-
-
Save atk/1053863 to your computer and use it in GitHub Desktop.
// Polyfill for Date.parse | |
Date.parse = Date.parse || function( | |
a // ISO Date string | |
){ | |
// turn into array, cutting the first character of the Month | |
a = a.split(/\W\D?/); | |
// create a new date object | |
return new Date( | |
// year | |
a[3], | |
// month (starting with zero) | |
// we got only the second and third character, so we find it in a string | |
// Jan => an => 0, Feb => eb => 1, ... | |
"anebarprayunulugepctovec".search(a[1]) / 2, | |
// day | |
a[2], | |
// hour | |
a[4], | |
// minute | |
a[5], | |
// second | |
a[6] | |
) | |
} |
Date.parse=Date.parse||function(a){a=a.split(/\W\D?/);return new Date(a[3],"anebarprayunulugepctovec".search(a[1])/2,a[2],a[4],a[5],a[6])} |
DO WHAT THE FUCK YOU WANT TO PUBLIC LICENSE | |
Version 2, December 2004 | |
Copyright (C) 2011 Alex Kloss <alexthkloss@web.de> | |
Everyone is permitted to copy and distribute verbatim or modified | |
copies of this license document, and changing it is allowed as long | |
as the name is changed. | |
DO WHAT THE FUCK YOU WANT TO PUBLIC LICENSE | |
TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION | |
0. You just DO WHAT THE FUCK YOU WANT TO. |
{ | |
"name": "parsedate", | |
"description": "Date.parse replacement", | |
"keywords": [ | |
"Date", | |
"parse" | |
] | |
} |
<!DOCTYPE html> | |
<title>Foo</title> | |
<div>Expected value: <b>Thu Mar 22 2007 12:56:06...</b></div> | |
<div>Actual value: <b id="ret"></b></div> | |
<script> | |
var myFunction = Date.parse=Date.parse||function(a){a=a.split(/\W\D?/);return new Date(a[3],"anebarprayunulugepctovec".search(a[1])/2,a[2],a[4],a[5],a[6])} | |
document.getElementById( "ret" ).innerHTML = myFunction(''+new Date(2007,2,22,12,56,6)); | |
</script> |
would still love to nail the timezone issue, i think it's more important than the assignment.
Problems to solve for timezone:
- we lose the +/- of the timezone direction in the splitting atm
- we need to compare to the current timezone :-(
especially the latter will be too long for the last 25 bytes.
good points, @atk.
is there any potential in a function(a){return new Date(a.replace(...))}
approach?
Alas, not. Either would it need a regexp to parse the whole string at once, which is 53 bytes long, and/or it would need a callback which resolves the month name, too, a part which currently takes amazing 41 bytes, add the 2 function statements and the new Date() and you're well over 140 bytes. It would look something like this:
function(a){return new Date(a.replace(/\w+ \w(\w+) (\d+) (\d+) (\d+):(\d+):(\d+) \w+(.\d+)/,function(a){a=arguments;[a[0],a[2]]=[a[2],a[0]];a[1]="anebarapayunulugepctovec".search(a[1])/2;return a.join(' ')}))}
Which does not yet take care of the timezone, btw.
i dunno, there's a bunch of stuff i thought was impossible before 140 bytes kicked off. what about something structured like this to reuse functions?
function a(b,c,d,e,f,g){return c?XXXXX:b.replace(/YYYYYY/,a)}
nice idea - I will look into it later.
Shouldn't April be "pr"?
Yes, you are right. Changed.
@jed: I just found out that new Date([string]) internally calls Date.parse - so if it was not available anyway, we could only use eval for call will not work on new Date, thus enlargening the function quite a bit. Anyway, I will try to get Timezone handling inside this later.
Maybe this will help somehow for timezones:
var timezone = -7;
var d = new Date();
d.setUTCFullYear(2011, 6, 1) /* 1 Jul 2011 */
d.setUTCHours(1 + timezone, 0, 0) /* 01:00:00 UTC -7 hours */
console.log(d.toUTCString());
// Thu, 30 Jun 2011 18:00:00 GMT
Works well in IE6+, Fx, Op, node.js
currently, I can cope for the given Date's timezone:
function(a){a=a.split(/[^\w-]\D?/);return new Date(a[3],"anebarprayunulugepctovec".search(a[1])/2,a[2],a[4]+(0|a[7].slice(2,5)),a[5],a[6])}
This doesn't take care of the local timezone, though - and I already got 139 bytes without the polyfill prefix.
Date.parse
has been around since ES1 making a fallback seems a bit out of place. Also the workings of ES5/5.1 Date.parse are a bit too complex to fit in 140bytes.
Great idea - and since we haven't got special chars here, we can use search instead of indexOf, too.