undefined | Better HN

0 pointsquickthrower23y ago0 comments

or 11-11-11 11:11:11

or 01-01-01 10:10:10

Interview question: write a program to list them all!

0 comments

yamtaddle3y ago

> write a program to list them all!

Me: Is it OK if it doesn't print until it's done?

Interviewer: Uh, sure, why not?

Me: while(1){}

Me: Finished. This produces the exact desired output of the program you specified.

Interviewer: GTFO.

quickthrower2OP3y ago

while(1) {} is the universal program to do anything. You can then optimise performance later.

Interestingly in Haskell the only pure program of type variable “a” where a means any type at all is the one that never quits:

    myProg: a

    myProg = myProg

Which amounts to the same thing: it can do anything - produce any type - but it takes literally forever.

carl_dr3y ago

Sounds like a great Code Golf contest.

ruby -e 'require"date";a=DateTime.new(0,1,1,0,0,0);(0..4e9).each{|d|b=(a+Rational(d,86400));c=b.strftime("%y%m%d%H%M%S");puts b.strftime("%y-%m-%d %H:%M:%S")if c==c.reverse}'

About the most inefficient anyone could make it, but I think it works!

AkshatJ273y ago

Some saves: `Rational(d,86400)` -> `d/86400r`

Get rid of the variable `a` since it is used in only one place.

`DateTime.new(0,1,1,0,0,0)` -> `DateTime.new(0,1,1)`

also in newer ruby versions you can use numbered parameters to replace d with _1 and remove `|d|`

Pretty sure there might be some more which i do not see immediately :)

carl_dr3y ago

ruby -e 'require"date";(0..4e9).each{b=DateTime.new(0,1,1)+_1/86400r;c=b.strftime("%y%m%d%H%M%S");puts b.strftime("%y-%m-%d %H:%M:%S")if c==c.reverse}'

151 bytes (including the ruby -e, down from 174.) Thanks!

1 more reply

PinkMilkshake3y ago

PowerShell 7, 108 chars

  0..2e9|%{date(date 0).AddSeconds($_)-f 'yy-MM-dd HH:mm:ss'}|?{$_-match'(.)(.)-(.)(.)-(.)(.) \6\5:\4\3:\2\1'}

lcnPylGDnU4H9OF3y ago

Looks like it would be defeated by the dreaded leap second.

SAI_Peregrinus3y ago

They're going away. Resolution 4 of the CGPM 2022[1]. They didn't want to deal with having to add a negative leap second, and it looks likely they'd have to if they didn't end them.

[1] https://www.bipm.org/documents/20126/64811223/Resolutions-20...

carl_dr3y ago

I don't think so. It would be only year 2006 (which map to second 60) which could be a problem, and there wasn't a leap year in that year.

j / k navigate · click thread line to collapse

0 comments

yamtaddle3y ago

> write a program to list them all!

Me: Is it OK if it doesn't print until it's done?

Interviewer: Uh, sure, why not?

Me: while(1){}

Me: Finished. This produces the exact desired output of the program you specified.

Interviewer: GTFO.

quickthrower2OP3y ago

while(1) {} is the universal program to do anything. You can then optimise performance later.

Interestingly in Haskell the only pure program of type variable “a” where a means any type at all is the one that never quits:

    myProg: a

    myProg = myProg

Which amounts to the same thing: it can do anything - produce any type - but it takes literally forever.

carl_dr3y ago

Sounds like a great Code Golf contest.

ruby -e 'require"date";a=DateTime.new(0,1,1,0,0,0);(0..4e9).each{|d|b=(a+Rational(d,86400));c=b.strftime("%y%m%d%H%M%S");puts b.strftime("%y-%m-%d %H:%M:%S")if c==c.reverse}'

About the most inefficient anyone could make it, but I think it works!

AkshatJ273y ago

Some saves: `Rational(d,86400)` -> `d/86400r`

Get rid of the variable `a` since it is used in only one place.

`DateTime.new(0,1,1,0,0,0)` -> `DateTime.new(0,1,1)`

also in newer ruby versions you can use numbered parameters to replace d with _1 and remove `|d|`

Pretty sure there might be some more which i do not see immediately :)

carl_dr3y ago

ruby -e 'require"date";(0..4e9).each{b=DateTime.new(0,1,1)+_1/86400r;c=b.strftime("%y%m%d%H%M%S");puts b.strftime("%y-%m-%d %H:%M:%S")if c==c.reverse}'

151 bytes (including the ruby -e, down from 174.) Thanks!

1 more reply

PinkMilkshake3y ago

PowerShell 7, 108 chars

  0..2e9|%{date(date 0).AddSeconds($_)-f 'yy-MM-dd HH:mm:ss'}|?{$_-match'(.)(.)-(.)(.)-(.)(.) \6\5:\4\3:\2\1'}

lcnPylGDnU4H9OF3y ago

Looks like it would be defeated by the dreaded leap second.

SAI_Peregrinus3y ago

They're going away. Resolution 4 of the CGPM 2022[1]. They didn't want to deal with having to add a negative leap second, and it looks likely they'd have to if they didn't end them.

[1] https://www.bipm.org/documents/20126/64811223/Resolutions-20...

carl_dr3y ago

I don't think so. It would be only year 2006 (which map to second 60) which could be a problem, and there wasn't a leap year in that year.

j / k navigate · click thread line to collapse