Google PageRank Checksum Algorithm

[doctorow]

attached file for alex

[cyberax]

Hi Guys,

I am new to this forum and was excited to see that people are really working together to find the PHP version of the google checksum algo.

I would also be part of this joint effort and would also love to see the working PHP version.

By far and large I know about the checksum calculation is that it is using Hash Function and instead of using the lengthy process if we could translate the Hash Function by Bob Jenkins at : http://burtleburtle.net/bob/c/lookup2.c to PHP it would really work good for us.

The Process:

GOOGLE_MAGIC = 0xe6359a60

The functions that would be required are: mix(a,b,c) and ub4 hash( k, length, initval)

would be called as: hash(url, strlen(url), GOOGLE_MAGIC));

We need to encode the url first.

The C version of the urlencode function is as follows:

void urlencode(char *u, char *e) {
while (*u != '\0') {
if (*u == '.') {
*e++ = '%';
*e++ = '2';
*e = 'E';
} else if (*u == ':') {
*e++ = '%';
*e++ = '3';
*e = 'A';
} else if (*u == '/') {
*e++ = '%';
*e++ = '2';
*e = 'F';
} else {
*e = *u;
}
u++;
e++;
}
}

I am already working on a C version of the same. And I hope it would be complete soon.

I hope this will help.

Cheers,
CyberaX

[cyberax]

Someone who have earlier hacked the Google Toolbar suggested we must also add "info:" before hashing the URL

Cheers,
Cyberax

[cyberax]

If I am not wrong then please also try to check the CH value with the following string:

$url = 'info%3Awww%2Edomb%2Ecam';

please revert back on the same.

Regards,

Originally Posted by cyberax

By far and large I know about the checksum calculation is that it is using Hash Function and instead of using the lengthy process if we could translate the Hash Function by Bob Jenkins at : http://burtleburtle.net/bob/c/lookup2.c to PHP it would really work good for us.

I am already working on a C version of the same. And I hope it would be complete soon.

i also ran across that code when i search for the "0x9e3779b9" but couldn't get it to work before. then when i saw your post i took another hack at it and it works. no code needed changing, just needed to prepend 'info:'. encoding the url breaks it.

[cyberax]

Hey thats gr8 guys

You all are too good

[cyberax]

So now can anyone translate it into a full working PHP code and post it on this forum ?

Try to make the function something like GetGoogleCH("www.cyberax.net") which should return the 11 digit CH value.

Looking forward to see that code soon :cool

Cheers
Cyberax

[doctorow]

That is great news... the php script Alex and I were working on will work soon, but of course to have a clean c version of the hash function makes it easier and prettier at the end.

cyberax, how did you find out that the hash function is the one by this Bob Jenkins?

[doctorow]

Originally Posted by rayg

i also ran across that code when i search for the "0x9e3779b9" but couldn't get it to work before. then when i saw your post i took another hack at it and it works. no code needed changing, just needed to prepend 'info:'. encoding the url breaks it.

Ahhh... perhaps I should have googled more

[alexstapleton]

Is that code THE googlechecksum though? It is substanitally different from the google code if im reading it right. In particular the values used for shifting

[alexstapleton]

we still have the slight problem of it producing minus numbers when it shouldnt too. today i will work converting the ASM code to inline MASM assmebly for writing a PHP module

[cyberax]

Probably I had an intuition about the same

Cyberax

[alexstapleton]

do you want to post your C source for the hash function?

[alexstapleton]

thanks. i shal work on turning it into php right now.

[cyberax]

hi alex.. we are all behind you