Fix endless update loop in the presence of empty DataSet

[pacher]

Describe the bug
Empty DataSet causes some kind of endless update loop somewhere.

To Reproduce
Create chart, add two empty datasets. Wait a second (to let JFX thread do pending tasks). After a second add a point to one of the datasets.

My reproducer in Kotlin

    val chart = frame("ChartFx", 1366.0 * 0.7, 768.0) {
        val xAxis = DefaultNumericAxis("x")
        val yAxis = DefaultNumericAxis("y")
        XYChart(xAxis, yAxis)
    }

    val first = DoubleDataSet("First")
    val second = DoubleDataSet("Second")

    withContext(Dispatchers.JavaFx) {
        chart.datasets.addAll(first, second)
    }

    delay(1000)
    first.add(150.0, 25.0)

Observed behavior
High CPU and GPU usage, exceptions like

java.util.concurrent.RejectedExecutionException: Task com.sun.javafx.tk.quantum.PaintRenderJob@3195e9e1[Not completed, task = java.util.concurrent.Executors$RunnableAdapter@75ae38c1[Wrapped task = com.sun.javafx.tk.quantum.PresentingPainter@1ed15372]] rejected from com.sun.javafx.tk.quantum.QuantumRenderer@643b1d11[Terminated, pool size = 0, active threads = 0, queued tasks = 0, completed tasks = 4074]
	at java.base/java.util.concurrent.ThreadPoolExecutor$AbortPolicy.rejectedExecution(ThreadPoolExecutor.java:2081)
	at java.base/java.util.concurrent.ThreadPoolExecutor.reject(ThreadPoolExecutor.java:841)

when trying to exit application.
Also endless loop (and the fix) was confirmed by a println statement inside updateAxisRange function.
The problem goes away as soon as a point added to the second empty dataset.

Fix
Suggested simple fix solves the problem.

[sonarcloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarCloud

[sonarcloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarCloud

[wirew0rm]

Thanks for reporting your findings and this fix.
The early return in the updateAxisRange call is a good call in any way, although one could argue that in any way it should not be possible to create an endless loop this way.
I assume there are still ways to trigger the original problem (e.g. a dataset with one Double.Inf) will not do the early return but probably still endlessly retrigger the computation as in this case there will be no valid updated range.
But as this check does not cause any unneeded cognitive and computational complexity and solves a real problem (or at the least makes it way less probable), I'll merge this and than the underlying problem can be resolved later on.

I also ran the code formatting on the change.