On 4/13/25 23:33, James Mulder wrote:
If I am reading the code correctly, I think that CONVTOD will give return
code x'14' if a date beyond Sept 17, 2042 for TODVAL is specified.
...
Will STCK set error status if issued on such a date? It would seem
to make sense if CONVTOD returned the same value and status as STCK.
...
STCKCONV is a more interesting question. Peter Relson added code ...
But we found that it caused a problem for some z/OS code that used STCKCONV in
an unconventional way
...
Ouch! It aint easy. In:
<https://www.ibm.com/docs/en/ztpf/1.1.2025?topic=clocks-remediation-methods-2042-tod-clock-rollover>
The z/TPF time-of-day (TOD) clock will overflow
on 17 September 2042 at 23:53:47.
All that precision and no timezone identified. Submitted Feedback.
Bracketed TOD clock
A bracketed TOD clock value, which is contained in the tpf_BrkTOD_type
data structure, is an 8-byte time-of-day (TOD) clock value. The raw
data in this structure is identical to the data of a standard TOD clock
value, but the range of years that the bracketed TOD clock format can
represent is offset by approximately 71 years into the future.
Hmmm. That makes the comparator work unsigned in 2043 but causes
more problems in software than it solves. Will any systems lacking
signed comparator be supported through 2042?
How will DB2 and DFSORT be affected? But that and many other products
are questions for a different list. -- Thanks, gil
