Time and date window glitch at boundary datetimes

[Richard bartosz]

Experimenting with positions of sky objects between Julian and Gregorian calendars at the October 4/15 1582 changeover, I found that going forwards by "days" from 4 October 1582 was working fine. However, going back the program stalls at 15 October 1582 i.e. it won't proceed any further back. Month and year steps work fine but not days.

The program itself doesn't stall and I can work around the glitch by stepping back a month and then progressing forwards by days.

I reinstalled stellarium-0.12.2dev5-win32.exe, which I was previously using, and found the same glitch.

I am using OS XP-SP3, version 2002, Home Edition.

Richard

[Bogdan Marinov (daggerstab)]

It's because of the StelUtils::changeDateTimeForRollover() function - it validates dates submitted by the Date and time window. When it comes to that date, it rounds it "up". There is a similar bug with the day/time of daylight saving time transition and cycling by hour/minute.

Problems like these will be addressed by the time zone code on which I'm working, so I'm assigning this bug to myself.

[Alexander Wolf (alexwolf)]

A fix has been committed as revision 8385 of the trunk branch of Stellarium's Bazaar repository at Launchpad: http://bazaar.launchpad.net/~stellarium/stellarium/trunk/revision/8385