Move xxh3_64 methods to vespalib. That also removes the need for workarounds for GCC false positives.

4 files changed, 27 insertions, 29 deletions

diff --git a/vespalib/src/vespa/vespalib/fuzzy/sparse_state.h b/vespalib/src/vespa/vespalib/fuzzy/sparse_state.h
index 7e381468fbe..e023f4d3de2 100644
--- a/vespalib/src/vespa/vespalib/fuzzy/sparse_state.h
+++ b/vespalib/src/vespa/vespalib/fuzzy/sparse_state.h
@@ -2,13 +2,12 @@
 #pragma once
 
 #include <vespa/config.h>
+#include <vespa/vespalib/stllike/hash_fun.h>
 #include <algorithm>
 #include <array>
 #include <cassert>
-#include <cstdint>
 #include <ostream>
 #include <span>
-#include <xxhash.h> // TODO factor out?
 
 namespace vespalib::fuzzy {
 
@@ -80,15 +79,8 @@ public:
         size_t operator()(const FixedSparseState& s) const noexcept {
             static_assert(std::is_same_v<uint32_t, std::decay_t<decltype(s.indices[0])>>);
             static_assert(std::is_same_v<uint8_t,  std::decay_t<decltype(s.costs[0])>>);
-            // FIXME GCC 12.2 worse-than-useless(tm) warning false positives :I
-#pragma GCC diagnostic push
-#ifdef VESPA_USE_SANITIZER
-#  pragma GCC diagnostic ignored "-Wstringop-overread" // https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98465 etc.
-#endif
-#pragma GCC diagnostic ignored "-Warray-bounds"
-            return (XXH3_64bits(s.indices.data(), s.sz * sizeof(uint32_t)) ^
-                    XXH3_64bits(s.costs.data(), s.sz));
-#pragma GCC diagnostic pop
+            return (xxhash::xxh3_64(s.indices.data(), s.sz * sizeof(uint32_t)) ^
+                    xxhash::xxh3_64(s.costs.data(), s.sz));
         }
     };
 };
diff --git a/vespalib/src/vespa/vespalib/stllike/hash_fun.cpp b/vespalib/src/vespa/vespalib/stllike/hash_fun.cpp
index e4911d2e0c6..6802e5f91a6 100644
--- a/vespalib/src/vespa/vespalib/stllike/hash_fun.cpp
+++ b/vespalib/src/vespa/vespalib/stllike/hash_fun.cpp
@@ -6,23 +6,21 @@
 namespace vespalib {
 
 size_t
-hashValue(const char *str) noexcept
-{
-    return hashValue(str, strlen(str));
+hashValue(const char *str) noexcept {
+    return xxhash::xxh3_64(str, strlen(str));
 }
 
-/**
- * @brief Calculate hash value.
- *
- * The hash function XXH3_64bits from xxhash library.
- * @param buf input buffer
- * @param sz input buffer size
- * @return hash value of input
- **/
-size_t
-hashValue(const void * buf, size_t sz) noexcept
-{
+namespace xxhash {
+
+uint64_t
+xxh3_64(uint64_t value) noexcept {
+    return XXH3_64bits(&value, sizeof(value));
+}
+
+uint64_t
+xxh3_64(const void * buf, size_t sz) noexcept {
     return XXH3_64bits(buf, sz);
 }
 
 }
+}
diff --git a/vespalib/src/vespa/vespalib/stllike/hash_fun.h b/vespalib/src/vespa/vespalib/stllike/hash_fun.h
index f8a8a06b921..8fecc41b4c1 100644
--- a/vespalib/src/vespa/vespalib/stllike/hash_fun.h
+++ b/vespalib/src/vespa/vespalib/stllike/hash_fun.h
@@ -2,7 +2,6 @@
 #pragma once
 
 #include <vespa/vespalib/stllike/string.h>
-#include <cstdint>
 
 namespace vespalib {
 
@@ -64,9 +63,18 @@ template<typename T> struct hash<const T *> {
     size_t operator() (const T * arg) const noexcept { return size_t(arg); }
 };
 
+namespace xxhash {
+
+uint64_t xxh3_64(uint64_t value) noexcept;
+uint64_t xxh3_64(const void *str, size_t sz) noexcept;
+
+}
+
 // reuse old string hash function
 size_t hashValue(const char *str) noexcept;
-size_t hashValue(const void *str, size_t sz) noexcept;
+inline size_t hashValue(const void *buf, size_t sz) noexcept {
+    return xxhash::xxh3_64(buf, sz);
+}
 
 struct hash_strings {
     size_t operator() (const vespalib::string & arg) const noexcept { return hashValue(arg.data(), arg.size()); }
diff --git a/vespalib/src/vespa/vespalib/util/shared_string_repo.cpp b/vespalib/src/vespa/vespalib/util/shared_string_repo.cpp
index a5942e1af9b..09f2bbd828d 100644
--- a/vespalib/src/vespa/vespalib/util/shared_string_repo.cpp
+++ b/vespalib/src/vespa/vespalib/util/shared_string_repo.cpp
@@ -1,7 +1,7 @@
 // Copyright Vespa.ai. Licensed under the terms of the Apache 2.0 license. See LICENSE in the project root.
 
 #include "shared_string_repo.h"
-#include <xxhash.h>
+#include <vespa/vespalib/stllike/hash_fun.h>
 #include <charconv>
 #include <cassert>
 
@@ -232,7 +232,7 @@ string_id
 SharedStringRepo::resolve(vespalib::stringref str) {
     uint32_t direct_id = try_make_direct_id(str);
     if (direct_id >= ID_BIAS) {
-        uint64_t full_hash = XXH3_64bits(str.data(), str.size());
+        uint64_t full_hash = xxhash::xxh3_64(str.data(), str.size());
         uint32_t part = full_hash & PART_MASK;
         uint32_t local_hash = full_hash >> PART_BITS;
         uint32_t local_idx = _partitions[part].resolve(AltKey{str, local_hash});