refactor(window): consume explicitly sorted input

[KKould]

What problem does this PR solve?

Issue link:

Window execution currently performs its own in-memory sort and materializes the full input. This couples Window to one sorting implementation and prevents the existing spill-backed ExternalSort from satisfying Window ordering.

What is changed and how it works?

Window now consumes explicitly sorted input instead of sorting internally. The binder inserts a Sort on PARTITION BY + ORDER BY; native builds with the spill feature therefore use the existing ExternalSort executor, while non-spill and WASM builds continue to use the in-memory Sort implementation.

The Window executor processes ordered input incrementally and retains only the rows required by the current function set: one row, one peer group, or one partition. Partition and peer boundaries are detected from cached sort values, and a single pending row carries a boundary across buffered output.

Code changes

• Has Rust code change
• Has CI related scripts change

Check List

Tests

• Unit test
• Integration test
• Manual test (add detailed scripts or steps below)
• No code

Window unit coverage verifies row-, peer-, and partition-retention behavior. SQL logic coverage verifies ranking, aggregate windows, multiple window groups, partition boundaries, peer ties, views, and explain output. The spill-enabled build also compiles successfully.

Side effects

• Performance regression: Consumes more CPU
• Performance regression: Consumes more Memory
• Breaking backward compatibility

Note for reviewer

The primary goal is to decouple Window from Sort. Window only requires ordered input; with spill enabled, the existing Sort execution path selects ExternalSort and provides spill support without adding spill-specific code to Window.

The repository-wide clippy target remains blocked by pre-existing warnings outside this change: an unused parser import, unused optimizer pattern/matcher code, and manual_is_multiple_of in the existing ExternalSort test.

[codecov]

Codecov Report

❌ Patch coverage is 99.21260% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 92.90%. Comparing base (506bd78) to head (449219b).

Files with missing lines	Patch %	Lines
src/execution/dql/window.rs	99.05%	1 Missing ⚠️

Additional details and impacted files

@@              Coverage Diff               @@
##           feat/spill     #378      +/-   ##
==============================================
- Coverage       92.91%   92.90%   -0.01%     
==============================================
  Files             253      253              
  Lines           43141    43190      +49     
==============================================
+ Hits            40084    40126      +42     
- Misses           3057     3064       +7     

Flag	Coverage Δ
rust	92.90% <99.21%> (-0.01%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

Files with missing lines	Coverage Δ
src/binder/window.rs	96.07% <100.00%> (+0.18%)	⬆️
src/execution/dql/window/function.rs	97.29% <100.00%> (+0.07%)	⬆️
src/execution/dql/window.rs	99.28% <99.05%> (+0.29%)	⬆️

... and 2 files with indirect coverage changes

🚀 New features to boost your workflow:

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