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

[Ian Batten]

On 10 Dec 2010, at 15:15, Peter Vince wrote:


> Hello Paul,

> 

>     I'd be interested if you have some examples of of Y2K bugs that

> were fixed before they became a problem.  In my very limited

> experience, I wasn't affected by any, nor aware of them.


The Burroughs "TMS" manufacturing system stored and processed dates as DD-MM-YY, and was therefore unable to plan manufacturing once the planning horizon crossed 01-01-00.  My then employer had to do a crash replacement (the problem was too pervasive for remediation) of it MRP system by Easter 1998, as it was operating of an 18-month planning horizon.

The Atlas EDI system had a similar problem, along with quite a few other EDI systems, in that there were two-digit dates being crudely coerced to four-digit by adding 1900.  For that we added a preprocessor that did 0..50 => 2000+X, 51..99 => 1900+X.

Like I suspect most businesses, we found there was an endless parade of internal code written by not-quite-wizards which displayed 19100 rather than 2000, because it did printf ("19%2d", t->tm_year) rather than printf ("%4d", 1900 + t->tm_year).

ian