One Day Off on Daylight Saving Time

[high42]

First of all, congratulations on this awesome library :)

To begin, I am working with Mexico's CST (Central Stanard Time) and CDT (Central Daylight Time) which you can see here.

The problem I ran into was that when a task starts before the timezone change in spring (April 3rd) the end date (to) is one day short. I found the problem to be when looping through the values of each data element and setting a row in the Fill the Chart section.

So each time a bar is created we have the following line of code:

_bar = core.createProgressBar(dl, day);

That means that dl (duration) sets the width of the bar with the following line of code:

dl = Math.round((cTo - cFrom) / tools.getCellSize()) + 1;

However, for a specific case (default case in switch) the calculation is the following:

dl = Math.floor((dTo - dFrom) / UTC_DAY_IN_MS) + 1;

In this specific case, the Math.floor function is taking for example a 12.958333333333334 to a 12. So the solution for this specific case would be to use Math.Round();

I wanted to post this here to see why we're rounding down in that specific case. What does it aim to do? I don't know if using Math.Round() will mess with something else...

[smb]

i had the same problem and initially changed it to ceil first which caused it to error (1 extra day calculated/painted) if you have an data element from sept -> nov because of the second DST change.

so changing it to round seems to fix it for me (i only use daily view)

[usmonster]

Hello, from 2018! Sorry for the massive delay in responding, but this seems to be a duplicate of #95 and #41, so I'll close this one for now. Please feel free to follow the other issues until a fix is implemented.