set_time() bizarre behaviour on NRF52832 (AKA "am I going mad")

[RobMeades]

Description

set_time() should set the time in seconds since 1st January 1970. However, it is behaving very strangely on NRF52832 (UBLOX_EVK_NINA_B1 platform, all default compilation switches): it seems to be adding an offset to the elapsed number of seconds since startup instead! Am I going mad!?

Here is some simple/boring code:

int main()
{
    time_t intendedTime;
    time_t currentTime;

    printf("\nStarting test.\n");

    // Set the time to 7 seconds after midnight on 1st Jan 1970
    intendedTime = 7;
    printf("@ %d setting time to %d.\n", time(NULL), intendedTime);
    set_time(intendedTime);
    currentTime = time(NULL);
    printf("time now is %d (difference %d), waiting 10 seconds...\n", currentTime, intendedTime - currentTime);

    // Wait 10 seconds
    wait_ms(10000);

    // Try again to set the time to 7 seconds after midnight on 1st Jan 1970
    intendedTime = 7;
    printf("@ %d setting time to %d again.\n", time(NULL), intendedTime);
    set_time(intendedTime);
    currentTime = time(NULL);
    printf("time now is %d (difference %d), waiting 10 seconds...\n", currentTime, intendedTime - currentTime);

    // Wait another 10 seconds
    wait_ms(10000);

    // Try setting the time to midnight on 1st Jan 1970
    intendedTime = 0;
    printf("@ %d setting time to %d now.\n", time(NULL), intendedTime);
    set_time(intendedTime);
    currentTime = time(NULL);
    printf("time now is %d (difference %d).\n", currentTime, intendedTime - currentTime);

    while(1);
}

...and here is the output:

Starting test.
@ 0 setting time to 7.
time now is 7 (difference 0), waiting 10 seconds...
@ 16 setting time to 7 again.
time now is 17 (difference -10), waiting 10 seconds...
@ 26 setting time to 0 now.
time now is 20 (difference -20).

• toolchain: GCC_ARM or ARM
• mbed-os version: 6165386 (HEAD -> master, origin/master, origin/HEAD) Merge pull request #7893 from kivaisan/improve_callback_descriptions

Issue request type

[ ] Question
[ ] Enhancement
[x] Serious bug
[ ] I'm going mad

[TacoGrandeTX]

@RobMeades It appears that you might be hitting the issue addressed by this PR: #7849. Could you please report back if that fixes the issue?

[RobMeades]

That looks plausible, I'll check it in the morning (UK time) and let you know.

[RobMeades]

Good stuff: #7849 fixes the problem, the output from my test code is now:

Starting test.
@ 0 setting time to 7.
time now is 7 (difference 0), waiting 10 seconds...
@ 16 setting time to 7 again.
time now is 7 (difference 0), waiting 10 seconds...
@ 16 setting time to 0 now.
time now is 0 (difference 0).

I suggest you get that fix in ASAP as this is a somewhat fundamental thing to leave broken.

[adbridge]

This has been marked as a duplicate of #7804
thus closing this one.