Use ReadLE64 in uint256::GetUint64 instead of duplicating logic #26105

No need to have a (naive) copy of the ReadLE64 logic inside uint256::GetUint64, when we have an optimized function for exactly that.

ACK 04fee75bacb9ec3bceff1246ba6c8ed8a8759548

The existing code assumes host endianness.

ACK 04fee75bacb9ec3bceff1246ba6c8ed8a8759548 review, this use of ReadLE64() is similar to the existing invocation by Num3072::Num3072(), sanity checked that before and after this change GetUint64() returns the same result (debug build, clang 13)

--- a/src/uint256.h
+++ b/src/uint256.h
@@ -7,6 +7,7 @@
 #include <crypto/common.h>
+#include <logging.h>
 #include <span.h>
 
@@ -85,7 +86,20 @@ public:
 
     uint64_t GetUint64(int pos) const
     {
-        return ReadLE64(m_data + pos * 8);
+        const uint8_t* ptr = m_data + pos * 8;
+        uint64_t x = ((uint64_t)ptr[0]) | \
+                     ((uint64_t)ptr[1]) << 8 | \
+                     ((uint64_t)ptr[2]) << 16 | \
+                     ((uint64_t)ptr[3]) << 24 | \
+                     ((uint64_t)ptr[4]) << 32 | \
+                     ((uint64_t)ptr[5]) << 40 | \
+                     ((uint64_t)ptr[6]) << 48 | \
+                     ((uint64_t)ptr[7]) << 56;
+        LogPrintf("GetUint64 old: %s\n", x);
+
+        const auto y = ReadLE64(m_data + pos * 8);
+        LogPrintf("GetUint64 new: %s\n", y);
+
+        assert(x == y);
+        return y;
     }

2022-09-16T09:13:53Z GetUint64 old: 2830827502690553650
2022-09-16T09:13:53Z GetUint64 new: 2830827502690553650
2022-09-16T09:13:53Z GetUint64 old: 13427380872610120290
2022-09-16T09:13:53Z GetUint64 new: 13427380872610120290
2022-09-16T09:13:53Z GetUint64 old: 12183277635375193944
2022-09-16T09:13:53Z GetUint64 new: 12183277635375193944
2022-09-16T09:13:53Z GetUint64 old: 16580367070613284548
2022-09-16T09:13:53Z GetUint64 new: 16580367070613284548
2022-09-16T09:13:53Z GetUint64 old: 16949092074016075712
2022-09-16T09:13:53Z GetUint64 new: 16949092074016075712
2022-09-16T09:13:53Z GetUint64 old: 14891467430351852074
2022-09-16T09:13:53Z GetUint64 new: 14891467430351852074
2022-09-16T09:13:53Z GetUint64 old: 2319828850847527094
2022-09-16T09:13:53Z GetUint64 new: 2319828850847527094
2022-09-16T09:13:53Z GetUint64 old: 921024485749193602
2022-09-16T09:13:53Z GetUint64 new: 921024485749193602

The existing code assumes host endianness.

No, it does not. Both the old and new code interpret the bytes as little-endian.

The existing code assumes host endianness.

No, it does not. Both the old and new code interpret the bytes as little-endian.

Right, I meant that the old code interprets the bytes as little endian and stores them as little endian, even if the host machine is big-endian. ReadLE64 invokes le64toh which does not make an assumption about "host" endianness.