Re: TIME opcode analogous to QTIME?

[Bruce Vining <bvining@xxxxxxxxxx>]

In V5R3, the adjtime(), gettimeofday(), and settimeofday() APIs have been
changed to
use the system clock rather than the software clock.  This is documented in
the V5R3
Memo to Users under Command and API changes.

>From the original question (SNTP client) V5R3 also changes which clock is
maintained
by the i5/OS supplied SNTP client.  With V5R3 the system clock is now being
maintained.
This is documented in the V5R3 Memo to Users under SNTP client changes the
system clock
instead of the software clock.

For both changes the Memo to Users also mentions how to get the pre-V5R2
behavior.

Also from the original question, the RPG TIME opcode is referencing the
"system" (not
software) clock.  This clock is the same as that used when referencing the
QTIME system
value.  It should also be noted that in V5R3 time is now internally
maintained with a
UTC base and the "system" time (QTIME, TIME opcode, and the like) is
calculated on the
fly based on the time zone specified with the QTIMZON system value.  New
interfaces are
provided to directly access this UTC time.  Prior to V5R3 the system
internally
maintained time based on the local time value.

As you can see there have been many time related changes in V5R3.

>You can be assured of accessing the software clock if you call the C
>function gettimeofday().  Using it in RPG is not very easy since
>gettimeofday() required a pointer to a struct timeval.

>> I'm using the SNTP client to access a machine configured as a time
server on
>> our LAN; then I just read Scott Klement's answer to a submitted question
in
>> ClubTech that the time being updated is OS/400's software clock and not
the
>> hardware clock.  He states that QTIME, etc. is tied to the hardware
clock.
>> I was wondering if the TIME opcode used the software or hardware clock.