1697 replies to this topic

[biggest yikes]

If you add some sort of Lua scripting language in KristScape, make sure you don't include _G in your sandboxing. Quest didn't do that, and it resulted in an exploit which allows malware to be downloaded on the computer.

Edited by Atenefyr, 12 June 2015 - 01:14 PM.

[クデル]

Atenefyr, on 12 June 2015 - 01:07 PM, said:

If you add some sort of Lua scripting language in KristScape, make sure you don't include _G in your sandboxing. Quest didn't do that, and it resulted in an exploit which allows malware to be downloaded on the computer.

Yeah, I would just set it up to where it runs in a completely new environment.

[3d6]

Sandstorm, on 12 June 2015 - 01:18 PM, said:

Atenefyr, on 12 June 2015 - 01:07 PM, said:

If you add some sort of Lua scripting language in KristScape, make sure you don't include _G in your sandboxing. Quest didn't do that, and it resulted in an exploit which allows malware to be downloaded on the computer.

Yeah, I would just set it up to where it runs in a completely new environment.

We may just make a new language and parse that, too.

[Yevano]

cossacksson, on 12 June 2015 - 04:41 PM, said:

We may just make a new language and parse that, too.

Why make a new language when you already have Lua, the most embeddable, sandbox-able scripting language there is?

[biggest yikes]

cossacksson, on 12 June 2015 - 04:41 PM, said:

We may just make a new language and parse that, too.

Yay, a new unnecessary language to learn!
In all seriousness, that's not really a good idea in my opinion. You have a scripting language already (Lua), no need to reinvent the wheel. Just make a sandboxed Lua environment, and you're set

Edited by Atenefyr, 12 June 2015 - 11:14 PM.

[クデル]

I have begun work on alternative browser, its more based on shopping so it might just evolve into a shop, so I dont know. But will the ?a= call only return a single record? if so, it would simply the process a lot.

[sci4me]

Yevano, on 12 June 2015 - 08:21 PM, said:

cossacksson, on 12 June 2015 - 04:41 PM, said:

We may just make a new language and parse that, too.

Why make a new language when you already have Lua, the most embeddable, sandbox-able scripting language there is?

this.

[PokeAcer]

Does anybody know the progress of this wallet?:
http://www.reddit.co...android_wallet/

I currently use https://github.com/k...ndroid/releases but I find it slightly lacking (no colourful GUI)

[クデル]

Can someone care to explain how a miner actually works in depth? like the workflow from from pressing start and getting a block. Wanting to re-work my OpenCL miner, but my mind has gone blank.

[InDieTasten]

Sandstorm, on 14 June 2015 - 07:39 AM, said:

Can someone care to explain how a miner actually works in depth? like the workflow from from pressing start and getting a block. Wanting to re-work my OpenCL miner, but my mind has gone blank.

Basically, it communicates to the krist site, what the hash of the last block was. it then takes this hash, your own krist address and some nonce(just a number), and put that into some block. You then get the hash value of that block, and if the hash value meets the requirement of the difficulty, you submit your block to the site. If not, you would increment your nonce by one and do that again, approximately 1 million times a second for the average miner. The "requirement" for your block to be valid, is that you truncate everything but the last few bytes of the hash value, and the few bytes left must then be of lower value than the current difficulty.

Thats how far I understand it. Not entirely sure where the value of the block is set, but I think that is some magic that happens at the krist site.

So basically, you change nonce until you find a hash value of (lastblock_hash,krist_address,nonce) that has a lower value than the difficulty when truncated.

For more technical stuff, like what bytes actually matter, and what bytes are truncated, you have to look into some other miners or look through pages in this forum. I once saw the process described fairly detailed.

Edited by InDieTasten, 14 June 2015 - 07:58 AM.

[クデル]

InDieTasten, on 14 June 2015 - 07:57 AM, said:

Sandstorm, on 14 June 2015 - 07:39 AM, said:

Can someone care to explain how a miner actually works in depth? like the workflow from from pressing start and getting a block. Wanting to re-work my OpenCL miner, but my mind has gone blank.

Basically, it communicates to the krist site, what the hash of the last block was. it then takes this hash, your own krist address and some nonce(just a number), and put that into some block. You then get the hash value of that block, and if the hash value meets the requirement of the difficulty, you submit your block to the site. If not, you would increment your nonce by one and do that again, approximately 1 million times a second for the average miner. The "requirement" for your block to be valid, is that you truncate everything but the last few bytes of the hash value, and the few bytes left must then be of lower value than the current difficulty.

Thats how far I understand it. Not entirely sure where the value of the block is set, but I think that is some magic that happens at the krist site.

So basically, you change nonce until you find a hash value of (lastblock_hash,krist_address,nonce) that has a lower value than the difficulty when truncated.

For more technical stuff, like what bytes actually matter, and what bytes are truncated, you have to look into some other miners or look through pages in this forum. I once saw the process described fairly detailed.

Okay, thanks a bunch.

[Tiin57]

It looks like no one has done this yet, so here is a Java 7 compilation of Yevano's Krist miner (at commit b47cec23336a062cec8c6545c9e4ad5284284b27).

By the way, this looks pretty neat. Well done, cossacksson.

[Yevano]

Tiin57, on 14 June 2015 - 12:03 PM, said:

It looks like no one has done this yet, so here is a Java 7 compilation of Yevano's Krist miner (at commit b47cec23336a062cec8c6545c9e4ad5284284b27).

By the way, this looks pretty neat. Well done, cossacksson.

Thanks a lot for doing that. If people keep using my miner, I might feel like updating it a bit.

[3d6]

Tiin57, on 14 June 2015 - 12:03 PM, said:

It looks like no one has done this yet, so here is a Java 7 compilation of Yevano's Krist miner (at commit b47cec23336a062cec8c6545c9e4ad5284284b27).

By the way, this looks pretty neat. Well done, cossacksson.

Works good

sci4me, on 13 June 2015 - 09:11 AM, said:

Yevano, on 12 June 2015 - 08:21 PM, said:

cossacksson, on 12 June 2015 - 04:41 PM, said:

We may just make a new language and parse that, too.

Why make a new language when you already have Lua, the most embeddable, sandbox-able scripting language there is?

this.

I'm just worried about possible security holes. Admittedly, it'd probably be just fine sandboxing Lua. We'll see when we get there ;P

[valithor]

cossacksson, on 14 June 2015 - 11:54 PM, said:

sci4me, on 13 June 2015 - 09:11 AM, said:

Yevano, on 12 June 2015 - 08:21 PM, said:

cossacksson, on 12 June 2015 - 04:41 PM, said:

We may just make a new language and parse that, too.

Why make a new language when you already have Lua, the most embeddable, sandbox-able scripting language there is?

this.

I'm just worried about possible security holes. Admittedly, it'd probably be just fine sandboxing Lua. We'll see when we get there ;P

Honestly I only see the functions from the term api and textutils api along with a few modified fs api functions being useful. Sandboxing it to those functions would not be very hard.

One of the main problems I see with the current way the parser works is the fact that everything runs in the order it is written, when we may want scripts to run after everything else runs.

Edited by valithor, 15 June 2015 - 12:04 AM.

[apemanzilla]

Woohoo! In some tests with a modified version of sci4me's miner, running on 6 cores @ 4.5 GHz, in Arch Linux, I managed to hit just shy of 40 MH/s on my home PC!

[クデル]

apemanzilla, on 15 June 2015 - 01:15 AM, said:

Woohoo! In some tests with a modified version of sci4me's miner, running on 6 cores @ 4.5 GHz, in Arch Linux, I managed to hit just shy of 40 MH/s on my home PC!

Holy shi. 6 core, AMD FX 6300?

[apemanzilla]

Sandstorm, on 15 June 2015 - 01:26 AM, said:

apemanzilla, on 15 June 2015 - 01:15 AM, said:

Woohoo! In some tests with a modified version of sci4me's miner, running on 6 cores @ 4.5 GHz, in Arch Linux, I managed to hit just shy of 40 MH/s on my home PC!

Holy shi. 6 core, AMD FX 6300?

Yup. It's an amazing chip for it's price.

Edited by apemanzilla, 15 June 2015 - 01:44 AM.

[クデル]

apemanzilla, on 15 June 2015 - 01:43 AM, said:

Sandstorm, on 15 June 2015 - 01:26 AM, said:

apemanzilla, on 15 June 2015 - 01:15 AM, said:

Woohoo! In some tests with a modified version of sci4me's miner, running on 6 cores @ 4.5 GHz, in Arch Linux, I managed to hit just shy of 40 MH/s on my home PC!

Holy shi. 6 core, AMD FX 6300?

Yup. It's an amazing chip for it's price.

Awesome, I was considering this for a new rig, im on a budget.

[apemanzilla]

Sandstorm, on 15 June 2015 - 01:47 AM, said:

apemanzilla, on 15 June 2015 - 01:43 AM, said:

Sandstorm, on 15 June 2015 - 01:26 AM, said:

apemanzilla, on 15 June 2015 - 01:15 AM, said:

Woohoo! In some tests with a modified version of sci4me's miner, running on 6 cores @ 4.5 GHz, in Arch Linux, I managed to hit just shy of 40 MH/s on my home PC!

Holy shi. 6 core, AMD FX 6300?

Yup. It's an amazing chip for it's price.

Awesome, I was considering this for a new rig, im on a budget.

Best price for performance I've ever seen - go for it.