[LEAPSECS] php breaks if UTC has no leap seconds?

[Warner Losh]

On 12/09/2010 17:35, Rob Seaman wrote:

> On Dec 9, 2010, at 3:53 PM, Steve Allen wrote:

>

>> This is the first example I've come across where a widely used API will break if UTC does not continue to have leap seconds.

> Has anyone even considered a Y2K style inventory?

>

> Absence of evidence is not...oh, what's the point?

Everybody said y2k was going to break everything.  In the end, it was a 
non-event :)  Apples are much redder than oranges, but that might not be 
a valid comparison.

We know the class of APIs that UTC is going to introduce errors into 
will be.  Astronomical events like these will degrade in accuracy over 
time, unless DUT1 is also passed in.  Technically, it needs to be passed 
in now, but the error is only second or so without it, so nobody cares 
today.  But the correction could easily be subtracted as DUT1 grows...

In fact, php could query what DUT1 is and make the corrections w/o any c

Warner