All...
Many thanks to all those who replied to my Posting, even the snotty replies.
Original Posting attached at the end of this mail.
Several people gave me a slapped wrist for not seeing this discussion
before. Apparently this has been raised in previous weeks. I'm sorry to
have asked the question again and wasted everybody's time. I don't always
have time to read all of the questions but always try to check the
summaries. I dint see a summary on this (but that's not to say there wasn't
one). Apologies once again if I missed this and caused such obvious offence
to some.
I was advised that you can find more information documented at:
http://www.merlyn.demon.co.uk/critdate.htm
However, I cant get to this page for some reason.
Seems I was worrying about nothing at all. Thanks for everybody who
replied. Too many to mention. Two very good concise replies are shown
below:
"Yes, the timestamp will go to 10 digits. NO THIS WILL NOT DO ANYTHING!
time() returns milliseconds since Jan 1, 1970. This number will not
overflow any integer wheels until July 8, 2038 (at 03:14:07 UTC if you want
to time it). By then, I fully expect everyone to be on at least 64-bit
platforms, which means it will cease to matter until well after we're dead
and gone. No one EVER writes code that "only looks at the first 9 digits"
because hardly anyone ever even uses the "real" value of time(). It is
usually converted into a human-readable form with the localtime()
function."
"It is probably worthwhile to run you in-house-developed apps on a test
server
with the date set after September. This is more for the sysadmin sanity
than
anything else. From what I understand, it takes a fair amount of skill to
write an app in such a way that it would fail..."
Best regards - Tony Miller
-----------------------------------------------Original Message-----
All...
I have heard a rumour from the SUN world that the unix time() value becomes
10 digits for the first time on Sun Sep 9 01:46:40 2001 "For the first time
in modern computer history, the timestamp
will be something besides 9 digits".
Is this true?
Is this only a problem likely to affect 32 bit Unix variants (i.e.., Solaris
2.6 and earlier)? Will dunix (64 bits) be ok or will it likely to suffer in
the same way (maybe only 32 significant bits of data??)?
Do we think that this could break anything? If so, is this fixed in any
particular patch kit? Are any specific releases of dunix likely to be
affected by this whilst other (later) releases are ok?
Sorry to pose so many vague questions, but any help you can give would be
appreciated.
regards - Tony
Received on Tue Jul 17 2001 - 16:09:05 NZST