Re: time_t size: The year 2038 bug?

From: Johan Kullstam (kullstam@ne.mediaone.net)
Date: Thu Jan 06 2000 - 18:47:40 EST


Dominik Kubla <dominik.kubla@uni-mainz.de> writes:

> On Wed, Jan 05, 2000 at 05:13:00PM -0600, Bill Wendling wrote:
>
> > My main point was that 38 years is a LONG TIME in the computer industry.
> > Are you still using Wordperfect 5.1 on a 386 DOS machine at 40 MHz? It
> > requires much less energy than the Windows boxen running at 200+MHz and
> > it has the added feature that secretaries really liked it and the Windows
> > version of the same product sucked (from what I understand). That was 10
> > years ago...
>
> But your are completely missing the point! What chip is in your cell phone?
> Most likely an ARM or Mips. 32bits. What chip is used in Fords electronic
> motor control? PPC4xx. 32bits. What chip is used in the german d-box (DVB
> settop box)? M68323 with the next generation going for PPC4xx. Still
> 32bits.

it matters not that the MIPS R3K is 32-bits or not. consider:

1) the lowly 8-bit motorola 6502 is capable of 64 bit math by chaining
   eight 8-bit operations.

2) even when CPUs are 64-bit, e.g., dec/compaq-alpha, time_t is still
   often 32-bits.

people do and will continue to do stuff like

  time_t tx;
  ...
  time(&tx);
  write(tx,sizeof(time_t),fd);

and this gets embedded into fileformats like tar and various databases
making it very hard to change.

just like y2k, it's not a question of cpu, but of data and file formats.

since the binary time format is so pervasive, i predict it will be
harder to fix than the y2k problem. in y2k it was mostly I/O format
strings and BCD dates in COBOL datastructures. 2038 will be all of
unix.

-- 
J o h a n  K u l l s t a m
[kullstam@ne.mediaone.net]
Don't Fear the Penguin!

- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/



This archive was generated by hypermail 2b29 : Fri Jan 07 2000 - 21:00:08 EST