refactor(port): delete unused multiply_u64_get_top_64

multiply_u64_get_top_64 is no longer used anywhere. Delete its
implementation and tests.

src/quick-lint-js/port/math.h

@@ -3,54 +3,13 @@
 
 #pragma once
 
-#include <cstdint>
-#include <quick-lint-js/port/have.h>
-#include <quick-lint-js/port/warning.h>
-
-#if QLJS_HAVE_INTRIN_H
-#include <intrin.h>
-#endif
-
 namespace quick_lint_js {
 // On some compilers, std::max is not constexpr. Define our own which is
 // constexpr.
 template <class T, class U>
 constexpr auto maximum(T x, U y) {
   return x < y ? y : x;
 }
-
-// Returns (lhs*rhs)>>64.
-inline std::uint64_t multiply_u64_get_top_64(std::uint64_t lhs,
-                                             std::uint64_t rhs) {
-  // Discussion of approaches: https://stackoverflow.com/q/28868367
-#if QLJS_HAVE_INT128
-  QLJS_WARNING_PUSH
-  QLJS_WARNING_IGNORE_GCC("-Wpedantic")
-
-  using U128 = unsigned __int128;
-  return static_cast<std::uint64_t>(
-      (static_cast<U128>(lhs) * static_cast<U128>(rhs)) >> 64);
-
-  QLJS_WARNING_POP
-#elif QLJS_HAVE_UMULH
-  return __umulh(lhs, rhs);
-#else
-  // Algorithm adapted from Fabian 'ryg' Giesen's public domain work:
-  // https://github.com/jkbonfield/rans_static/blob/576e9db9e02659a5797ce5e21dd99f6ce0285727/rans64.h#L51-L64
-  auto lo = [](std::uint64_t x) -> std::uint64_t { return x & 0xffff'ffff; };
-  auto hi = [](std::uint64_t x) -> std::uint64_t { return x >> 32; };
-  std::uint64_t a = hi(lhs);
-  std::uint64_t b = lo(lhs);
-  std::uint64_t c = hi(rhs);
-  std::uint64_t d = lo(rhs);
-  std::uint64_t x0 = d * b;
-  std::uint64_t x1 = lo(d * a) + lo(c * b) + hi(x0);
-  std::uint64_t x2 = hi(d * a) + hi(c * b) + lo(c * a) + hi(x1);
-  std::uint64_t x3 = hi(c * a) + hi(x2);
-  // x3, lo(x2), lo(x1), lo(x0)
-  return (x3 << 32) + lo(x2);
-#endif
-}
 }
 
 // quick-lint-js finds bugs in JavaScript programs.

test/CMakeLists.txt

@@ -127,7 +127,6 @@ quick_lint_js_add_executable(
   test-lsp-uri.cpp
   test-lsp-workspace-configuration.cpp
   test-math-overflow.cpp
-  test-math.cpp
   test-narrow-cast.cpp
   test-options.cpp
   test-output-stream.cpp