Randall S. Becker

> > There are many tokenization solutions out there. If it were me, I would suggest building an API to allow different token providers to plug in rather than building...

Please be aware that `restrict` is not necessarily portable. There should be a mechanism for disabling its use on platforms that do not support strict c99 features.

> @rsbeckerca do you have any idea what could be the cause? Apparently the timezone variable contains something wrong in NonStop. Does NonStop have timegm() function? If so would you...

> Could you please dump what are the contents of timestamp_tm and timezone when the timestamp_local is calculated? Sure. Will try to find an appropriate place - suggestions would save...

> For example here: > > https://github.com/openssl/openssl/blob/e9492d1cecf459261f1f5ac0eb03e9c631600537/crypto/asn1/a_time.c#L632 Thanks. Have to rebuild with `-O0`. Stand by.

> Could you please dump what are the contents of timestamp_tm and timezone when the timestamp_local is calculated? This does not look right. `tm_min` cannot possibly be `59` when I...

What the system actually reports is (when just run now): ``` now=1646664684 tm_sec=24 tm_min=51 tm_hour=14 tm_mday=7 tm_mon=2 tm_year=122 tm_wday=1 tm_yday=65 tm_isdst=0 ```

@t8m The call that is failing is `asn1_string_to_time_t (asn1_string=0x8001350 "210328005959Z")`. This time is across the Epoc boundary and cannot be represented in 32-bit time_t. I don't have a variable `timezone`...

timezone is inferred to be: `25200` from: ``` $4 = 1616914799 (xInspect 2,696):n * 635 timestamp_utc = timestamp_local - timezone; (xInspect 2,696): * 638 ASN1_TIME_free(timestamp_asn1); (xInspect 2,696):p timestamp_utc $5 =...

> > This time is across the Epoc boundary and cannot be represented in 32-bit time_t. > > Ah right, but the failure is not associated with that fact. As...