WTB Updater.

[Anarchy]

3 minutes ago, Rootware said:

 

If you know the better method for calculate hash for file with less chance of collision than MD5 i will glad to use it. otherwise you confirm what your application have a bit more chances for got collisions and will skips in perspective for update some client files.

bro, it's an updater not a crypto application wtf

file hashing doesn't need to be secure if you're only checking if a file is out of date or not, updaters aren't implementing forced file validation to stop people doing ini hacks so why would u care this much?

 

use adler32 from zlib and done, fastest you'll find with relatively low collision rate for any file you're gonna encounter in l2, jesus christ guys

[xxdem]

1 minute ago, Rootware said:

 

I spoke about probability. The more files, the greater the likelihood of a collision. Still LMAO?

 

HAHAHAHAHAHAHAH OKAY MATE UR A TROL

[Rootware]

2 minutes ago, xxdem said:

 

You don't care for safety, you care only for speed. You are not creating a hashmap here, hashes are individual for each file on the updater.

 

I know that. You didn't understood me.

 

Still await your video with 2-3 seconds for 2000 files.

[xxdem]

2 minutes ago, Anarchy said:

bro, it's an updater not a crypto application wtf

file hashing doesn't need to be secure if you're only checking if a file is out of date or not, updaters aren't implementing forced file validation to stop people doing ini hacks so why would u care this much?

 

use adler32 from zlib and done, fastest you'll find with relatively low collision rate for any file you're gonna encounter in l2, jesus christ guys

 

You still don't care if itemname-e collides with weapongrp-e or whatever else.

If they both have a hash of 1234 you don't care and it doesn't break the update process because they sre different files not the same file with different version

 

Now if itemname gets changes and has AGAIN hash 1234 this breaks the updater, but the chance for a file to collide itself on another version is zero.

 

Confirm this if you understand what im talking about

[BabayG]

U2 need a real translator. Google is shit. Pazalysta

[Rootware]

9 minutes ago, Anarchy said:

bro, it's an updater not a crypto application wtf

file hashing doesn't need to be secure if you're only checking if a file is out of date or not, updaters aren't implementing forced file validation to stop people doing ini hacks so why would u care this much?

 

use adler32 from zlib and done, fastest you'll find with relatively low collision rate for any file you're gonna encounter in l2, jesus christ guys

 

Thanks for tip. I will check it.

[Anarchy]

3 minutes ago, xxdem said:

 

You still don't care if itemname-e collides with weapongrp-e or whatever else.

If they both have a hash of 1234 you don't care and it doesn't break the update process because they sre different files not the same file with different version

 

Now if itemname gets changes and has AGAIN hash 1234 this breaks the updater, but the chance for a file to collide itself on another version is zero.

 

Confirm this if you understand what im talking about

i just don't get this obsession with overengineering a god damn updater :D it's just not that critical guys, if it works no one gives a shit players aint gonna notice the difference between secure vs not secure damn hash algo wtf

[Rootware]

5 minutes ago, xxdem said:

 

You still don't care if itemname-e collides with weapongrp-e or whatever else.

If they both have a hash of 1234 you don't care and it doesn't break the update process because they sre different files not the same file with different version

 

Now if itemname gets changes and has AGAIN hash 1234 this breaks the updater, but the chance for a file to collide itself on another version is zero.

 

Confirm this if you understand what im talking about

 

I'm confirm what you meant. But i told about - more files, more chance of potencial collision for each. If you have changes only for system folder than no matter what the hash method you are used.

[Anarchy]

4 minutes ago, Rootware said:

 

I'm confirm what you meant. But i told about - more files, more chance of potencial collision for each. If you have changes only for system folder than no matter what the hash method you are used.

dude you don't understand what he's saying, you dont JUST check hash, you're not comparing hash of itemname-e.dat to LineageMonsters5.ukx so collision even between files doesn't matter

[Rootware]

Just now, Anarchy said:

dude you don't understand what he's saying, you dont JUST check hash, you're not comparing hash of itemname-e.dat to LineageMonsters5.ukx so collision even between files doesn't matter

 

Yes. This was implied in my words.

[Anarchy]

1 minute ago, Rootware said:

 

Yes. This was implied in my words.

except you're still saying "yes if only system" lol 

[xxdem]

3 minutes ago, Anarchy said:

dude you don't understand what he's saying, you dont JUST check hash, you're not comparing hash of itemname-e.dat to LineageMonsters5.ukx so collision even between files doesn't matter

 

For gods sake I was really starting to believe that I am retarded

[Rootware]

Just now, Anarchy said:

except you're still saying "yes if only system" lol 

 

You don't understand what than more files than the chance increses. Ofcourse if you edit only system folder files, you are right. I'm edit many files from client. And it's not in system folder only. From all folders. And for me important safety check.

[Anarchy]

i'm just gonna face-desk and leave the topic

 

guys updaters just aren't that complicated that's the last ima say on it :D