refactor(juno): stop support for pending block

[MaksymMalicki]

Summary

StarkNet 0.14.0 introduced a new pending block format. The legacy Pending block has been replaced by a two-phase PreConfirmed and PreLatest. This PR removes support for Pending block in the synchroniser and the RPC. Since the network has been running 0.14.0+ for a long time, nodes no longer receive or produce the old-style pending blocks, meaning the code that handled them was dead weight — untriggered branches, split test cases, and a pending polling phase that never ran.

What changed

Sync

The sync layer's polling pipeline previously had two phases: one for the legacy pending format and one for the new pre-confirmed format. The legacy phase, along with all helpers and state logic it relied on, has been removed. Nodes now follow a single polling path (which is runPreConfirmedPhase()) for pending data.

RPC handlers (v9, v10)

The handlers no longer need to detect which pending format they received and adapt accordingly. Every code path that previously forked on "old format vs new format" has been collapsed into one - the support for pending Variant has been dropped and only the PreConfirmed is supported. Tests that covered the old-format path have been removed; existing tests have been updated to use PreConfirmed + PreLatest only.

RPC v8

The v8 spec still exposes a "pending" block ID that maps to the old format. The legacy type is therefore kept alive — but marked as deprecated — solely to serve v8 responses. It will be deleted when v8 is retired. Removed onPendingData field, pending data subscription, and the pending event loop case, as it always returns nil for 0.14.0+ versions.

Notes

• Due to the Deprecated tag being added to the Pending type and NewPending constructor, the linter issues will be present in the CI.
• Reviewers should not spend much time reviewing rpc/v6 package, as its flagged for deprecation
• Removal of the PendingData interface was moved to a separate PR to reduce the noise and make reviews easier: refactor(juno): obsolete the PendingData interface #3512

[codecov]

Codecov Report

❌ Patch coverage is 69.79167% with 29 lines in your changes missing coverage. Please review.
✅ Project coverage is 73.20%. Comparing base (9206c9a) to head (3bffac0).
⚠️ Report is 1 commits behind head on main.

Files with missing lines	Patch %	Lines
sync/pendingdata/helpers.go	25.00%	2 Missing and 4 partials ⚠️
rpc/v9/trace.go	57.14%	0 Missing and 3 partials ⚠️
sync/data_source.go	25.00%	2 Missing and 1 partial ⚠️
consensus/datasource/consensus_data_source.go	0.00%	2 Missing ⚠️
rpc/v10/pending_data_wrapper.go	33.33%	1 Missing and 1 partial ⚠️
rpc/v10/trace.go	84.61%	0 Missing and 2 partials ⚠️
rpc/v9/pending_data_wrapper.go	33.33%	1 Missing and 1 partial ⚠️
sync/sync.go	60.00%	1 Missing and 1 partial ⚠️
core/pending_state.go	0.00%	1 Missing ⚠️
rpc/v10/helpers.go	80.00%	1 Missing ⚠️
... and 5 more

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #3498      +/-   ##
==========================================
- Coverage   75.51%   73.20%   -2.32%     
==========================================
  Files         384      384              
  Lines       34973    35173     +200     
==========================================
- Hits        26411    25749     -662     
- Misses       6700     7139     +439     
- Partials     1862     2285     +423     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[thiagodeev]

Hey. We are on the right track.
This is a partial review, I haven't finished yet.
Requesting changes since we still have mentions to pending in the starkinetdata/feeder pkg

[EgeCaner]

First round

[thiagodeev]

LGTM

[EgeCaner]

Looks good, thanks! I love this PR