You probably have missed my locationservice extension
I'm currently working on implementing itoo functionality...
Taifun
You probably have missed my locationservice extension
I'm currently working on implementing itoo functionality...
Taifun
Can i have a test of your extension. just to make sure your ext can send live to clouddb in my phone, Before buying full version
No you can't... the sane restrictions also apply for my extension...
In case ther iis a method, which uses the web component to connect to CloudDB, then it would be possible...
Taifun
In the meantime i think i can do store property in list method and fetch the data when app is alive that is possible right.
Or broadcast block.
Its not live data.
I need help tracking down a timing/out-of-sync problem that I have illustrated via the attached simplified case [please ignore certain text that would apply to the app that it was derived from ]
Here is the aia and blocks (sorry blocks are so small, I used blocks editor save as image option):
DataMinder_ButtonCase.aia (903.1 KB)
The troublesome behavior centers around the NotificationStyle1.ActionNotification with 3 user response buttons and the button response event NotificationStyle1.CallbackButtonAction, which is registered (from 'run') with itoo.RegisterEvent and procedure NotifyStyleCbButtonProcessing.
Initially the Clock.Timer event TimerInterval is 20 sec (20000). When the app starts and "Start" is touched, about 20 secs later the Timer event processing issues the following notification showing the time issued and the value of the TimerInterval (ignore the text that follows):
The user is presented with a choice:
If "Half" is selected two or three times (reducing the interval to 5000 or 2500), and then "Double" is selected the same number of times, the TimerInterval does not return to 20000.
Evidently I have done something(s) that does not assure synchronous behavior.
This is mitigated (perhaps solved?) by the introduction of Notifier1.loginfo messages which are "disabled" in the blocks.
What have I done? not done?
Kind regards,
Randal
Debug your logic using Logcat. To find possible errors of your blocks in the background, let me suggest to filter the logcat output for the tags, Notifier , ItooCreator and System.err
Taifun
I think I have done about all I can to sort this one. I may have one more idea in me :-), but I am not hopeful
Below find the updated the test case code and blocks image.
I have also included an annotated logcat output filtered on several character strings grep -e %%% -e TimerInterval -e Notifi -e System.err -e CallbackButtonAction
.
The notations in the logcat files (included below) trace the following behavior:
See attached -hidden.pdf file.
DataMinder-wlogs3_20240210-%TimerNotSysCallonly-hidden.pdf (82.4 KB)
To see intermediate/additional filtered logcat entries, see un-hidden rows via:
DataMinder-wlogs3_20240210-%TimerNotSysCallonly.pdf (97.7 KB)
Unfiltered logcat entries are hare:
DataMinder-wlogs3-51-53.txt (71.2 KB)
What I cannot fathom is #13, above: why does the ButtonCallback execute two additional times when there is/was no further user input to the NotificationAction window.
It is as if, when no user input was made (items #8 and #10) that these CallBack "events" were held and then were executed immediately after the expected one (#12), with the last user selected action Button (Double) being used. I have tried many ways to cancel the NotificationAlerts that were "stale" not responded to (see blocks). But they seem to stick around and then mis-behave.
Any help would be appreciated,
-Randal
DataMinder_ButtonCase.aia (903.6 KB)
I would advise you to not use the CallBackButtonAction feature for now, we are not sure how it behaves in the background.
PS: I'm sorry I'm preparing for my exams right now and cannot reply you.
What would be the "normal" way to communicate a change in parameters to a background process that requires periodic user input ?
The blocks below show what I did that seems to avoid the multiple-queuing of unwanted Button-actions. Note the use of itoo1 database item "ActionNotificationQueued" as a flag in the notification sending procedure (LargeIconNotificationWithOptions), and in the button callback processing procedure (NotityStyleCbButtonProcessing) .
It allows only one outstanding notification to be "queued" (flag set by b/g timer processing). If it does not get cleared before the next periodic cycle (via ButtonCallback processing) , the notification is canceled, and the flag is cleared. Then on the following periodic cycle a new notification is issued.
Since I do not understand how all this works I suppose this may not be bullet proof. But I am hoping that by allowing an entire period/cycle to pass with no new outstanding notification, there will be time for any CallBack processing already started to complete, awaiting the next cycle.
Another approach (that I have not tried) is to post the same notification for user input when the b/g process is started, setting the "queued" flag. Then use Notifer (not reliable?) and SimpleToast to periodically notify the user when input is suggested (e.g., "Stop sending these notices for 2 hours!!").
At any point the user could navigate to the notification that was posted on Start and supply what is needed. This would be processed by ButtonCallback clearing the flag when done. The next periodic cycle that detects a cleared flag would post another notification and set the flag.
Thanks for any comments and help you might offer, kind regards,
Randal
Thanks for the message and I wish you very good success on exams!!!
-Randal
Hi Randal, rather than the explanation (what you have done so far), it would also help to know what you are trying to do so far.
Also since I assume you are touching the callback button features of the notification style component, it's not guaranteed to work in the background (that's why dont use it for now).
I'll be able to take a look into this thoroughly next month.
Hello Kumaraswamy,
I know you are especially busy, so thanks for the reply.
You asked:
what you are trying to do
I want to periodically check the data transport mechanism. If it is cellular, I want to notify the user and give him the opportunity to increase the time between checks or to stop checking ("snooze the app") for a certain period of time. This could help those without unlimited data plans to be more aware of their data use.
I want it to start when the phone is started.
Eventually, I want the user to be able to set various parameters that determine 1) how frequently to be notified, 2) for what reason (under what conditions) to be notified besides simply that data is in use, and 2) how to be notified (a notification, toast, notifier, other?).
Again, I wish you good success in your academic achievements and upcoming exams.
Randal
You could have the user do this kind of configuration before starting the background service and store these choices using the StoreProperty method. Like this you do not have to use the CallBackButtonAction feature
Taifun
Hi Taufun,
Thanks for helping me think through this one.
I am thinking about your suggestion, which is a good one. Except I do not think I can design a configurable set of parameters that would be sufficient without requiring user input in real time.
The main challenge is different (even changing) user circumstance, preference, and intent. In one instance, the user might be intentionally using cell data for what he anticipates to be a very short time and does not want to forget to turn it off when he resumes watching his movie or listening to his "tunes" which he "thinks" are downloaded onto his phone... so for that situation, he would not want to "snooze" the monitor but be willing to put up with repeated reminders (a minute or two apart).
In another instance, the user may intend to use cell data for an extended period since he realizes he may not be near wifi for hours...so he would want to "snooze" the monitor for those number of hours (or hour by hour until wifi is available - which the monitor could communicate).
Another example is whether or not to include a sound in the notice - preference might vary depending on the circumstance.
I suppose, when the user wanted to make these changes in monitor "parameters", he could simply restart the app, make changes, and then close it (without stopping the b/g task)... this might work well enough. Perhaps this is what you had in mind.
In an earlier approach, I experimented with your alarmmanager facility. Frankly, I forget what problems I ran into (perhaps of my own making and ignorance), but the itoo background approach seemed more of a fit - until I started using the NotificationStyle callback events
BTW, is there a proper way to communicate reliably between the user and a background task?
Thanks for your thoughts and suggestions,
Randal
Notification Style extension's callbacks just isnt able to properly work in background. I'll write a new notifications extension if possible next month.
You mean live communication or communication by constantly reading/writing? Both are possible.
Check the RegisterBroadcast
, Broadcast
, UnregisterBroadcast
blocks for live communication or FetchProperty
and StoreProperty
for storing/retrieving data.
I think that would be quite nice ... and I know you'd rather work on that than prep for exams but "nose to the grindstone" (pardon the English idiom)
I have employed FetchProperty
and StoreProperty
successfully....What I am stumbling on is the "live communication".
Yes, I must investigate this.
My concern is that, while the blocking behavior would guarantee synchronization of communication, I would expect to need a "timeout" parameter, so that user inattention could not indefinitely hang the cyclical monitor (Timer). But I need to use these to see how they work. Thanks for the suggestion.
I think I have found an example here that uses Broadcast:
I'll see what I can come up with ...
Kind regards,
Randal
How could I replace the /!\ (exclamation point circumscribed by a triangle) notification icon in the following with one of my own choosing?
and
The blocks used in the app are substantially the same as those shown earlier in this thread.
-Randal
Well, I'm sorry I had to remove that feature in the previous update due to some reasons. I'll re-add it back in the next version (just few weeks more until my exams end)
No problem - I'm quite happy at the prospect that it may be possible
Thanks for taking time to reply .... Hang in there!
Kind regards,
-Randal
Hello @Kumaraswamy! I am trying to reduce the amount of logging my app is creating in logcat, in order to improve the app performance.
The lines I am trying to get rid of in logcat look like this:
03-09 15:45:38.879 2128 2128 D ItooCreator: AEvent(BytesReceived=BluetoothLE1) args [0000ff01-0000-1000-8000-00805f9b34fb, 0000ff02-0000-1000-8000-00805f9b34fb, (153 19 0 32 192 224 176 200 7 219 ...)] listener = xyz.kumaraswamy.itoo.Itoo$1@294738c
03-09 15:38:27.522 2038 2038 D ItooCreator: AEvent(RssiChanged=BluetoothLE1) args [-28] listener = xyz.kumaraswamy.itoo.Itoo$1@b27b2c7
Do you have any hints/guidance on how I can achieve this? So far, I looked into the source code to see where this line is logged, but wasn't able to. I am thinking maybe it's somewhere in the itoo-x repo.