Where communities thrive

Join over 1.5M+ people
Join over 100K+ communities
Free without limits
Create your own community

Explore more communities

CodeChain-io/codechain

Programmable multi-asset chain

People

Repo info

Activity

Mar 20 06:28

MSNTCS opened #1962
Mar 20 06:27

MSNTCS commented #1961
Mar 20 06:25

MSNTCS closed #1961
Mar 19 08:02

MSNTCS review_requested #1961
Mar 19 08:02

MSNTCS opened #1961
Mar 03 02:37

ScarletBlue edited #1948
Mar 03 02:37

ScarletBlue edited #1931
Mar 02 00:41

ScarletBlue edited #1915
Mar 02 00:41

ScarletBlue edited #1915
Feb 27 00:09

sgkim126 closed #1959
Feb 27 00:09

sgkim126 on master

Bug fix in on_request_commit_me… (compare)
Feb 25 09:41

majecty labeled #1960
Feb 25 09:41

majecty opened #1960
Feb 17 02:10

majecty commented #1959
Feb 14 23:13

sgkim126 review_requested #1959
Feb 14 23:13

sgkim126 review_requested #1959
Feb 14 23:13

sgkim126 opened #1959
Feb 14 23:13

sgkim126 labeled #1959
Feb 14 06:30

majecty closed #1911
Feb 14 06:30

majecty commented #1911

Park Juhyung

@majecty

Yes, let's talk about how to change current ics-poc branch not to generic today.

In the connection spec, they are saying, "Each chain has knowledge of and has agreed to its identifier on the other chain."
We might need some authorization logic to make only allowed users be able to create connections.

junha1

@junha1

Then how about just having that (set of identifiers) in state since the Genesis?

Byeongjee Kang

@Byeongjee

I'm implementing BLS signature aggregation scheme in foundry. Introducing new hash types such as H384 and H768 will be really helpful. Is there any nice way to do this?
Currently I forked rust-codechain-primitives repository and added the new types with fixed_hash library. However I had to copy impl_serde macro from ethereum-types crate. There is impl_serde crate, but I couldn't make use of it because ethereum-types library uses a very old version of fixed_hash library, so they were not compatible.
I also had to fork and add code in codechain rlp to give the new types Encodable/Decodable traits.

Park Juhyung

@majecty

Then how about just having that (set of identifiers) in state since the Genesis?

Maybe that can be a solution. I'll create a GitHub issue about this.

Byeongjee Kang

@Byeongjee

Anyone know about history of parameter names in execute method of ActionHandler trait?

I think the name 'fee_payer' makes it hard to understand. The address can be read as fee payer for money transfer, but I think it should be 'deposit_owner' in self nomination and 'informant' in double vote reporting.

Does the fee_payer actually pay fee? If not, I will make a PR to change its name to more understandable one, such as 'sender_address'.

Park Juhyung

@majecty

Here is the history. In CodeChain, anyone can create assets and give them to anyone. When transferring an asset, the asset owner and a fee payer should sign the transaction. So 'send_address' is a confusing name.
In Foundry, since we don't have a native asset system (anyone can create asset using module), we can use the "sender_address"

junha1

@junha1

How about getting ibc directory out of core? And we can make it as a separate crate. @majecty

Park Juhyung

@majecty

@junha1 I agree with you.
It seems we don't need to hurry. How about moving it after PoC implementation?

junha1

@junha1

No problem

Park Juhyung

@majecty

@MSNTCS Please add the yarn.lock file in this PR: CodeChain-io/foundry#130
Let's check the CI tests after uploading it.

junha1

@junha1

Do you think that user interface implementer will directly access to host to get information? Or they should all request something to some module? @majecty

Park Juhyung

@majecty

The UI application implementer will use RPC to Foundry to get information. Foundry will query to a module about the information.

@junha1 Could you check these two PRs? I want to merge them today.
CodeChain-io/foundry#157
CodeChain-io/foundry#161

Mohammadsadegh Najafi

@MSNTCS

@majecty I add the changed yarn.lock. We should wait for the CI result.

Park Juhyung

@majecty

👍

junha1

@junha1

Me and @majecty think that

stand alone light client will be likely to be embedded in host
All each module will be loaded regardless of host's client mode
All each module is agnostic of client mode.
But not finalized about this idea

junha1

@junha1

What do you think of having some categorization of RPCs about light client support? Not all but at least gradually as we add some new one. @majecty

Mohammadsadegh Najafi

@MSNTCS

@majecty in foundry-rpc, chain.getBlockbyHash and chain.getBlockbyNumber return Block type. This Block type has not been defined as an export type. I have a plan to change foundry-rpc to export it.

Park Juhyung

@majecty

@MSNTCS OK.

@junha1 I think it is not a good time to do it. Since our team is modifying a lot of foundry code, it can change many RPCs. Let's do it after Foundry is stabilized.

junha1

@junha1

Park Juhyung

@majecty

It seems that there is a bug in the RlpEncodable and RlpDecodable macro.

junha1

@junha1

How?

Park Juhyung

@majecty

When I used the two proc macro in a struct, It failed to pass rlp_encode_and_decode_test test. I'm figuring out it now.

Mohammadsadegh Najafi

@MSNTCS

@majecty error Incorrect hash when fetching from the cache for foundry-rpc. I cannot find the source of problem.

Park Juhyung

@majecty

@MSNTCS I'll check it now.

Mohammadsadegh Najafi

@MSNTCS

@majecty thanks

Park Juhyung

@majecty

@MSNTCS Could you try again after removing this line?
https://github.com/CodeChain-io/foundry/blob/b20fe1304b35efffd363e13db8157848c77f3216/.travis.yml#L98
It seems that Travis' cache is not working well.

Mohammadsadegh Najafi

@MSNTCS

Alright

Should I remove it and push it as changes in PR?

Park Juhyung

@majecty

Yes. Create a new commit for it, and include it in your PR.

I found that we can remove cache in Travis. I'll try it.

Park Juhyung

@majecty

https://blog.travis-ci.com/2016-05-03-caches-are-coming-to-everyone#managing-caches

Park Juhyung

@majecty

When I used the two proc macro in a struct, It failed to pass rlp_encode_and_decode_test test. I'm figuring out it now.

It was my fault. 😭

junha1

@junha1

What was the reason? I'm curious about that

Park Juhyung

@majecty

I used append instead of append_single_value

junha1

@junha1

I'll remove foundry directory and kind for that. This poc is not considered to support other chain. @majecty

Park Juhyung

@majecty

OK.

Mohammadsadegh Najafi

@MSNTCS

@majecty I think, I have to update codechain-primitives to use some of its function like decode, fromWord. What is your idea?

junha1

@junha1

To 0.2?

Park Juhyung

@majecty

@MSNTCS Do you mean codechain-primitives-js or rust-codechain-primitives?

Mohammadsadegh Najafi

@MSNTCS

@majecty codechain-primitives-js

Park Juhyung

@majecty

Let me summarize. You need decode and fromWord functions in the Foundry tests. So, you want to use the recent version of codechain-primitive-js in Foundry test code.
It looks good to me. Go ahead.

@Byeongjee Could you give the link to the code you showed me before?

Mohammadsadegh Najafi

@MSNTCS

Let me summarize. You need decode and fromWord functions in the Foundry tests. So, you want to use the recent version of codechain-primitive-js in Foundry test code.
It looks good to me. Go ahead.

Yes, I need to add some exports in Index.ts. Do I need to change the version of the package and publish another version? or use a branch in my fork repository?

Byeongjee Kang

@Byeongjee

@majecty https://github.com/Byeongjee/rust-codechain-primitives/blob/master/src/hash.rs This is my implementation of serde, and other types are declared in ethereum-types package https://docs.rs/crate/ethereum-types/0.4.0/source/src/hash.rs . I copy-and-pasted impl_serde macro from ethereum-types to rust-codechain-primitives, so that may cause the error, but I'm not sure.

junha1

@junha1

Can I have a definition of action_data in general? It confuses me..

Park Juhyung

@majecty

@junha1 action_data is used by custom action, which will be removed soon by Seulgi.

junha1

@junha1

Ah thanks. It appears in many places.