2038 Bug

[redhat81]

Will this be automatically fixed when we all start using 64 bit systems or will operating systems have to be patched to prevent this?

[mdwatts]

I get the feeling this is more of a 'off topic' question, so I'll move it over to /dev/random to give that crowd something to do. ;)

[CMonster]

If the world survives to 2038 and I am still alive at 79-years old... I'll worry about it then :D

[ricstr]

I think it will be solved with 64bit systems. But 2038 is 35 years from now people might not use linux/unix then.

[bwkaz]

Not to mention that we can double the time until this happens by simply interpreting time_t as an unsigned number rather than a signed number, even on 32-bit systems... which will put it closer to 2098 than 2038 (another 60 years or so).

Moving to a signed 64-bit time_t will double the interval, 32 times (which will assuredly be more than enough). But not everything is 64-bit, and it probably won't be for a while. In any case, changing from signed to unsigned (why would this time value ever be negative, anyway?) will add another ~60 years.

[Alex Cavnar, aka alc6379]

I can see it now! Y2k38: The second bugging!!!


Does this mean that 35 years from now, airplanes will fall out of the sky and we'll all have to live in holes, just like we were supposed to for Y2K?

:D