From bff97833b62e5b8a8f2cbbbd408f46846e47020d Mon Sep 17 00:00:00 2001
From: ddidderr <ddidderr@paul.network>
Date: Sun, 19 Jul 2026 19:08:13 +0200
Subject: [PATCH] feat(compress): move long-match update arithmetic into Rust

Move the bounded nextToUpdate catch-up arithmetic used after very long
matches into Rust. C continues to calculate the source index against the
private match-state window and retains all surrounding assertions and setup;
Rust owns the exact wrapping U32 threshold and 192-byte catch-up policy.

Test Plan:
- cargo fmt --manifest-path rust/Cargo.toml -- --check
- ulimit -v 41943040 && CARGO_BUILD_JOBS=1 cargo test --manifest-path rust/Cargo.toml limit_next_to_update --lib (1 passed)
- ulimit -v 41943040 && CARGO_BUILD_JOBS=1 cargo test --manifest-path rust/Cargo.toml --lib (692 passed)
- ulimit -v 41943040 && CARGO_BUILD_JOBS=1 cargo clippy --manifest-path rust/Cargo.toml --all-targets -- -D warnings
- ulimit -v 41943040 && MAKEFLAGS=-j1 make -B -C programs -j1 zstd
- ulimit -v 41943040 && MAKEFLAGS=-j1 make -C tests -j1 test-zstream ZSTREAM_TESTTIME=-T1s (84 tests and both short fuzz rounds passed)
---
 lib/compress/zstd_compress.c |  4 ++--
 rust/src/zstd_compress.rs    | 29 +++++++++++++++++++++++++++++
 2 files changed, 31 insertions(+), 2 deletions(-)

diff --git a/lib/compress/zstd_compress.c b/lib/compress/zstd_compress.c
index f434d7249..e2b9eb83d 100644
--- a/lib/compress/zstd_compress.c
+++ b/lib/compress/zstd_compress.c
@@ -604,6 +604,7 @@ size_t ZSTD_compressStream2_c(ZSTD_CCtx* cctx,
                               ZSTD_EndDirective endOp);
 int ZSTD_rust_simpleCompress2Level(const void* cctx, size_t srcSize);
 int ZSTD_rust_simpleCompressStream2Level(const void* cctx, size_t srcSize);
+U32 ZSTD_rust_limitNextToUpdate(U32 curr, U32 nextToUpdate);
 void ZSTD_rust_reduceIndex(U32* hashTable, U32 hashSize,
                            U32* chainTable, U32 chainSize,
                            U32* hashTable3, U32 hashSize3,
@@ -4721,8 +4722,7 @@ static void ZSTD_rust_buildSeqStore_prepareMatchState(void* context,
         const BYTE* const istart = (const BYTE*)src;
         const U32 curr = (U32)(istart-base);
         if (sizeof(ptrdiff_t)==8) assert(istart - base < (ptrdiff_t)(U32)(-1));
-        if (curr > ms->nextToUpdate + 384)
-            ms->nextToUpdate = curr - MIN(192, (U32)(curr - ms->nextToUpdate - 384));
+        ms->nextToUpdate = ZSTD_rust_limitNextToUpdate(curr, ms->nextToUpdate);
     }
 }
 
diff --git a/rust/src/zstd_compress.rs b/rust/src/zstd_compress.rs
index 552c408f4..4d3d7e5f5 100644
--- a/rust/src/zstd_compress.rs
+++ b/rust/src/zstd_compress.rs
@@ -5238,6 +5238,26 @@ pub extern "C" fn ZSTD_rust_windowNeedOverflowCorrection(
     ) as u32
 }
 
+#[inline]
+fn limit_next_to_update(curr: u32, next_to_update: u32) -> u32 {
+    let const_gap = 384u32;
+    let const_step = 192u32;
+    if curr > next_to_update.wrapping_add(const_gap) {
+        curr.wrapping_sub(const_step.min(curr.wrapping_sub(next_to_update).wrapping_sub(const_gap)))
+    } else {
+        next_to_update
+    }
+}
+
+/// Limit a stale matchfinder update point after a very long match.
+///
+/// C computes the source index against its private window base; Rust owns the
+/// wrapping U32 threshold and bounded catch-up arithmetic.
+#[no_mangle]
+pub extern "C" fn ZSTD_rust_limitNextToUpdate(curr: u32, next_to_update: u32) -> u32 {
+    limit_next_to_update(curr, next_to_update)
+}
+
 #[inline]
 fn next_input_size_hint(
     in_buffer_mode: c_int,
@@ -10803,6 +10823,15 @@ mod tests {
         assert!(!window_need_overflow_correction(27, true, 1024, false));
     }
 
+    #[test]
+    fn limit_next_to_update_preserves_threshold_and_bounded_catch_up() {
+        assert_eq!(limit_next_to_update(884, 500), 500);
+        assert_eq!(limit_next_to_update(885, 500), 884);
+        assert_eq!(limit_next_to_update(1000, 500), 884);
+        assert_eq!(limit_next_to_update(1200, 500), 1008);
+        assert_eq!(ZSTD_rust_limitNextToUpdate(300, u32::MAX - 100), 283);
+    }
+
     #[test]
     fn window_overflow_need_scales_the_frequency_with_correction_count() {
         assert!(window_can_overflow_correct(27, 3, 16, 0, 0));