Aggregation v2 #296

Riateche · 2024-07-09T10:10:00Z

Include oracle program as a dependency
Move all pyth functions from bank.rs to a submodule and remove unnecessary arguments
Aggregate prices and create price feed messages for price feeds with v2 aggregation enabled
Merge v1 and v2 messages and generate message accounts and wormhole accounts for all of them

ali-behjati

Nice.

runtime/src/bank/pyth_accumulator.rs

ali-behjati · 2024-07-23T15:59:18Z

runtime/src/bank/pyth_accumulator.rs

+    );
+    pub static ref ORACLE_PID: Pubkey = env_pubkey_or(
+        "ORACLE_PID",
+        "FsJ3A3u2vn5cTVofAjvy6y5kwABJAqYWpe4975bi2epH"


not for now, but seeing this i'm wondering whether in the future it makes sense to move the entire code in this file to pythnet_sdk

ali-behjati · 2024-07-23T16:06:16Z

runtime/src/bank/pyth_accumulator_tests.rs

+            // Insert into message buffer in reverse order to test that accumulator
+            // sorts first.


where are you reversing it :?

This was left over after copy-paste. Removed.

ali-behjati · 2024-07-23T16:06:35Z

runtime/src/bank/pyth_accumulator_tests.rs

+    // TODO: New test functionality here.
+    // 1. Create Price Feed Accounts owned by ORACLE_PUBKEY
+    // 2. Populate Price Feed Accounts
+    // 3. Call update_v2()
+    //    - Cases:
+    //      - No V1 Messages, Only Price Accounts with no V2
+    //      - No V1 Messages, Some Price Accounts with no V2
+    //      - Some V1 Messages, No Price Accounts with no V2
+    //      - Some V1 Messages, Some Price Accounts with no V2
+    //      - Simulate PriceUpdate that WOULD trigger a real V1 aggregate before End of Slot
+    //      - Simulate PriceUpdate that doesn't trigger a real V1 aggregate, only V2.


i think this is outdated :?

Yes. Removed.

ali-behjati

Nice!

guibescos · 2024-07-24T13:50:29Z

runtime/src/bank/pyth_accumulator.rs

+    ]
+}
+
+pub fn update_v1<'a>(


lifetime seems unnecessary

guibescos · 2024-07-24T16:21:06Z

runtime/src/bank/pyth_accumulator_tests.rs

+        PostedMessageUnreliableData::deserialize(&mut wormhole_message_account.data()).unwrap();
+
+    // Create MerkleAccumulator by hand to verify that the Wormhole message
+    // contents are correctg.


typo: correctg

guibescos · 2024-07-24T16:26:58Z

runtime/src/bank/pyth_accumulator_tests.rs

+        .sorted_unstable()
+        .dedup()
+        .collect::<Vec<_>>();
+


nitpick: I'd assert the length of messages here, to confirm they haven't been deduplicated

…gation

… tree

* feat: accumulator v2 (wip 1) * feat: accumulator v2 (wip 2) * chore: fix some warnings and fix validator build * feat: accumulator v2 (wip 3) * feat: skip message buffer aggregation if all price feeds use v2 aggregation * feat: combine v1 and v2 aggregation price feed messages in the merkle tree * refactor: clean up and avoid extra vector * fix: don't aggregate if already aggregated in the program * test: bring back old pyth tests (wip) * test: fix old pyth tests * test: add initial v2 tests to validator * refactor: unify env var handling * chore: fix build for stable Rust by ignoring lints * test: fix and expand accumulator-v2 tests * chore: clean up comments * chore: update dependency * chore: fix clippy and fmt * test: add extra check * chore: bump version to 1.14.175 * feat: add more traces for accumulator * fix: load program accounts by index * feat: add checks to make sure indexes are set * chore: bump oracle Co-authored-by: Reisen <[email protected]>

ali-behjati reviewed Jul 12, 2024

View reviewed changes

runtime/src/bank/pyth_accumulator.rs Outdated Show resolved Hide resolved

runtime/src/bank/pyth_accumulator.rs Show resolved Hide resolved

runtime/src/bank/pyth_accumulator.rs Outdated Show resolved Hide resolved

Reisen force-pushed the accumulator-v2 branch from 4dc85c5 to f3deb0f Compare July 17, 2024 11:44

ali-behjati reviewed Jul 23, 2024

View reviewed changes

ali-behjati approved these changes Jul 23, 2024

View reviewed changes

guibescos reviewed Jul 24, 2024

View reviewed changes

guibescos approved these changes Jul 24, 2024

View reviewed changes

Riateche and others added 19 commits July 24, 2024 17:45

feat: accumulator v2 (wip 1)

1106de9

feat: accumulator v2 (wip 2)

af6e6cb

chore: fix some warnings and fix validator build

266b7b1

feat: accumulator v2 (wip 3)

62b19ab

feat: skip message buffer aggregation if all price feeds use v2 aggre…

0582205

…gation

feat: combine v1 and v2 aggregation price feed messages in the merkle…

c7e3834

… tree

refactor: clean up and avoid extra vector

4e730d0

fix: don't aggregate if already aggregated in the program

c64a43b

test: bring back old pyth tests (wip)

159a5ed

test: fix old pyth tests

c8f9b8d

test: add initial v2 tests to validator

3daebe2

refactor: unify env var handling

9c49fcb

chore: fix build for stable Rust by ignoring lints

abffa06

test: fix and expand accumulator-v2 tests

5cd4657

chore: clean up comments

287cb06

chore: update dependency

9f1b93a

chore: fix clippy and fmt

65b352e

test: add extra check

4565643

chore: bump version to 1.14.175

bd4012a

Riateche force-pushed the accumulator-v2 branch from 61ff045 to bd4012a Compare July 24, 2024 16:50

ali-behjati approved these changes Jul 24, 2024

View reviewed changes

Riateche merged commit a378583 into pyth-v1.14.17 Jul 24, 2024

Riateche deleted the accumulator-v2 branch July 24, 2024 17:59

		// Insert into message buffer in reverse order to test that accumulator
		// sorts first.

Aggregation v2 #296

Aggregation v2 #296

Uh oh!

Conversation

Riateche commented Jul 9, 2024

Uh oh!

ali-behjati left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

ali-behjati left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants