IsValidDate fails after the Year 9999

[Nigel Thomas 28]

DateUtils.IsValidDate checks if the Year is <= 9999. This means the test will fail in the year 10000 and forward. Shouldn't it be fixed now, to avoid a new "Millenium Bug" panic?

[programmerdelphi2k 235]

Independent of the System.DateUtils functions, I have tested the minimum and maximum value accepted by Delphi to inform a valid date and time, being:

• Delphi: of course, all should knows it
  • Min: 30/12/1899 12:00 AM
  • Max: 31/12/9999  23:59:59.999
• TDateTime var using a "double" value directly
  • Min: 01/01/0001    00:00:00.001  =         -693593.00000001   =       693,593 days   = 1900 years
  • Max: 12/31/65535  23:59:59.990  = +669528949.999999940  =  23,920,275 days   = 65,535 years
  • Total: 24,613,868 days   = 67,435 years

then, the limitation is just from Delphi procedure!

 

using TDatePicker (MSWin10 component) =  Min: 01/01/0001  Max: 12/31/9999   = 64bits

using TDateTimePicker                                = Min: 01/01/1601  Max: 12/31/9999   = 32bits

 

NOTE: in function to Date/time, Delphi use "look" just first 4 digits for "years", then, the year "65535" stay as "6553"! here is the limitation!

Edited February 2, 2023 by programmerdelphi2k

[Ian Branch 121]

I was there for Y2K.  Not sure I will be around for Y10K... 😉

[David Heffernan 2300]

5 hours ago, Nigel Thomas said:

DateUtils.IsValidDate checks if the Year is <= 9999. This means the test will fail in the year 10000 and forward. Shouldn't it be fixed now, to avoid a new "Millenium Bug" panic?

No. 

[Fr0sT.Brutal 897]

Not only this function... just think about all these dd-mm-yyyy patterns that will break

[Attila Kovacs 617]

That's why I'm using stardate.pas

[Alexander Elagin 143]

There is a nice RFC 2550 dealing with dates up to year 10E30 and beyond (dated 1 April 1999 )

[Sherlock 634]

3 hours ago, David Heffernan said:

No. 

To elaborate on this correct and concise answer: One should only fix things that a really broken or occur during the life cycle of the software. You simply cannot think of all eventualities - one of them being: will there still be a Delphi compiler in the year 9999? If not, how can I make sure that my code can still be compiled? Are we all gonna die? OMG! Panic! Enter into catatonic state now.

 

To use popular acronyms: This is both YAGNI and KISS.

[Uwe Raabe 1891]

11 hours ago, Nigel Thomas said:

DateUtils.IsValidDate checks if the Year is <= 9999.

The underlying reason is the Windows date format being limited to four digits 

[David Heffernan 2300]

1 hour ago, Sherlock said:

To elaborate on this correct and concise answer: One should only fix things that a really broken or occur during the life cycle of the software. You simply cannot think of all eventualities - one of them being: will there still be a Delphi compiler in the year 9999? If not, how can I make sure that my code can still be compiled? Are we all gonna die? OMG! Panic! Enter into catatonic state now.

 

To use popular acronyms: This is both YAGNI and KISS.

I definitely hope that no Delphi code is running by 9999. I'd be amazed if there were any humans left on the planet when that year comes around.

[Stefan Glienke 1905]

1 minute ago, David Heffernan said:

I'd be amazed if there were any humans left on the planet when that year comes around.

Planet of the Firemonkeys?

[Rollo62 502]

10 hours ago, Ian Branch said:

I was there for Y2K.  Not sure I will be around for Y10K... 😉

If there will be any issues in the future, I'm very sure I can fix them the day after

[David Heffernan 2300]

16 minutes ago, Stefan Glienke said:

Planet of the Firemonkeys?

Yeah, defo

[programmerdelphi2k 235]

for sure, this will be the new "visual" for "Nigel 10K"

 

[Fr0sT.Brutal 897]

19 minutes ago, Stefan Glienke said:

Planet of the Firemonkeys?

Riding Firebirds 🙂

[Alexander Elagin 143]

Quoting the aforementioned RFC 2550:

Quote

 

   There are a number date comparison problems that are beyond the scope
   of this specification.
   (...)
   3) Continued existence of Earth-centric time periods (year, day,
      etc.) are problematical past the up-coming destruction of the
      solar system (5-10 billion years or so).  The use of atomic-time
      helps some since leap seconds are no longer an issue.
   4) Future standards and methods of synchronization for inter-planetary 
      and inter-galactic time have not been agreed to. 
   5) Survivability of dates past the end of the universe is uncertain. 

 

 

[Lars Fosdal 1732]

The next Y2K like problem is already in 2038, when the 32-bit int that holds a Unix timestamp wraps to a negative number.

Aka The Epochalypse.