Scroll evm executor/v35 #1

lispc · 2024-04-22T11:21:06Z

No description provided.

# Conflicts: # Cargo.lock # crates/precompile/Cargo.toml # crates/primitives/Cargo.toml # crates/primitives/src/env.rs # crates/primitives/src/state.rs # crates/primitives/src/utilities.rs # crates/revm/Cargo.toml # crates/revm/src/db/in_memory_db.rs

# Conflicts: # Cargo.lock # crates/precompile/Cargo.toml # crates/primitives/src/env.rs # crates/primitives/src/state.rs # crates/primitives/src/utilities.rs # crates/revm/Cargo.toml # crates/revm/src/db/states/cache_account.rs # crates/revm/src/evm_impl.rs # crates/revm/src/handler.rs # crates/revm/src/handler/mainnet.rs # crates/revm/src/journaled_state.rs

# Conflicts: # Cargo.lock # crates/interpreter/Cargo.toml # crates/precompile/Cargo.toml # crates/primitives/Cargo.toml # crates/primitives/src/env/handler_cfg.rs # crates/revm/Cargo.toml

# Conflicts: # Cargo.lock # crates/precompile/Cargo.toml # crates/primitives/Cargo.toml # crates/primitives/src/env.rs # crates/revm/Cargo.toml

# Conflicts: # crates/primitives/src/env/handler_cfg.rs

# Conflicts: # crates/revm/src/context.rs

# Conflicts: # crates/revm/src/handler.rs

lispc · 2024-04-22T11:24:43Z

crates/precompile/src/blake2.rs

+#[cfg(feature = "scroll")]
+pub const BERNOULLI: PrecompileWithAddress = PrecompileWithAddress(
+    crate::u64_to_address(9),
+    Precompile::Standard(|_input: &Bytes, _gas_limit: u64| Err(PrecompileError::OutOfGas)),


not sure whether use "PrecompileError::OutOfGas" is good. let us recheck later. Err(PrecompileError::NotImplemented)?

lispc · 2024-04-22T11:26:52Z

crates/primitives/Cargo.toml

@@ -41,6 +41,10 @@ serde = { version = "1.0", default-features = false, features = [
    "rc",
 ], optional = true }

+# scroll
+halo2_proofs = { git = "https://github.com/scroll-tech/halo2.git", branch = "v1.1", optional = true }
+hash-circuit = { package = "poseidon-circuit", git = "https://github.com/scroll-tech/poseidon-circuit.git", branch = "scroll-dev-1201", optional = true }


branch = "main"

hash-circuit or halo2_proofs?

hash-circuit

lispc · 2024-04-22T11:32:30Z

crates/primitives/src/specification.rs

+    GRAY_GLACIER = 14,
+    MERGE = 15,
+    SHANGHAI = 16,
+    BERNOULLI = 17,


lispc · 2024-04-22T11:33:46Z

crates/primitives/src/state.rs

+        #[cfg(all(debug_assertions, feature = "scroll"))]
+        if eq {
+            assert_eq!(self.keccak_code_hash, other.keccak_code_hash);
+        }


great code!

very clear, congrats

lispc · 2024-04-22T11:40:30Z

Cargo.toml

@@ -22,3 +22,6 @@ debug = true
 [profile.ethtests]
 inherits = "test"
 opt-level = 3
+
+[patch."https://github.com/privacy-scaling-explorations/halo2.git"]
+halo2_proofs = { git = "https://github.com/scroll-tech/halo2.git", branch = "v1.1" }


hahah so we need this scroll-tech/poseidon-circuit#28, then we can only use the poseidon primitive crate, no longer needs halo2_proofs, quite confusing.

re-export peer dependency is a good pattern

i mean we should get rid of halo2_proofs in future. How to do this? Split poseidon-circuit crate into 2, one is focusing on hashing computation, not need of halo2, one is focusing on hahsing circuit, need halo2. Here we only need the former

Re-exporting peer dependencies might seem like a shortcut to a clean and concise codebase. After all, it keeps your imports tidy and avoids users from wrestling with dependency conflicts. But this convenience comes at a hidden cost, potentially leading you down a path of dependency management nightmares.

Loss of Version Control: Peer dependencies are meant to be managed by the consumer, not the provider. By re-exporting, you're taking control over a dependency version that another package might need a specific version of. This can lead to conflicts and unexpected behavior when multiple packages using your library have different version requirements for the peer dependency.
Imagine this: Your library re-exports a peer dependency, say mathjs, at version 5.0.0. Package A, which uses your library, needs mathjs at version 4.2.1 for specific functionality. Package B, also using your library, needs the latest mathjs (version 6.0.0). This creates a conflict. Package A might break because it relies on features removed in 5.0.0, while Package B might malfunction due to incompatibilities introduced in 6.0.0.

Fragile Ecosystem: Re-exported dependencies become tightly coupled to your library's version. If you update your library and that update requires a newer version of the peer dependency, all your consumers are forced to upgrade as well, even if their code perfectly works with the older version. This creates a domino effect, potentially causing ripple effects throughout your project's ecosystem.
Think of it this way: You build a house on a foundation designed for a specific weight limit. By re-exporting a peer dependency as part of your house, you're changing the foundation's specifications. If you later decide the house needs a stronger foundation, all the houses built on your design (consumers of your library) would need significant work, even if their original design was perfectly functional.

The Illusion of Clarity: While re-exporting might seem to simplify your code on the surface, it actually obscures the true dependency picture. Consumers have no way of knowing which version of the peer dependency they're actually getting or why it's there. This can make debugging issues and understanding code flow more difficult.
Consider this analogy: You borrow a toolbox from a friend. They've added their own custom wrench but haven't mentioned it. When you use the wrench on a project and it breaks something, troubleshooting becomes harder because you weren't aware of the extra tool and its potential impact.

hahahaha so we need this scroll-tech/poseidon-circuit#28, then we can only use the poseidon primitive crate, no longer needs halo2_proofs, quite confusing.

lispc · 2024-04-22T11:41:08Z

great! high readable codes, with good docs and tests.

diamondo6 · 2024-04-22T19:59:31Z

Block hash upgrade....

* fix ci * update branch * fix clippy,test,docs * fix clippy * fix clippy

lispc · 2024-05-10T07:46:40Z

since CI is fixed, maybe we can mark this not draft PR?

supernovahs · 2024-05-10T15:29:56Z

crates/interpreter/src/host/dummy.rs

+    #[inline]
+    #[cfg(feature = "scroll")]
+    fn code_hash(&mut self, __address: Address) -> Option<(B256, bool)> {
+        Some((POSEIDON_EMPTY, false))
+    }
+
+    #[inline]
+    #[cfg(feature = "scroll")]
+    fn keccak_code_hash(&mut self, __address: Address) -> Option<(B256, bool)> {
+        Some((KECCAK_EMPTY, false))
+    }
+


Using both feature and not feature seems like an antipattern. Will be a drag in long term maintainance.

#[inline] fn code_hash(&mut self, __address: Address) -> Option<(B256, bool)> { #[cfg(feature = "scroll")] Some((POSEIDON_EMPTY, false)) Some((KECCAK_EMPTY, false)) }

this will also remove un- needed tags on top of code that uses code_hash function also

* fix extcodesize regression * fix test

lispc and others added 25 commits April 16, 2024 15:59

fix modexp and pairing gas

2ef1c0c

use poseidon

098f35e

port changes

ae94add

add scroll feature

301846b

add scroll handler

391421f

Merge branch 'see/v28-base' into see/v28

66a396a

Merge branch 'see/v29-base' into see/v29

63d75b2

# Conflicts: # Cargo.lock # crates/interpreter/Cargo.toml # crates/precompile/Cargo.toml # crates/primitives/Cargo.toml # crates/primitives/src/env/handler_cfg.rs # crates/revm/Cargo.toml

Merge branch 'see/v30-base' into see/v30

c98e0cb

# Conflicts: # Cargo.lock # crates/precompile/Cargo.toml # crates/primitives/Cargo.toml # crates/primitives/src/env.rs # crates/revm/Cargo.toml

Merge branch 'see/v31-base' into see/v31

131facf

# Conflicts: # crates/primitives/src/env/handler_cfg.rs

Merge branch 'see/v32-base' into see/v32

d0c493f

# Conflicts: # crates/revm/src/context.rs

Merge branch 'see/v34-base' into see/v34

66e9e83

Merge branch 'see/v35-base' into see/v35

290b5a3

# Conflicts: # crates/revm/src/handler.rs

impl l1fee

c9858f7

disable some precompiles

337a2f8

disable SELFDESTRUCT

591ee51

BLOCKHASH

476d7ed

fix modexp

fddc9dd

fix modexp

91433b3

fix poseidon hash

f6d4bd8

cleanup

d06633a

use keccak code hash for evm

c554acf

disable BASEFEE

06857da

handle l1 tx

61d8ca7

lispc commented Apr 22, 2024

View reviewed changes

crates/primitives/src/specification.rs

GRAY_GLACIER = 14,

MERGE = 15,

SHANGHAI = 16,

BERNOULLI = 17,

Copy link

Author

lispc Apr 22, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

hahaha

lispc commented Apr 22, 2024

View reviewed changes

add NotImplemented error

b57e7d9

lightsing and others added 6 commits April 26, 2024 15:54

fix pairing

9ecb479

fix bins

74c5a77

remove halo2proofs dep

dae1ae4

update poseidon dep

e1e8f7a

[fix] ci (#4)

1d44eb0

* fix ci * update branch * fix clippy,test,docs * fix clippy * fix clippy

add faster-hash gate (#3)

abb93af

supernovahs reviewed May 10, 2024

View reviewed changes

fix: extcodesize regression (#5)

09daf95

* fix extcodesize regression * fix test

lightsing deleted the branch v35-upsteam July 8, 2024 07:30

lightsing closed this Jul 8, 2024

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Scroll evm executor/v35 #1

Scroll evm executor/v35 #1

lispc commented Apr 22, 2024

lispc Apr 22, 2024

lightsing Apr 22, 2024

Badbunnybb Apr 22, 2024

lispc Apr 22, 2024

lightsing Apr 22, 2024

lispc Apr 22, 2024

lispc Apr 22, 2024

lispc Apr 22, 2024

glogo2 Apr 22, 2024

lispc Apr 22, 2024

lightsing Apr 22, 2024

lispc Apr 22, 2024

glogo2 Apr 22, 2024

ogulcanckr Apr 22, 2024

lispc commented Apr 22, 2024

diamondo6 commented Apr 22, 2024

lispc commented May 10, 2024

supernovahs May 10, 2024

Scroll evm executor/v35 #1

Scroll evm executor/v35 #1

Conversation

lispc commented Apr 22, 2024

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

lispc commented Apr 22, 2024

diamondo6 commented Apr 22, 2024

lispc commented May 10, 2024

Choose a reason for hiding this comment