Re: Files are being stamped with a future time to the system time

ibuprofin@xxxxxxxxxxxxxxxxxxxxxx (Moe Trin) writes:

On Sun, 08 Apr 2007, in the Usenet newsgroup comp.os.linux.misc, in article
<nPcSh.3946$Uf7.2273@edtnps89>, Unruh wrote:

ibuprofin@xxxxxxxxxxxxxxxxxxxxxx (Moe Trin) writes:

Mark Hobley wrote:

Sat Apr 7 22:33:01 BST 2007
Sat Apr 7 21:33:01 UTC 2007

That looks fine - but TZ is a shell environmental variable that takes
precedence over a tzset (timezone) setting, AND it can be invoked on a
'by shell' basis. Try doing the time check in the same terminal you are
editing those files.

I would find it very ver weird if TZ were producing a 75 sec or 2 min
problem.

man tzset

Setting the TZ variable for a two minute error is trivial. Using
the more recent version (ex. the one in tzcode2007e), even a 75 second
offset is trivial - try 'man 3 newtzset'

The chances of his having tzcode2007e which has been out for what is it, a
day, is negligible.

Again I would guess that he would have remembered setting up such a TZ.


The OP would have had to have created his own tzdata file giving the
time as 0:02 from UTC and then compiled it with zic and installed it.
Surely he would have remembered doing so.

man tzset If he had compiled a zone with a bogus offset, unless
he were only declaring that tzfile through a TZ variable, it would
effect all shells the same way. Note the time shown at the top of
this article - one hour exactly.

Yes, that was what I meant-- his TX variable pointed at this "special" zone
file.


You wrote in your original post that this is a "local ext2 filesystem".
Are all of the files that 'make' is going to use located on "this"
computer? If not, that's likely a network time issue.

Sounds like it is more likely. Not sure which time is used for a file
on an nfs filesystem where the times on the local machine on which the
file was created and the remote machine on which the files were
actually stored differ.

This depends on the implementation. I see file times on the server
as being server time - they go over the net as UNIX time (seconds
since 00:00:00 UTC 1/1/1970) and get a local correction via the
timezone or TZ variable ON THE CLIENT. This has always been a picky
problem with network file systems. That's why we've been using a
time sync every hour on all servers since, I dunno, 1985? RFC0868
dates back to 1983.

OK, just tried it. It is the time on the server that the file gets. (set
the time ahead on a client, touched a file on the nfs system, and got the
mtime on the file as the server's time)
( at least on Mandriva 2007).


Old guy
.

Relevant Pages

  • Re: Converting the time from one timezone to another
    ... I understand the concept of UTC, ... | a way to enter the timezone as a parameter in order to find the offset. ... Carry out some calculation at the server using a date/time pasesed ...
    (microsoft.public.dotnet.framework)
  • Re: Files are being stamped with a future time to the system time
    ... man tzset ... time as 0:02 from UTC and then compiled it with zic and installed it. ... I see file times on the server ... timezone or TZ variable ON THE CLIENT. ...
    (comp.os.linux.misc)
  • Re: Selecting rows where a datetime column is in the past
    ... For another part of the app it will need to be timezone aware.. ... current time as reported by the server; as long as the dates that are ... I usually encourage people to store their data in the database as UTC ...
    (microsoft.public.sqlserver.programming)
  • Converting local time to UTC
    ... I need to convert the local-time of the server on which I run my website ... to UTC and I don't want to rely on the server local time and timezone (i.e. ...
    (microsoft.public.inetserver.asp.general)
  • Re: Requiring a date be entered as mm/dd/yy on a form field.
    ... specify that they cannot calculate to that precision. ... Furthermore, UTC ... use whatever standard they are given. ... Most areas have some name that they use to refer to their timezone. ...
    (microsoft.public.scripting.jscript)