World: r3wp

>> age: 28-Dec-1923
== 28-Dec-1923
>> difference now age
** Math Error: Math or number overflow
** Near: difference now age

It appears to be a problem with the +-50 year window around the current 
year which Rebol uses to assign the century to two-digit years:

>> date-of-birth: 28-dec-1923
== 28-Dec-1923
>> age: func [dob [Date!] /local cutoff] [
[    cutoff: now
[    cutoff/year: cutoff/year - 50
[    either dob < cutoff [
[        return (now - cutoff) + ((cutoff - 1) - dob)][
[        return now - dob]
[    ]
>> age date-of-birth
== 30262

The calculation of the cutoff is not quite correct as on checking 
I found that Rebol sets the 50 year window at the start of the current 
year not from the current date.

>> date-of-birth: 28-dec-1923
== 28-Dec-1923
>> age: func [dob [date!] /local cutoff] [
[    cutoff: 1-1-06
[    cutoff/year: now/year - 50
[    either dob < cutoff
[    [return (now - cutoff) + ((cutoff - 1) - dob)]
[    [return now -dob]
[    ]
>> age date-of-birth
== 30262

should this be rambo'd ?

If not in Rambo already, yes.

graham, i think the problem there is that time! values cannot hold 
that much hours.

>> now - 28-dec-1923
== 30263

(days)

>> difference now 1-nov-1938
== 596192:12:37

there's probably not much more room than some 600k hours.

When I did a date-to-epoch func, I used ATTEMPT with DIFFERENCE and, 
if that failed, used subtraction to get the number of days, then 
multiplied by 86400.0.

time fot 64 bits ints.... and people asked... do we need them?

I'm not sure this issue means we need 64-bit ints.

by all accounts, time is expressed in seconds internally .  and guess 
what! 600000 hours is 2.16 Gs..  just above 2^31 (2.14 Gs)..   so 
if we had 64 bits, then we could hold the date and calculate it with 
other values.

now/precise

precise add floating point sub seconds.

I think the minimum time unit is milliseconds - thousandths of a 
second.

Is time quantisized? ;-)

Anton, that might be right under Windows. I think, under UNIX (Linux, 
OS X, etc.) the minimum time unit is less than that. Under OS X:
>> now/time/precise
== 17:28:10.349125

That's interesting. I don't know about OS X, but that *could* be 
just the rebol OS X beta molding of the seconds floating point number.

time really is quantized depending on the OS. XP SP 2 has got a bigger 
quantum than XP SP 1, which was 10 milliseconds IIRC

Linux has much more precise timer (or it just simply returns more 
digits for now/time/precise)

not just more digits, it has got a "lower basic quantum" too

Max, what I mean is that this isn't a showstopper. Adding 64-bit 
ints for this case doesn't seem worth it. There are other, more important, 
reasons to go there; I just don't think this issue is that important.

OUCH !!! to-integer error! 

>> to-integer "2147483648"
==  2147483648.0

@ Gregg: well it depends on what you are doing.  any serious date 
management will eventually break if its trying to calculate anything 
remotely old  :-(


having to work around something as basic like this is anoying.. (although 
like you say, possible)

I just ramboed my discovery, as I did not find any reference to it 
in searching to-integer

this is a very dangerous bug for any dialect writer... beware!

to-integer - what do you want to get?

an error! obviously.

It will only break if you try to do it this way. I haven't heard 
of many places where this is an issue, and it isn't hard to work 
around if you really need to.

if I try to use the decimal and converting it to integer it will 
return the error.

to-integer  2147483648.0
** Math Error: Math or number overflow
** Where: to-integer
** Near: to integer! :value

so we must not allow different type! representation of a value be 
considered different through the most basic type converting natives.

its an issue when you load data and make a dialect loader.  It just 
fucked up a week's work ... now I have to completely change the way 
I handle Integers...  had this broken early on, it would not even 
have slown me down.

:-(

I have to agree that this is not correct behavior.

the data is not of REBOL origin.  and has to be spitted out the same, 
cause it will break the external tool's understanding of the value...

What's the actual issue? I agree that it's not correct, but maybe 
there's an easier way to solve your current problem.

Or maybe we can all learn what not to do if we try something similar.

I can work around it (obviously).. but its just a core REBOL consistency 
issue.


its semantically unacceptable for a function called  to-INTEGER  
 to return anything else than an integer.  imagine if to-string decided 
to return words sometimes... its the same dangerous issue.

many places this can be a total fuck up
imagine this:

as-integer: func [value][
	any [
		attempt [to-integer value]
		0
	]
]


here we end up with a broken logic and impending doom...not of our 
fault.

so for now, we should all do this IMHO:

to-integer to-integer value

how could that help you?

it raises the error overflow error.

the way to-integer is implemented right now it should be called
to-number

to-integer is consistent as long as the value being given to it is 
within the bounds of 32-bit integer; outside that range, all bets 
are off. Your solution isn't going to help you, or catch other errors, 
e.g.:

>> to-integer to-integer "2147483646344.0"
== -1656
>> to-integer to-integer "2147483648343.0"
== 343

Only Carl can say why he designed it this way or, indeed, if it's 
intentional

my god this horror is worse than I thought !

thanks for catching the 32 bit rollover

I don't know if I'd call it "horror", though it's obviously not what 
some people will expect. It just means having to deal with those 
cases differently; use LOAD or TO DECIMAL! and check the range, etc.