I did some mistakes in documenting the process mainly in counting zeros and setting length of key and position which result in inconsitency in implementation results and documented results. Finally spotted them and applied the changes in week 14.
Allowing to get storage using eth_getStorageAt
.
Doing some cleaning up and redoing some design decisions.
Accessing values of map of contracts isn't very straight forward, for example let's say there is a map of address to balance in a contract
mapping(address => uint256) balances
then to access that value, one needs to put slot number to retrieve it from storage. There are 3 params consumed by it. Out of them 2 are required i.e., contractAddress
(whose stored value here balance we want to access) and then position of variable. Let's say there is only the balances
map in the contract so it would be at position 0
. But if you just try to think for a moment then all of the map can't be stored at single position. So position of the particular value for a given key(here any wallet address) can be accessed by a specially crafted hash through series of steps given below:
Let's say I want to read balance of address 0x00000000000000000000000000000000deadBeef
stored in the contract through that balances
map.
Step 1: Left pad the address to make it of 32 bytes.
00000000000000000000000000000000000000000000000000000000deadBeef
Step 2: Left pad the position of map you want to access here it is 0
.
0000000000000000000000000000000000000000000000000000000000000000
Step 3: Concatenate them.
00000000000000000000000000000000000000000000000000000000deadBeef0000000000000000000000000000000000000000000000000000000000000000
Step 4: Keccak hash of this hex encoded value.
> web3.sha3("00000000000000000000000000000000000000000000000000000000deadBeef0000000000000000000000000000000000000000000000000000000000000000", {"encoding": "hex"})
"0x49361c85d50c86f43fdb7ff3f85f3cac00c47bda054279b6563a02b9b214ccf4"
Now this value is passed to eth_getStorageAt
in the place of position to retrieve the value i.e., balance of the address.
In the current situation, user is required to do all of this before calling the rpc. Now, ethRPCtoREST will do the heavy lifting for user allowing them to put required values and get appropriate response automatically.
/eth/storage/:address/:position?map=${key}
will internally compute that hash by consuming position
and key
. Then it will proceed to call rpc with parameters plugged appropriately.
Supports eth_getBlockTransactionCountByHash
and eth_getBlockTransactionCountByNumber
Returns latest block number through eth_getBlockNumber
Allows getting the balance of given address at given block or in the latest block by default.
Using eth_getTransactionCount
, this allows retreiving number of transactions sent from address.
I introduced a new feature - ENS name is also supported for address fields i.e., GET /eth/balance/vitalik.eth
will give back balance of the resolved address at latest block through eth_getBalance
.
Wrapper around eth_getCode
method.
Till now whole emphasis was on simplification of requests. Now it's time for response simplification.
Following are the results:
Before | After |
---|---|
![]() |
![]() |
Did you read
extraData
? By coincidence, I found an interesting block.
I also did some re-arrangement and making dev setup more easy by incorporating AIR.
I realized when I would build binary from main.go
It would also consist of RPC_URL provided by me through environment variable!
If someone wants to use their personal RPC_URL they will have to build the binary themselves. This didn't sound modular enough. Now it would consume it through command-line argument.
This week I would release first ever version of the binary. 0.0.1
Allowing use of decimal numbers wherever BlockNumber
or index
are used instead of hexadecimal numbers. commit
This basically allows request methods like
curl -X POST --data '{"jsonrpc":"2.0","method":"eth_getBlockByNumber","params":["0x1b4", false]}'
are now simplified to
GET /eth/block/0x1b4
or
GET /eth/block/436
Proposing more simplification ideas to our mentor like one example of eth_call
for interacting with contracts
Original form
curl -X POST --data '{"jsonrpc":"2.0","method":"eth_call","params":[{
"from": "0x<sender's hash>",
"to": "0x<contract hash>",
"value": "0x1234",
"gas": "0x55555",
"data": "0x<first 4 bytes of Keccak hash of ASCII function signature>+<parameters in hex each padded with prefix zeros to 32 bytes>"
}],"id":1}'
can be simplified to
GET /eth/call/from="0x<sender's hash>"&to="0x<contract hash>"&data="foo(uint32 69, bool true)"
Added support for default block parameters like "earliest", "safe" etc. parameters
Looking at few examples like web3go and description of ethclient methods. To create API, I started with eth_getBlockBy*
methods which don't return full transaction with false
parameter as mentioned in the specs but the twist is it doesn't returns all transactions even when true
is passed into it. Yay! New easter egg unlocked! Really? I also looked at ethClient package and it mentions that you would need to send request two times if you want to get all transactions and uncles.
Currently I moved forward with what is JSON RPC doing. Simply shuttling the responses as it is by making use of HeaderBy{Hash/Number}.
Following methods got covered till now, corresponding commits are linked into them:
Learning Fiber and Golang at the same time 🫣
https://dev.to/fenny/getting-started-with-fiber-36b6
"Learn GO" course from https://boot.dev
Mapping golang commands to analogous nodeJS commands in my mind.
NodeJS Universe | Golang Universe |
---|---|
npm init |
go mod init <module name> |
npm install |
go mod tidy |
node server.js |
go run main.go |
package.json & package-lock.json |
go.mod & go.sum |
Finalised Project proposal which got merged. Wrote a blog post to quantify the impact of the project and increasing awareness for the purpose of it.
Creating a presentation which was presented during office hours.
eth_getUncleByBlockHashAndIndex
and eth_getUncleByBlockNumberAndIndex
are wrapped by above api.
This one binds to eth_getTransactionReceipt
eth_getUncleCountByBlockHash
and eth_getUncleCountByBlockNumber
got covered by it.
Also proposed implementation of them which makes interacting with execution-apis
more user friendly by using a boolean flag like ?useInt
for using integers in place of hashes for parameters like Index
.
For example, currently if one wants to call eth_getTransactionByBlockHashAndIndex
then according to current specifications, request would look like following:
GET /eth/transaction/block/{32-bytes BlockHash}/0x41
Notice number mentioned at the place of Index
in above request isn't simply 41 but it's in hexadecimal and is actually 65 of decimal.
So, I wanted to provide a simple layer of easeness included in wrapper which will allow to call above mentioned call as
GET /eth/transaction/block/{identifier}/65?useInt
or
GET /eth/transaction/block/{identifier}/65?useInt=true
So, after reading stories (history) on inevitable ethereum, learning about EIPs and having a read on JSON RPCs when I was going through cohort-four repository again to see - What other things are there? At the end, the purpose is to explore and research well before diving into a specialised area. This was probably 4th or 5th time when I was re-reading project ideas already proposed and this time I felt ideas were looking a lot familiar and clear than my last visit.
And then I stumbled upon JSON RPC improvement ideas, for some reason this time Rest Wrapper made lot of sense for me. It was very clear that, given the fact - I had visited JSON RPCs recently I was able to draw a clear picture of requirements and responsibilities under the idea.
I expressed my interest to intended mentor and got started same day for REST API on OPENAPI yaml specifications.
I made first ever implementation with one API for each method like for eth_getBlockByHash
which takes two parameters one Hash
and another boolean value for whether to includeTransaction
or not.
So, I proposed API like following:
GET /eth_getBlockByHash/{32-Bytes Hash}?includeTxn={boolean}
Then our mentor suggested to rather implement a more generalised API. For example, eth_getBlockByHash
and eth_getBlockByNumber
can be merged into:
GET /eth/block/hash/{HashOrNumber}
Also speced out eth_getTransactionByBlockHashAndIndex
and eth_getTransactionByBlockNumberAndIndex
under
And eth_getTransactionByHash
by
There was lot of other stuff going behind the scenes like getting something like swagger editor which can allow working offline or locally and can give visual docs for the APIs. Downloading literally every random binary from aur and pacman
which mentions OPENAPI in its description because of the inconsistency of schema around versions or absence of a standard.
I had heard lot about ethereum whitepaper but haven't ever read it ever just because I always felt that those things are for researchers or academics purposes and are of no use for builders.
After reading I feel I was quite right. Certainly there was a good refresher of UTXO which I had last studied when learning about zk applications and trying to understand how Tornado protocol works.
However, Ethereum whitepaper is one of those resources which will give a broad and deep enough understanding of ethereum and its most intuitive applications which can be helpful from day one and are much better solution compared to current options, without getting into buzzwords in a single read.