No last tick on X time scale.

[Kopleman]

Hello and thank you for great Chart.js.
I dont know exactly is is a bug or maybe i missed some option but I have a strange behaviour of X time-scale - every chart that i have with huge amount of points/time(thats way i cant create plunker, sorry) have same thing there is no last time tick on X scale

[etimberg]

@Kopleman I think this is an rounding error in the code. There is some code to determine if the 2nd last and last labels will overlap. If so, the code prioritizes the last label for display. Looks like this is accidentally triggering in this case

[awallat]

I have the same problem and I created a jsbin for this behaviour:
http://jsbin.com/jofocuditu/edit?html,output

[leonexcc]

This can also happen if you use labels and autoSkip kicks in:
http://jsbin.com/lixixekuja/1/edit?html,output
(You might have to adjust the browser window because jsbin adjusts the canvas automatically. If the canvas gets small enough you can see the same problem as above)

[theawesomenayak]

This and #2964 are the same issue. Its caused due to miscalculation in the scaleSizeInUnits variable. If you look at the buildTicks function in scale.time.js, the value of me.scaleSizeInUnits is derived from lastTick.

me.scaleSizeInUnits = me.lastTick.diff(me.firstTick, me.tickUnit, true);

But the value of lastTick itself is changed much later down the function.

I tried moving the lastTick calculation to the top, but it breaks 3 test cases.

• 'Should use the isoWeekday option'
• 'should get the correct pixel for a value'
• 'should get the correct label for a data value'

The charts come out perfecly though. Any thoughts @etimberg ?

[etimberg]

What do the test failures look like?

[theawesomenayak]

PFA the error log. All other tests in the same script which test for the ticks pass.
time-scale-error.txt

[etimberg]

@sibennayak two of those tests look like something being undefined. The third looks like some pixel values changed (which would be kind of worrying)

[etimberg]

@sibennayak I have a PR based on your suggestion in #3064