I am building an app which list a set of files. When the user performs an action to one of the files, the action is process intensive (performs multiple ajax requests, updates an offline sqlite database, performs various checks on the file, etc). The user can trigger the said action on multiple files at a time.
The problem is that when the action is performed on more than one file, it blocks the UI. Digging down into the issue, removing certain part of the action (i.e. if I remove the section that performs a verification of the file size, type, and calculate the meta when missing, etc), it wears down the effect of UI lock, though it is still there.
From what I read so far, it is suggested to extract such logic out of the main process i.e. to have a separate process to handle those. The proposed solutions I've read are
use worker thread
create a child process
create a server
I've been struggling to understand and come up with a working example for each. I came across this module named piscina which says to generate workers (my understanding it is a wrapper to create worker thread?).
On the repo here I've setup an example which simulates the issue I have. So if I trigger multiple instances of the action by pressing the "Test" button at the bottom. And then try to increment the count at the top, I can experience the UI lock/lags. It takes approximately 10-15 clicks on the "Test" button to replicate. I can also hear the CPU fan increasing when performing this.
Are there any projects/examples around which gives a simplified implementation on how to mitigate or work around this problem?
EDIT
After #tromgy suggestions I've performed the following updates/test:
Adjusting the await run to a regular promise did seem to solve the issue.
But triggering multiple instances simultaneously, for example, 20, still causes UI lock. Attempting the same on my actual project causes memory heap errors.
I tried moving the Piscina instance outside, rather than creating a new instance on each event, it solves the memory leap, but not the UI lock/lag/freeze issue.
I've tried converting invoke to an actual event, the UI issue still persist.
Related
I have a project where I have 4 tab bars, and when i switched tabs, my api to get the API request to update my view is in the method viewDidAppear().
However this will make a very bad UX experience for user as whenever they switch tab, a loading icon and some times it load for 0.5 seconds but that instant appearance and disappearance of the loading icon is really bad for UX perspective. How can i implement a view where whenever the database value change, my view is automatically updating itself?
I have thought of using Timer() to implement where it calls the api every second but this is not practical as it will be constantly calling the API every second.
Does anyone have any suggestion? Thank you
There's a lot to unpack here, but I'll try to provide a generalized answer and point to several possible solutions to the UX problem.
There's several aspects that can play a role here:
Do you control the backend the data comes from, i.e. is it possible to follow paiv's comment and implement push notification? Or perhaps use some other means than pure HTTP(S) requests, like a websocket?
How frequently does data change? Is it realistic that it does so in between switching from one tab to another?
On that note, what do you do when data changes while your user is on one tab? Does that get refreshed in any way? Is it a problem if they don't get informed about this immediately? Is there a refresh button?
Do you persist the data beyond a view's on-screen time at all? Is it scrapped when a user re-visits a tab, i.e. does a tab always start "empty" until the API response is handled or do users see the values from the last time they visited that tab?
What best to do depends a lot on the exact answers to these questions. If you cannot enable the backend to do anything besides serving HTTP requests AND have to keep all data up-to-date constantly then obviously you have to do something like the polling you described. The update-interval depends on how important this is for your users. Even if it is 1 second, be aware that this means they could visit a tab before the next update has been fetched.
If you can adapt the backend and up-to-date data is a must, then push notifications or perhaps websockets are a way to go (keep in mind that websockets mean one open-connection per active user then!).
Push notifications would probably break down when the update interval is too high (i.e. if data gets changed in the backend very quickly and you keep spamming them), but would work nicely otherwise, plus they would update views that are already on screen if used correctly.
In 90 % of the cases I've seen, however, going with regular HTTP requests and a properly designed UI is the best choice. Properly decoupling the UI code from the data loading code and the code that displays the data is also very important. Here's how I would handle this:
Properly persist the data in some way in the app that does not require on views owning the data alone. If you want to persist beyond app runtime that means CoreData (or an equivalent), otherwise an in memory storage may be fine (that is kind of a global state then).
In your UI, show any data you currently have, even if it is "outdated" (in your case: "left over from the last time the user was on this tab"). Design the UI so that this is clearly apparent (grey it out, put tiny activity indicators next to the data while it is being loaded anew/updated, or something similar).
As said in 2. do show if the app is loading data. Initiating the loading when the users go to a tab is okay, but don't make a UI that is completely dominated by it (like one giant, empty view with a spinner on it).
Functionally, the loading should not directly update any views, instead it writes new data to your storage and just informs views and view controllers that they need to update themselves (if you're using CoreData and/or Combine you get a lot of this logic basically for free, especially for a SwiftUI app).
Such a design also lends itself to later being adapted for push notifications: While at first the re-loading is triggered by a tab switch (or a "refresh" button, for example), once (silent) PNs are available, they trigger the reloading. Even websockets can be integrated into this more or less, though there's probably a bit more to do that.
In any way, as said it all depends on your use-case and also available resources, but the gist of it is probably "If the current loading icon you have destroys the UX, redesign it properly". My gut feeling here would be that anything else might be over-engineering for your needs...
The Observable module uses change method as a way of toggling when the observers should get updates of any change in state in the Subject. However, it seems redundant to me, as notify_observers(self) it is going to be called intentionally. Is there a situation where having change makes a difference?
It allows for a split in timing, and for multiple changes without side effects.
The state of an object may or may not change during a sub-process of an application. If it changes, then .change can be invoked without any side effects. It is also possible to call it multiple times.
Later, notifications can be sent. As long as they are sent before any dependent actions are taken, then anything that depends on updates due to the original change should be able to make its update correctly
You might do this for example if processing following a change had a high cost (I/O, CPU time, network bandwidth), but only needed to be done before a second section of code, not simply every time a change occurred.
You might also do this if updates are not strictly necessary in all code paths. I.e. sometimes you care about updating due to a change, but other times it is not necessary, and any change can be ignored.
An example might be if you need to re-generate an XML document every time a content object used by the XML creation code has a property change. You would place calls to .change in the important property setters of the content object, hiding them from the external caller. You don't want to generate a new XML document each time you call content.property= (it could be very slow), instead you wait until you are finished making updates and place a single call to content.notify_observers after all possible changes.
We had to use CoFreeUnusedLibrariesEx for fixing a bug with heap not being cleared after using a MSXML library
Refer this link:
http://blogs.msdn.com/b/marcelolr/archive/2008/11/13/msxml-heaps-not-being-released.aspx
But this caused another issue with TTimers which takes while to show up and disappears when Delphi app is bounced and again shows after a while.
This app uses TTimers to schedule it's job like running a XML transform .
Here is the issue:
When TTimer.Enable is called it throws an error not enough Timer available.
I know this is a masked error and I would have to figure out how to get to the actual error.
This is a single threaded application with only one timer.
Here are the links I looked into
Most Common reason seems to be invalid windows handle
https://groups.google.com/forum/#!topic/borland.public.delphi.winapi/UrIskaFZggU
There are other threads that suggested that OS ran out of resources for TIMERS I'm not sure if that is relevant to me.
I'm Just trying to understand what is Interaction between CoFreeUnusedLibrariesEx and TTimers that it gradually some how robs it of resources and makes us bounce the app to get it working.
How do I go about solving this issue , I'm looking for some directions?
CoFreeUnusedLibrariesEx should not affect TTimers. But if loading and unloading a (buggy) dll leaks any user objects (this includes timers, window handles, ...) then I could imagine that you run out of user objects.
Use Windows Task Manager and configure it so it will show the "USER Objects" in the "Processes" tab. Then compare the number of user objects when you call CoFreeUnusedLibrariesEx and when you don't call CoFreeUnusedLibrariesEx.
Since I've started developing my Blackberry app, the biggest problems I've encountered all had to do with SQLite Databases.
Right now I'm putting my app through a stress test, and when problems pop up I address them by printing out statuses to the console and taking care of things line by line. Right now (after mashing buttons on my app) I received a "Database is locked" error and I'm not sure what to do.
It seems that once the database is locked it's locked for good until it is unlocked........ my question is how can I unlock it?? First of all, how can I check to see if it's locked??
I'm sure our users won't be mashing buttons like I did, but you never know. I want to account for every possible scenario.
Thanks
EDIT: This is what happens in my application..... When I launch it starts a thread, this thread performs a cleanup on one of my tables based on how old certain pieces of data are (uses DELETE). The thread then continues to get a USER object from my DB (read only), it then uses this USER object as a parameter to call a web service. The data retrieved from the web service is INSERTED into my database. (It's a little more complex than that as a few read/write operations are performed at this time. After that, the thread fires a callback method to update my UI.
This all works fine. I can exit the app WHILE the thread is running and relaunch and a flag will prevent it from starting a new instance of the same thread (unless the other one is done of course).
Now my problem: My app's home screen is a list of buttons, when the user clicks one of these buttons another, more detailed list is loaded (this requires a READ ONLY call to the database). When I launch the app (firing the web service calling thread) and then click a button on the main screen right away, the table gets locked. (Not always, sometimes it takes 4 or 5 tries, sometimes more, sometimes less). But if I keep doing this it WILL eventually lock making it impossible to make any calls to my DB, hence no more UI (which depends on the DB).
The DB call that populates the UI on the second screen is READ ONLY, can't I have as many of these as I need?? What causes the DB to lock?? What's the difference between a DB lock and File System error (12)??
I seemed to have fixed the problem. I was under the impression that if a read/write connection was open then a read-only connection could be created safely.
This doesn't seem to be the case. If I have a read/write connection open then no other connections can open until that one is finished.
I basically created one read/write connection, set a flag to identify it as open, and during my read connection use the same Database object if the flag is open, or create a read only if it's closed.
So far so good.
Sqlite does not support concurrent modification. In practice on BlackBerry, this means you can only open the database from one part of the code at a time. To maintain this one-at-a-time access, you need to close the database when you are done with it, as #AnkitRox points out.
However you also need to guard against concurrent access. Even if your code properly closes the database, it is possible for two different threads to access the database. In that case, you will need one to wait. In Java-ME this is typically accomplished through the 'synchronized' keyword, and using the same lock object for all database access.
Check properly that, you are opening and closing database before and after execution of query respectively.
Because if Database is going to open without closing it properly, then it gives errors.
Like with browser games. User constructs building, and a timer is set for a specific date/time to finish the construction and spawn the building.
I imagined having something like a deamon, but how would that work? To me it seems that spinning + polling is not the way to go. I looked at async_observer, but is that a good fit for something like this?
If you only need the event to be visible to the owning player, then the model can report its updated status on demand and we're done, move along, there's nothing to see here.
If, on the other hand, it needs to be visible to anyone from the time of its scheduled creation, then the problem is a little more interesting.
I'd say you need two things. A queue into which you can put timed events (a database table would do nicely) and a background process, either running continuously or restarted frequently, that pulls events scheduled to occur since the last execution (or those that are imminent, I suppose) and actions them.
Looking at the list of options on the Rails wiki, it appears that there is no One True Solution yet. Let's hope that one of them fits the bill.
I just did exactly this thing for a PBBG I'm working on (Big Villain, you can see the work in progress at MadGamesLab.com). Anyway, I went with a commands table where user commands each generated exactly one entry and an events table with one or more entries per command (linking back to the command). A secondary daemon run using script/runner to get it started polls the event table periodically and runs events whose time has passed.
So far it seems to work quite well, unless I see some problem when I throw large number of users at it, I'm not planning to change it.
To a certian extent it depends on how much logic is on your front end, and how much is in your model. If you know how much time will elapse before something happens you can keep most of the logic on the front end.
I would use your model to determin the state of things, and on a paticular request you can check to see if it is built or not. I don't see why you would need a background worker for this.
I would use AJAX to start a timer (see Periodical Executor) for updating your UI. On the model side, just keep track of the created_at column for your building and only allow it to be used if its construction time has elapsed. That way you don't have to take a trip to your db every few seconds to see if your building is done.