…value count (#7708)

#7376 made the RLE miniblock decoder reject chunks whose run lengths sum
past the declared value count. The legacy (pre run-length-width)
miniblock encoder rolled back to a power-of-2 checkpoint after a run had
already crossed it, so existing files contain chunks that declare 2048
values but hold runs summing slightly more; the excess values are
re-encoded at the start of the next chunk. The legacy decoder truncated
the excess, so those files were valid when written — after #7376 they
fail to read with:

```
Invalid user input: RLE decoding overflowed expected value count: produced at least 2080, expected 2048
```

This restores the truncating behavior for miniblock decoding: clamp the
final crossing run at the declared count and ignore trailing legacy
runs, while still rejecting underflow and zero run lengths. Block
payloads have no chunk boundaries, so an overflow there can only be
corruption and remains a hard error. Current encoders emit exact chunks,
so this only changes how legacy files are read. Includes regression
tests for the legacy chunk-boundary overflow and the strict block path.

<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
## Summary by CodeRabbit

* **Bug Fixes**
* Improved run-length (RLE) decoding to cap results to the expected
number of values, avoiding failures when runs exceed the requested
range.
* Enhanced compatibility with legacy data by clamping miniblock overflow
at chunk boundaries and ensuring decoded output matches the expected
length and values.
* Updated miniblock test coverage to reflect the new overflow semantics,
and added a regression test for boundary-crossing truncation.
* Preserved strict overflow handling for full block payloads, which
still reports an error.
<!-- end of auto-generated comment: release notes by coderabbit.ai -->

## Summary
- fail RLE miniblock encoding when chunk selection cannot advance
instead of returning partial buffers with the original value count
- assert that raw encoded run lengths match each declared chunk count
across U8, U16, and U32 run-length widths
- retain the exact-prefix implementation introduced in #7376; this does
not change the file format

## Context
`LANCE_MINIBLOCK_MAX_VALUES=1` is currently accepted, but a multi-value
RLE page cannot represent a non-final one-value chunk because
`log_num_values == 0` identifies the final chunk. The encoder previously
broke out of the loop and returned incomplete data while retaining the
original `num_values`.

## Test plan
- [x] `cargo test -p lance-encoding` (430 passed, 5 ignored)
- [x] `cargo fmt --all -- --check`
- [x] `cargo clippy --all --tests --benches -- -D warnings`

Made with [Cursor](https://cursor.com)

<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->

## Summary by CodeRabbit

* **Bug Fixes**
* Prevented incomplete or invalid compressed data from being returned
when encoding cannot make progress.
* Improved error reporting with contextual details for encoding
failures.
  * Fixed run-length handling across the 2,048-value boundary.

* **Tests**
* Added coverage for multiple run-length widths and zero-progress
encoding scenarios.

<!-- end of auto-generated comment: release notes by coderabbit.ai -->

Co-authored-by: Cursor <[email protected]>