Generator for pageTokens for the Youtube API

pageTokenGen.js

// workaround for https://code.google.com/p/gdata-issues/issues/detail?id=7163

var d0 = "AEIMQUYcgkosw048";
var d1 = "ABCDEFGHIJKLMNOPQRSTUVWXYZ";
var d2 = d1;
var d1c = 0;
var d2c = 0;
var overflowSuffix = "Q";
var direction = "AA";
var d2OverflowCounter = 0;
var pageSize = 50;

for(i = 0; i < 1024; i++) {
	if(i % pageSize == 0) console.log("C" + d1.charAt((d1c / d0.length) % d1.length) + d0.charAt(i % d0.length) + overflowSuffix + direction, ":", i);
	if(++d1c % (1 << 8) == 0) d1c = 1 << 7;
	if(++d2c % (1 << 7) == 0) overflowSuffix = d2.charAt(++d2OverflowCounter) + "E";
}

Hum... interesting!
How did you find the way tokens work?

by looking at a couple of valid ones and guessing how the algorithm might work.

This would be nice in Python with page and max_result as a parameter :)

i=400 isn't working. Token should be CJADEAA but calculates CJAAEAA

pageSize = 50
i = 400

Further tests reviled that var d1 = "ABCDEFGHIJKLMNOP"; should be correct. After P the iteration starts with I again. So d1 has only a length of 16.

@bromix: the sequence is calculated correctly on my end, just pasted the code into the chrome dev console. gives me CJADEAA : 400. Please keep in mind that the algorithm is not very smart and requires to iterate beginning with 0.

And yes, you are right. d1 might be trimmed to [A..P] but d2 may not. I way lazy. Feel free to fork ;)

@Pulsar: thanks for the feedback...already figured it out for python. But your code was inspiring and a lot of help :) no need for iterations on my end now :)

https://gist.github.com/bromix/d159787a63f9d7b00036

awesome! ;)

even with the code examples above, i seem to get stuck at 46 pages (with a max of 50 per page). Can you get more than 46 pages returned?

I got stuck at page 22, after 1000 position. But it may be a mostake on my porting.@chuckstarl can you give a list of codes until 46?