely have to. For example, you could have separate API methods for the "collections" of circles, squares, etc, store them a single list object ... and implement the methods via some complicated filtering or wrapping.

The bottom line is this: You need to understand the requirements that your API must meet ... and (ideally) reasonably anticipated generalizations of those requirements¹ ... and design accordingly.

^{1 - For example, a generalization of your ToyBox requirements that you should anticipate is that the user may want the API to be able to handle triangles.}

By the way, this won't work:

One way -- perhaps not the most elegant, but serviceable -- is to recreate the results of your tab year1 command using egen's rowtotal() function: