x day/week/year ahead bug

ArchivesWebsite problems & bugs → x day/week/year ahead bug

CLOSED

I’ve been having this weird bug lately on the website. The pictures should pretty much explain the situation. Is it just me or does it happen to everyone?

edit: I have no idea how to reproduce this, beacause it appears random to me.

edit 2: according to bleh this number is a timestamp for Sun Jul 17 16:39:44 2022

example 1: pic1 example 2: pic2 example 3: pic3 example 4: pic4 it just happened again so example 5: pic5 last but not least example 6: pic6

Ain’t it weird, I thought it was my system time at first too. Then I checked and it was correct.
Wow. Ur living far in the future huh.
Nah I’m pretty sure this is all on your end. Whilst writing this reply, I thought of the possibility that it’s that many days left until that 2035 date thing (the one where computer clocks stop and they need to update the way clocks work to 64 I think) to effectively fix the clock thing)

And another weird thing: It doesn’t consistantly say “day ahead” but “week ahead” and “year ahead” as well (and it says day instead of day ==s== )…

d=((new Date).getTime()-c.getTime())/1e3,e=Math.floor(Math.abs(d))

Yeah I see it using Math.abs there but… Math.abs will return a negative number in ONE single case: Tossing it 100…000 (32 bit range all zero except the first, smallest int possible) will return that exact same number.

The times I see there are “-1645416736 day ahead”, “-1645351200 day ahead” and “-1645351200 week ahead” and “-1645351200 year ahead”. As the first three examples are all taken from the same thread I assume it all happened at once, he just made a screenshot and cut it into pieces, therefor the same number on all three. I’d think the numbers actually represent milliseconds, as the difference from the first to the second number is 65,536 [which would be almost a day when it was seconds (86,400)], but the original post and the edit are just about an hour apart so… yeah. Weird. With milliseconds that would be about a minute, @psrcek would have to confirm those being as close together as about a minute…

Still that shouldn’t return a negative number on Math.abs as it’s not the min int value :S

That’s just my thoughts on it :P

Greetings

~Pepich~

[EDIT]: Fuck you too, formatting <3 [EDIT2]: Le derp fucked up smallest int :P

Might be. Weird coincidenes happen. I suppose example four, five and six are all from the same screen as well. And it’s 2^16-1 so the biggest unsigned 16bit int you can get. Sill if psrcek could tell us approximately how much the difference was (a minute or an hour is close enough) that’d maybe help a lot already. But really, the last 16 bits of the number inverted? Weird…

What results did you get on Math.abs(Integer.min) or however you get the smallest int in js… It would either throw an exception or give you the nimber itself. Mathematically abs is defined as:

abs(x): if x > 0: return x; else: return -x;

Which would give you Integer.min when given Integer.min cause if you negate 10…00 you get 01…11 + 1 = 10…00 again. Or an exception if the language is smart enough (browser dependant, 32/64 bit environment or similar maybe…?). I’m too tired to put way more thoughts into this right now, I just wanted to put out my ideas ;)

Greetings

~Pepich~

[EDIT]: The formatting again. Damn it -.-

i get Sun Dec 20 2015 20:39:34 GMT+0100 (CET) and i’m using qupzilla on linux mint 17.1, fully updated