I have a JavaFX UI that displays a relatively large data set. When the data set is changed, the changes are queued up and then the changes need to be "pushed" to the UI. Often there are more changes in the queue than I can push to the UI in one Platform.runLater() -- trying to push them all causes the JavaFX window to stutter or freeze.
Since the changes are in a queue, it's easy for me to break them into a series of batches. I would like to run a batch each time the javaFX ui thread loops.
How do I connect to the FX thread so it calls my method once every loop? In addition, about how many milliseconds can I responsibly use per-loop to update the data set?
I have a small IDE for a modeling language I wrote, implemented in PyQt/PySide, and am trying to implement a code navigator that let's you jump to different sections in the file being edited.
The current implementation is: (1) connect to QPlainTextEditor.textChanged, (2) any time a change is made, (sloppily) parse the file and update the navigator pane
It seems to work OK, but I'm worried this could cause major performance issues for large files on slower systems, in particular if more stuff is connected to textChanged in the future.
My question: Has anybody here implemented a delayed reaction to events, so that multiple events (i.e. keystrokes) within a short period only trigger a single update (say once per second)? And is there a proper QT way of doing this?
Thanks,
Michael
You can try using timers if you want some "delay".
There would be 2 ways to use them (with different results).
One is only parse after no input has been done for a certain amount of time
NOTE: I only know C++ Qt but I assume the same things are valid for pyqt so this is kind of "pseudocode" I hope you get the concept though.
QTimer timer; //somewhere
timer.setSingleShot(true); //only fire once
connect(timer,QTimer::timeout(),OnTimerDone(...);
OnTextChanged(...)
{
timer.start(500); //wait 500ms
}
OnTimerDone(...)
{
DoStuff(...);
}
This will restart the timer every input, so when you call that and the timer is not done the timeout signal is not emitted. When no input is done for an amount of time the timer timeouts and you parse the file.
The second option would be to have a periodic timer running (singleShot(false)).
Just start the timer for like each second. and timeout will be called once a second. You can combine that with a variable which you set to true when the input changes and to false when the file is parsed. So you avoid parsing when nothing has changed.
In C++Qt you won't have to worry about multi-threading because the slot gets called in the GUI thread. I assume it is the same for python but you should probably check this.
A while ago I wrote a little RAII class to wrap the setOverrideCursor() and restoreOverrideCursor() methods on QApplication. Constructing this class would set the cursor and the destructor would restore it. Since the override cursor is a stack, this worked quite well, as in:
{
CursorSentry sentry;
// code that takes some time to process
}
Later on, I found that in some cases, the processing code would sometimes take a perceptible time to process (say more than half a second) and other times it would be near instantaneous (because of caching). It is difficult to determine before hand which case will happen, so it still always sets the wait cursor by making a CursorSentry object. But this could cause an unpleasant "flicker" where the cursor would quickly turn from the wait cursor to the normal cursor.
So I thought I'd be smart and I added a separate thread to manage the cursor override. Now, when a CursorSentry is made, it puts in a request to the cursor thread to go to the wait state. When it is destroyed it tells the thread to return to the normal state. If the CursorSentry lives longer than some amount of time (50 milliseconds), then the cursor change is processed and the override cursor is set. Otherwise, the change request is discarded.
The problem is, the cursor thread can't technically change the cursor because it's not the GUI thread. In most cases, it does happen to work, but sometimes, if I'm really unlucky, the call to change the cursor happens when the GUI thread gets mixed in with some other X11 calls, and the whole application gets deadlocked. This usually only happens if the GUI thread finishes processing at nearly the exact moment the cursor thread decides to set the override cursor.
So, does anyone know of a safe way to set the override cursor from a non-GUI thread. Keep in mind that most of the time, the GUI thread is going to be busy processing stuff (that's why the wait cursor is needed after all), so I can't just put an event into the GUI thread queue, because it won't be processed until its too late. Also, it is impractical to move the processing I'm talking about to a separate thread, because this is happening during a paint event and it needs to do GUI work when its done (figuring out what to draw).
Any other ideas for adding a delay to setting the override cursor would be good, too.
I don't think there is any other way besides a Signal-Slot connection going to the GUI thread followed by a qApp->processEvents() call, but like you said, this would probably not work well when the GUI thread is tied up.
The documentation for QCoreApplication::processEvents also has some recommended usages with long event processing:
This function overloads processEvents(). Processes pending events for
the calling thread for maxtime milliseconds or until there are no more
events to process, whichever is shorter.
You can call this function
occasionally when you program is busy doing a long operation (e.g.
copying a file).
Calling this function processes events only for the
calling thread.
If possible break up the long calls in the paint event and have it periodically check to see how long it has been taking. And in any of those checks, have it set the override cursor then from in the GUI Thread.
Often a QProgressBar can go a long way to convey the same information to the user.
Another option that could help quite a bit would be to render outside of the GUI thread onto a QImage buffer and then post it to the GUI when it is done.
I have been involved in building a custum QGIS application in which live data is to be shown on the viewer of the application.
The IPC being used is unix message queues.
The data is to be refreshed at a specified interval say, 3 seconds.
Now the problem that i am facing is that the processing of the data which is to be shown is taking more than 3 seconds,so what i have done is that before the app starts to process data for the next update,the refresh QTimer is stopped and after the data is processed i again restart the QTimer.The app should work in such a way that after an update/refresh(during this refresh the app goes unresponsive) the user should get ample time to continue to work on the app apart from seeing the data being updated.I am able to get acceptable pauses for the user to work-- in one scenario.
But on different OS(RHEL 5.0 to RHEL 5.2) the situation is something different.The timer goes wild and continues to fire without giving any pauses b/w the successive updates thus going into an infinite loop.Handling this update data definitely takes longer than 3 sec,but for that very reason i have stopped-restarted the timer while processing..and the same logic works in one scenario while in other it doesnt.. The other fact that i have observed is that when this quick firing of the timer happens the time taken by the refreshing function to exit is very small abt 300ms so the start-stop of the timer that i have placed at the start-and-end of this function happens in that small time..so before the actual processing of the data finishes,there are 3-4 starts of the timer in queue waiting to be executed and thus the infinite looping problem gets worse from that point for every successive update.
The important thing to note here is that for the same code in one OS the refresh time is shown to be as around 4000ms(the actual processing time taken for the same amount of data) while for the other OS its 300ms.
Maybe this has something to do with newer libs on the updated OS..but I dont know how to debug it because i am not able to get any clues why its happening as such... maybe something related to pthreads has changed b/w the OSs??
So, my query is that is there any way that will assure that some processing in my app is timerised(and which is independent of the OS) without using QTimer as i think that QTimer is not a good option to achieve what i want??
What option can be there?? pthreads or Boost threads? which one would be better if i am to use threads as an alternate??But how can i make sure atleast a 3 second gap b/w successive updates no matter how long the update processing takes?
Kindly help.
Thanks.
If I was trying to get an acceptable, longer-term solution, I would investigate updating your display in a separate thread. In that thread, you could paint the display to an image, updating as often as you desire... although you might want to throttle the thread so it doesn't take all of the processing time available. Then in the UI thread, you could read that image and draw it to screen. That could improve your responsiveness to panning, since you could be displaying different parts of the image. You could update the image every 3 seconds based on a timer (just redraw from the source), or you could have the other thread emit a signal whenever the new data is completely refreshed.
In our game project we did have a timer loop set to fire about 20 times a second (the same as the application framerate). We use this to move some sprites around.
I'm wondering if this could cause problems and we should instead do our updates using an EnterFrame event handler?
I get the impression that having a timer loop run faster than the application framerate is likely to cause problems... is this the case?
As an update, trying to do it on EnterFrame caused very weird problems. Instead of a frame every 75ms, suddenly it jumped to 25ms. Note, it wasn't just our calculation claimed the framerate was different, suddenly the animations sped up to a crazy rate.
I'd go for the Enter frame, in some special cases it can be useful to have two "loops" one for logic and one for the visuals, but for most games I make I stick to the Enter frame-event listener. Having a separate timer for moving your stuff around is a bit unnecessary since having it set to anything except the framerate would make the motion either jerky or just not visible (since the frame is not redrawn).
One thing to consider however is to decouple your logic from the framerate, this is most easily accomplished by using getTimer (available in both as2 and as3) to calculate the time that has expired since the last frame and adjusting the motions or whatever accordingly.
A timer is no more reliable than the enter frame event, flash will try to keep up with whatever rate you've set, but if you're doing heavy processing or complex graphics it will slow down, both timers and framerate.
Here's a rundown of how Flash handles framerates and why you saw your content play faster.
At the deepest level, whatever host application that Flash is running in (the browser usually) polls flash at some interval. That interval might be every 10ms in one browser, or 50ms in another. Every time time that poll occurs, Flash does something like this:
Have (1000/framerate) miliseconds passed since the last frame update?
If no: do nothing and return
If yes: Execute a frame update:
Advance all (playing) timelines one frame
Dispatch all events (including an ENTER_FRAME event
Execute all frame scripts and event handlers with pending events
Draw screen updates
return
However, certain kinds of external events (such as keypresses, mouse events, and timer events) are handled asynchronously to the above process. So if you have an event handler that fires when a key is pressed, the code in that handler might be executed several times between frame updates. The screen will still only be redrawn once per frame update, unless you use the updateAfterEvent() method (global in AS2, attached to events in AS3).
Note that the asynchronous behavior of these events does not affect the timing of frame updates. Even if you use timer events to, for example, redraw the screen 50 times per second, frame animations will still occur at the published framerate, and scripted animations will not execute any faster if they're driven by the enterFrame event (rather than the timer).
The nice thing about using enter frame events, is your processing will degrade at the same pace as the rendering and you'll get a screen update right after the code block finishes.
Either method isn't guaranteed to occur at a specific time interval. So your event handler should be determining how long it's been since it last executed, and making decisions off of that instead of purely how many times it's run.
I think timerEvent and Enter Frame are both good options, I have used both of them in my games. ( Did you mean timerEvent by timer loop? )
PS: notice that in slow machines the timer may not refresh quick enough, so you may need to adjust your code to make game work "faster" in slow machines.
I would suggest using a class such as TweenLite ( http://blog.greensock.com/tweenliteas3/ ) which is lightweight at about 3kb or if you need more power you can use TweenMax, which i believe is 11kb. There are many advantages here. First off, this "engine" has been thoroughly tested and benchmarked and is well known as one of the most resource friendly ways to animate few or even many things. I have seen a benchmark, where in AS3, 1,500 sprites are being animated with TweenLite and it holds a strong 20 fps, as where competitors like Tweener would bog down to 9 fps http://blog.greensock.com/tweening-speed-test/. The next advantage is the ease of use as I will demonstrate below.
//Make sure you have a class path pointed at a folder that contains the following.
import gs.TweenLite;
import gs.easing.*;
var ball_mc:MovieClip = new MovieClip();
var g:Graphics = ball_mc.graphics;
g.beginFill(0xFF0000,1);
g.drawCircle(0,0,10);
g.endFill();
//Now we animate ball_mc
//Example: TweenLite.to(displayObjectName, totalTweeningTime, {someProperty:someValue,anotherProperty:anotherValue,onComplete:aFunctionCalledWhenComplete});
TweenLite.to(ball_mc, 1,{x:400,alpha:0.5});
So this takes ball_mc and moves it to 400 from its current position on the x axis and during that same Tween it reduces or increases the alpha from its current value to 0.5.
After importing the needed class, it is really only 1 line of code to animate each object, which is really nice. We can a also affect the ease, which I believe by default is Expo.easeOut(Strong easeOut). If you wanted it to bounce or be elastic such effects are available just by adding a property to the object as follows.
TweenLite.to(ball_mc, 1,{x:400,alpha:0.5,ease:Bounce.easeOut});
TweenLite.to(ball_mc, 1,{x:400,alpha:0.5,ease:Elastic.easeOut});
The easing all comes from the gs.easing.* import which I believe is Penner's Easing Equations utilized through TweenLite.
In the end we have no polling (Open loops) to manage such as Timer and we have very readable code that can be amended or removed with ease.
It is also important to note that TweenLite and TweenMax offer far more than I have displayed here and it is safe to say that I use one of the two classes in every single project. The animations are custom, they have functionality attached to them (onComplete: functionCall), and again, they are optimal and resource friendly.