[ILUG] Linux's 2038 Bug

[Paul O'Malley]

Richard Eibrand wrote:

> Hi all,
>
>
> Just read this article
> http://us.rediff.com/money/2005/may/06linux.htm,  and was wondering
> what peoples thoughts are with regards this problem.  Some more scare
> mongering?
>
> Here is a snippet to get a gist,
>
> {snip}
>
> R
>
In the spirit of not agreeing with everything that is said:
I do not concur that the article is FUD - it is a little alarmist,
however I have five questions for you:

(A) how much of todays technology will be use in 2038?
less than 10% I would guess
(B) of that fictional 10% how much will actually have time_t in it?
less than the 10%
(C) for any unit or device in that small grouping where time is
measured, how often is it a function of day and date that exists there?
Mostly time is a matter of seconds - so it is not still much of an issue
(D) do you think that manafacutrers are not aware of this?
I remember getting a video recorder in spring 1992 and it was able to
run to do dates up to 2050 it is has already gone to the great recycling
home in the sky
(E) if you still think that the article is fud read this bit:
"How can the problem be sorted? Modern Linux programs could use 64-bit
or longer time_t data storage to overcome the problem. As for the
existing systems, the way the C language stores time_t data could be
changed and then all the programs could be recompiled. All this is
easier said than done."

When 64bit is normal there will be people who will want to use 64bit
apps on 32bit machines - so a patch will be made for the kernel.

So what falls from all of this is as follows, in the rare case where the
technology still exists and is operational in 33 years it may break in
several ways, interesting or not someone will deal with it.

[joke]
all of the above fixes will not happen for you if you are running Debian
stable
[/joke]

Regards,

Paul