bug in dateTime(now(), ...)

[johnslemmer]

Timezone localTime;

void setup()
{
  setupWIFI();
  waitForSync();
  localTime.setDefault();
    localTime.setLocation("America/Los_Angeles");
  Serial.print("Current time:       ");
  Serial.println(dateTime(ISO8601));
  Serial.print("Current time now(): ");
  Serial.println(dateTime(now(), ISO8601));
}

results in the following output:

Current time:       2019-11-11T17:57:26-0800
Current time now(): 2019-11-12T01:57:26-0800

I believe this is a bug with dateTime(). The docs make it seem like these two calls to dateTime should be equivalent (unless I'm reading the docs incorrectly). The Current time: is correct for when I ran it. The problem seems to be that when a time_t is given to dateTime it incorrectly translates it.

[underwoodblog]

Doesn't need localTime.setLocation("America/Los_Angeles");to be beforelocalTime.setDefault();?

I've noticed that putting this in setup, doesn't always work, even if Timezone is global, but in loop it does.

[johnslemmer]

@underwoodblog

Doesn't need localTime.setLocation("America/Los_Angeles");to be beforelocalTime.setDefault();?

No it doesn't. setDefault just sets a certain Timezone instance to the default. You can set/change the location as much as you want after that. At least that is the behavior I have seen in my testing.

I've noticed that putting this in setup, doesn't always work, even if Timezone is global, but in loop it does.

I have edited my original post to more clear about what is happening. But I haven't noticed any issue with how I'm doing it.

[ropg]

That does look like unexpected behaviour indeed. I just did a round of fixes, but let me look at this when I make some time for ezTime again. At that point I should be doing a larger overhaul to address larger structural issues as well, as the way I now deal with time_t in timezones is simply wrong. (It was an attempt to stay as close as possible to earlier arduino time libs, it turned into a bit of a mess.)

[vi7]

@johnslemmer thanks for the fix! @ropg I confirm the bug either, looking forward for the PR merging and new version release