[CI] Ignore ruff D209, rejoin the docstrings it split, and silence incompatible-rule warnings (#57212)

Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
This commit is contained in:
Harry Mellor
2026-09-16 10:57:24 -07:00
committed by GitHub
co-authored by Claude Opus 5
parent 1085b64425
commit 651a88c09a
693 changed files with 2088 additions and 4171 deletions
@@ -847,8 +847,7 @@ class FlashMLASparseImpl(SparseMLACommonImpl[FlashMLASparseMetadata]):
prefill_meta: "FlashMLASparseMetadata.FP8SeparatePrefillDecode.Prefill",
) -> tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
"""All-gather this rank's upconverted KV shard so the chunk's rows attend
the whole context, and map their top-k onto the rank-major result.
"""
the whole context, and map their top-k onto the rank-major result."""
shard_rows = int(chunk.chunk_tot_seqlen)
assert self.gathered_kv_workspace is not None
gathered_kv = self.gathered_kv_workspace[: self.dcp_world_size * shard_rows]