[BAILOPAN]

Lt.RAT: Deleting your post since I don't want to encourage this. I'll just say, Lysis is designed to be able to defeat obfuscation, and if there is something Lysis cannot handle, SourceMod should just move to requiring source code to load plugins.

[Lt.RAT]

Then add whole source code in smx file, hash it etc. But also breakable
But it`s still simpler then writing decompiler, which could be avoided by ways in my deleted post.

Sometimes better to hide smth, than show it (like with my post ), and it`s bad that we can`t hide and secure some information in our code (when we really need it). All could be avoided and hacked, but more complicate way to hack - better, and with decompiler it would be too simple

[Despirator]

need add support for switch(integer)

[Despirator]

Originally Posted by Starbish 3. Or, other etc secure Plugin like 'Korean World Ban', or included IP, MySQL ID/PASS in .sp

it's not too hard to use socket, curl or steamtools extension and make web php script which will check a player's ban

Example: mysite.com/check.php?steam=STEAM_0:0:1234
banned

[necavi]

That's pretty much what I was thinking, or if inserting is needed, then give each server a unique ID/pass or something similar? So if they abuse the system they get revoked

Also: If it is decided to to move to requiring only scripting files, how exactly would it be done? A straight compile on server start? (I would likely prefer that, as there is a good chance that it would be faster)

[Starbish]

Originally Posted by Despirator it's not too hard to use socket, curl or steamtools extension and make web php script which will check a player's ban Example: mysite.com/check.php?steam=STEAM_0:0:1234 banned

but as i know, it needs main server.
so, anyway, Main server must be exposured...;;

[BAILOPAN]

necavi: yes. not sure whether it'd be faster to load (parsing is expensive), but it'd be a chance to overhaul how compilation works and greatly improve the performance of generated code. i was unable to do this in the existing JIT because the binary format is too low-level, the instructions in it defy optimization.

a good deal of the expense could be mitigated by parsing includes once and sharing the results across all .sp compilations.

[GoD-Tony]

Another nice use-case of Lysis would be binary diffs. Many plugins are uploaded with separate .sp and .smx files because it doesn't compile on the forum. Nothing really stops a user from uploading an .smx that's different than the posted source.

A self-compiled version of a plugin should be equal to the uploaded .smx, right? Or could they be different depending on how they were compiled? Would a simple file hash be a better idea?

[TnTSCS]

ya, I always wondered about that... could the web compiler include sdkhooks, colors, smlib, and a few other includes that are most popular? That way, it wouldn't require the posting of the .smx at all.

Or, if someone is to post something that cannot be compiled on the web compiler, they upload a zip file with all the necessary includes in the appropriate folder with cstrike being the home folder, and when uploaded (with a check box in a "cannot compile on web" the attachment service includes the compile.exe in the zip, that way, people can compile themselves?

just thinking out loud - throwing out ideas... maybe one of them will spark a good one

[Fyren]

Originally Posted by GoD-Tony A self-compiled version of a plugin should be equal to the uploaded .smx, right? Or could they be different depending on how they were compiled? Would a simple file hash be a better idea?

There's metadata in the plugins (like dates, off the top of my head) that will not make the binaries match. Parts of the binaries are also gzipped, but I don't know if the metadata is. If it is, it could make the entire plugin look different due to the compression.