EH>
  >  JdeBP> A time_t is a time_t, and (for your compiler) it
  >  JdeBP> is always a POSIX 1003.1 seconds-since-The-Epoch number.  The
  >  JdeBP> value of the TZ environment variable doesn't change the number.
  >
  > Then I don't understand that FleetStreet does not
  > understand the number (posix date value) I write to a *.MSG
  > file when the TZ variable is set.
  > In the case the TZ variable is not set FleetStreet does
  > understand the posix date value written to the *.MSG file...
EH>

  Is it the same value in the file in both cases ?  Did you check ?  You
  certainly haven't said whether it is or not.

  You aren't expressing yourself very well.  All that I can infer from
  "FleetStreet doesn't understand the number" is that it prints the wrong
  human-readable calendar time when displaying the message, because
  obviously FleetStreet *does* understand the number (how can a program
  "not understand" a 32-bit number ?).

  It's very important to accurately describe the symptoms when you are
  asking for help with a problem.  Vague terms like "doesn't understand"
  are as useless as "doesn't work".  Exact behavioural descriptions please.

  ... preferably accompanied by the values in the file, and some snippets
  from the code that you are using to obtain those values in the first
  place.

  Anyway, to print a human-readable calendar time when you have a `time_t',
  one usually calls localtime() and strftime().  In the next paragraph of
  my original message, you will find that I said :

EH>
  > JdeBP> The value of the TZ environment variable changes the way that this
  > JdeBP> number is translated into a "broken-down time"
(i.e. a `struct tm')
  > JdeBP> by the localtime() function.
EH>

  And the reverse is true, as well.  The TZ environment variable affects
  the way that a "broken-down time" (i.e. a `struct tm') representing
  local time is translated into a `time_t' by the mktime() function.

EH>
  > JdeBP> The TZ environment variable is defined by POSIX, but not all C/C++
  > JdeBP> compilers use it the same way (or even at all), since it isn't part
  > JdeBP> of the Standard C Language (the conversion that localtime()
  > JdeBP> performs is implementation-defined in Standard C, AFAIAA).
  >
  > What do you mean by implementation-defined: the exact code
  > lines are defined, or the results are defined?
EH>

  The conversion that localtime() performs betweem its argument and its
  result is implementation-defined.

  > JdeBP <
___
 X MegaMail 2.10 #0:
--- Maximus/2 3.01