Fix std/hashes completely ignoring endianness (#25386)

This is a problem on big-endian CPUs because you end up with nimvm
computing something different than Nim proper, so e.g. a const table
won't work.

I also took the liberty to replace a redundant implementation of load4
in murmurHash.

(Thanks to barracuda156 for helping debug this.)

(cherry picked from commit a061f026a8)

lib/pure/hashes.nim

@@ -304,6 +304,35 @@ else:
 proc rotl32(x: uint32, r: int): uint32 {.inline.} =
   (x shl r) or (x shr (32 - r))
 
+proc load4e(s: openArray[byte], o=0): uint32 {.inline.} =
+  uint32(s[o + 3]) shl 24 or uint32(s[o + 2]) shl 16 or
+  uint32(s[o + 1]) shl  8 or uint32(s[o + 0])
+
+proc load8e(s: openArray[byte], o=0): uint64 {.inline.} =
+  uint64(s[o + 7]) shl 56 or uint64(s[o + 6]) shl 48 or
+  uint64(s[o + 5]) shl 40 or uint64(s[o + 4]) shl 32 or
+  uint64(s[o + 3]) shl 24 or uint64(s[o + 2]) shl 16 or
+  uint64(s[o + 1]) shl  8 or uint64(s[o + 0])
+
+when declared(copyMem):
+  from std/endians import littleEndian64, littleEndian32
+
+proc load4(s: openArray[byte], o=0): uint32 {.inline.} =
+  when nimvm: result = load4e(s, o)
+  else:
+    when declared copyMem:
+      result = uint32(0)
+      littleEndian32(addr result, addr s[o])
+    else: result = load4e(s, o)
+
+proc load8(s: openArray[byte], o=0): uint64 {.inline.} =
+  when nimvm: result = load8e(s, o)
+  else:
+    when declared copyMem:
+      result = uint64(0)
+      littleEndian64(addr result, addr s[o])
+    else: result = load8e(s, o)
+
 proc murmurHash(x: openArray[byte]): Hash =
   # https://github.com/PeterScott/murmur3/blob/master/murmur3.c
   const
@@ -320,24 +349,10 @@ proc murmurHash(x: openArray[byte]): Hash =
     h1: uint32 = uint32(0)
     i = 0
 
-
-  template impl =
-    var j = stepSize
-    while j > 0:
-      dec j
-      k1 = (k1 shl 8) or (ord(x[i+j])).uint32
-
   # body
   while i < n * stepSize:
-    var k1: uint32 = uint32(0)
+    var k1 = load4(x, i)
 
-    when nimvm:
-      impl()
-    else:
-      when declared(copyMem):
-        copyMem(addr k1, addr x[i], 4)
-      else:
-        impl()
     inc i, stepSize
 
     k1 = imul(k1, c1)
@@ -384,32 +399,6 @@ const k0 = 0xc3a5c85c97cb3127u64 # Primes on (2^63, 2^64) for various uses
 const k1 = 0xb492b66fbe98f273u64
 const k2 = 0x9ae16a3b2f90404fu64
 
-proc load4e(s: openArray[byte], o=0): uint32 {.inline.} =
-  uint32(s[o + 3]) shl 24 or uint32(s[o + 2]) shl 16 or
-  uint32(s[o + 1]) shl  8 or uint32(s[o + 0])
-
-proc load8e(s: openArray[byte], o=0): uint64 {.inline.} =
-  uint64(s[o + 7]) shl 56 or uint64(s[o + 6]) shl 48 or
-  uint64(s[o + 5]) shl 40 or uint64(s[o + 4]) shl 32 or
-  uint64(s[o + 3]) shl 24 or uint64(s[o + 2]) shl 16 or
-  uint64(s[o + 1]) shl  8 or uint64(s[o + 0])
-
-proc load4(s: openArray[byte], o=0): uint32 {.inline.} =
-  when nimvm: result = load4e(s, o)
-  else:
-    when declared copyMem:
-      result = uint32(0)
-      copyMem result.addr, s[o].addr, result.sizeof
-    else: result = load4e(s, o)
-
-proc load8(s: openArray[byte], o=0): uint64 {.inline.} =
-  when nimvm: result = load8e(s, o)
-  else:
-    when declared copyMem:
-      result = uint64(0)
-      copyMem result.addr, s[o].addr, result.sizeof
-    else: result = load8e(s, o)
-
 proc lenU(s: openArray[byte]): uint64 {.inline.} = s.len.uint64
 
 proc shiftMix(v: uint64): uint64 {.inline.} = v xor (v shr 47)

tests/pragmas/thintprocessing.nim

@@ -3,7 +3,7 @@ discard """
   matrix: "--hint:processing"
   nimout: '''
 compile start
-...
+....
 warn_module.nim(6, 6) Hint: 'test' is declared but not used [XDeclaredButNotUsed]
 compile end
 '''