Skip to content

Checking mergability - #5

Closed
heemankv wants to merge 1037 commits into
release/13.2.1-newfrom
main
Closed

Checking mergability#5
heemankv wants to merge 1037 commits into
release/13.2.1-newfrom
main

Conversation

@heemankv

Copy link
Copy Markdown
Collaborator

Short description of what this PR does.


A longer description, include motivation and intent if possible. Detail any caveats or follow-up steps still required.


Consider motivating any new dependencies.


Delete once completed:

  • link any issues closed by this pull-request
  • add all user-facing changes to CHANGELOG.md
  • new dependencies were added to the workspace toml

vbar and others added 30 commits June 24, 2025 09:25
…-max-l2-limit-fix

fix(executor): correct max L2 limit in L2 gas estimation logic
…ansaction-status

fix(rpc): PreConfirmed transaction status
Both the empty state update we're constructing and the pre-committed
state update needs to have the parent state commitment field set to the
state commitment of the parent block.

This is especially important for the pre-committed state update, as the
parent state commitment there is the _only_ way to identify which
parent block the state update is for (the block number is not
included in the response for `starknet_getStateUpdate`).
Consensus: add write-ahead log (WAL) support
…-timeout-mgr

Consensus: Add `ResetTimeouts` effect handling
… data

And make sure we return the INVALID_TRANSACTION_NONCE error in case
the gateway is returning an invalid nonce error.

This is slightly complicated by the fact that the gateway is returning
a "validation failure" error instead of the "invalid nonce" error, so
the only way for us to detect this is to check the error message
for the presence of the "invalid nonce" string.
…-attestation-link

chore(docs): add link to validator attestation tool
…ssages-status

fix(rpc): get_messages_status returning execution_status for v09
…ix-pre-committed-state-update-parent-state-commitment

fix(rpc/pending): properly set parent state commitment in state update
…nvalid-transaction-nonce-error

feat(rpc/v09): update INVALID_TRANSACTION_NONCE error with additional data
…d-block

fix: download block reorg handling
…der-rpc-api

feat(rpc): remove obsolete pathfinder_* APIs
… without execution status

For JSON-RPC 0.9.0 we should never return transactions without an execution
status, since `execution_status` is a required field in the response.
…tionReceipt`

On JSON-RPC 0.9, we have to use `ACCEPTED_ON_L2` if the transaction
receipt was found in the pending block, and `PRE_CONFIRMED` if it was
in the pre-confirmed block.
…ransaction-status

feat(rpc/v09): implement PRE_CONFIRMED and CANDIDATE transaction statuses
We treat it as a synonym for "pending" right now: the internal representation
is BlockId::Pending, because that's what our code handles.

Semantic differences between "pre_confirmed" and "pending" are already
handled based on the actual type of the pending data we have.
…lock

The expected behavior is to return an empty pending block for pre-0.9
JSON-RPC versions.
Pre-0.9 JSON-RPC versions require `PendingWatcher::get()` to return an
empty _pending_ block instead of an empty _pre-confirmed_ block in case
we _do_ have a pre-confirmed block.
The expected behavior is to return an empty pending block for pre-0.9
JSON-RPC versions.
CHr15F0x and others added 28 commits August 11, 2025 12:20
…idation

feat: integrate consensus p2p with the consensus engine
…events-finality

feat(rpc): starknet_subscribeEvents accepts & returns finality
…be-tx-status-follow-up

fix(rpc): subscribe tx status follow-up
…onrpc-0.9.0-spec

chore(spec): update JSON-RPC spec to 0.9.0 final
When processing a query with a block range that involves empty blocks
(blocks with zero transactions / events) `starknet_getEvents` would
stop scanning at the first empty block.

This is bad, because it doesn't even return a continuation token so
users would have no way of knowing that there are more events.

The issue is that `query_events_by_block()` is returning an `Option`,
and `None` is being interpreted as "no block found",
which causes the scan to stop prematurely when encountering empty blocks.

To fix this, we need to ensure that we handle no events correctly in
`events_for_block()` and just return an empty vector of events instead
of returning `None`.
Ever since we've added `rust-toolchain.toml` to the repository, Docker
builds do re-download the toolchain even though we should just be using
the one already installed in the image.

This commit updates the Dockerfile to exclude the `rust-toolchain.toml`
file from the build context, so that the toolchain is not reinstalled.
…nt-reinstall-rust-toolchain

fix(docker): don't reinstall Rust toolchain
…s-fix-for-empty-blocks

fix(rpc/method/get_events): fix `starknet_getEvents` for empty blocks
Otherwise we just get a parse error from Docker BuildKit.
…e-add-version

chore(Dockerfile): add version to dockerfile
@heemankv
heemankv deleted the branch release/13.2.1-new August 14, 2025 07:24
@heemankv heemankv closed this Aug 14, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

9 participants