• Replies 11 replies
  • Subscribers 18 subscribers
  • Views 1941 views
  • Users 0 members are here
Options
Related

Different WF behavior in VA3M, VA3 and SIM

bine8x8
I developed WF and tried it in SIM as VA3. When you change the settings of the App Settings Editor, all settings are perfectly changed and saved.
When testing this WF on my VA3M, there are no problems either.
However, I receive reports from VA3 users from the fact that the settings for this WF are not saved. If the settings are not saved on a real device, they should not be saved in the SIM. In fact, the functioning of the SIM does not coincide with the functioning of the VA3, which causes problems with the development and maintenance of applications/WF.
  • I just tried some of my own WFs on a va3, and I can change settings fine with GCM/Android. Is the user running GCM/iOS? Is it all settings that don't apply or just some? I'm on the 5.90 firmware
  • Non-saving settings occurs with Android GCM. Windows SDK 3.0.8, 5.90 firmware. The number of settings is quite large, about 80 points. With so many settings, the VA3M and SIM in VA3 mode work correctly, the settings are saved. In VA3, the same settings are not saved. There is a fact of different behavior of a real VA3 device and its simulator. Thus, when testing on a PC, I do not have the opportunity to determine the incorrect operation of WF on VA3.
  • Maybe related to this with the large number of settings? The ones I tested that work have around 20 settings.
  • Maybe related to this with the large number of settings? The ones I tested that work have around 20 settings.

    I report not that a large number of parameters are not saved, but that the SIM and the real device work in different ways. If WF being executed under the control of the SIM will not save the settings, I will see this and take action (for example, I will reduce the number of settings). However, the SIM saves these settings with confidence, creating the illusion that my real device will also save them. However, saving on a real device does not occur and there is only one way to make sure of this - install WF on a real device.

  • It's not a bug in the sim. It's likely a bug on the watch, and with 80 properties, that could be involved, as mine with a smaller number of properties works fine.
    I myself don't think the sim should replicate FW bugs. The Garmin folks will want to replicate and pass it on if it's not in the VM on the watch. Is there a link in the store for a watch face that shows this?
  • I myself don't think the sim should replicate FW bugs.

    In my opinion, the behavior of the SIM and the real device should be exactly the same. Otherwise, there is no confidence in the suitability of the distributed application / WF.

  • I really don't think the sim should replicate FW bugs. Case in point, the va3 with 5.70 FW, watch faces would sometimes vanish for no reason and be replaced with a native WF. Not something you'd want in the sim, and that's fixed in 5.90

    When I support devices you don't have, I keep an eye on those devices forums, especially when there's new FW to see what users might be seeing with other apps, and reference those posts if a user sees the same thing with one of mine.

    I don't think having 80-100+ properties is that typical, and it looks like a FW problem that many users won't see. It's working on the watch for me just fine with fewer properties, so I wouldn't want the sim to change..
  • One way or another - the problem exists. If this is a bug in FW - it should be fixed in FW.
  • Sorry to jump in here, but since it was brought up, I don’t like the established precedent where devs have to monitor the forums like a hawk to make sure they aren’t running into known issues, or look for old threads especially in the case of old devices. In the latter case, there is no way to know if the issue was fixed or not (see: animation bug which was fixed for some old devices but not all. If you assumed the bug was either fixed for all or none, you’d be wrong.)

    I do appreciate that questions along those lines are usually answered by very helpful people, but should that be the main process for avoiding device quirks? What happens if those two or three people decide to stop developing for CIQ?

    Furthermore, the format for the forums is not ideal for questions and answers. I see the same questions coming up over and over again, even on other sub forums. When useful info is posted, it’s rarely stickied, and essentially disappears within days.

    And of course ppl often respond “please do a search first”. Well obviously the search feature doesn’t work so great, or ppl just aren’t inclined to use it. Ppl obviously aren’t going to change, so maybe the format of the forums needs to change. I don’t know why CIQ related stuff is so dead on stackoverflow, which would be the perfect place to accumulate the kind of Q and A knowledge that’s found on the CIQ forums.

    I get that this is never gonna happen, but I’d like to see greater transparency for known bugs, especially those which will never be fixed. And yes, I would like to see the sim replicate certain kinds of firmware bugs, especially those that will never be fixed. (E.g. Long press not detected for older devices.) I get that this is impractical for many reasons.

    Barring that, it sure looks like the only way to do 100% foolproof ciq development is to own every device. I’m sure the same could be said of every ecosystem. But Garmin has so many classes of devices....
  • I've made a ticket to look into this. Do you know what version of GCM android was used?