math.bits: port changes from e66e996, so that -cstrict -cc gcc-11 passes for markdown as well

vlib/math/bits/unsafe_bits.v

@@ -3,13 +3,29 @@
 // that can be found in the LICENSE file.
 module bits
 
+union U32_F32 {
+	u u32
+	f f32
+}
+
+union U64_F64 {
+	u u64
+	f f64
+}
+
 // f32_bits returns the IEEE 754 binary representation of f,
 // with the sign bit of f and the result in the same bit position.
 // f32_bits(f32_from_bits(x)) == x.
 @[inline]
 pub fn f32_bits(f f32) u32 {
-	p := *unsafe { &u32(&f) }
-	return p
+	$if tinyc {
+		return *unsafe { &u32(&f) } // this is faster for tcc, but causes `error: dereferencing type-punned pointer will break strict-aliasing rules` on gcc, with -cstrict
+	}
+	return unsafe {
+		U32_F32{
+			f: f
+		}.u
+	}
 }
 
 // f32_from_bits returns the floating-point number corresponding
@@ -18,8 +34,14 @@ pub fn f32_bits(f f32) u32 {
 // f32_from_bits(f32_bits(x)) == x.
 @[inline]
 pub fn f32_from_bits(b u32) f32 {
-	p := *unsafe { &f32(&b) }
-	return p
+	$if tinyc {
+		return *unsafe { &f32(&b) }
+	}
+	return unsafe {
+		U32_F32{
+			u: b
+		}.f
+	}
 }
 
 // f64_bits returns the IEEE 754 binary representation of f,
@@ -27,8 +49,14 @@ pub fn f32_from_bits(b u32) f32 {
 // and f64_bits(f64_from_bits(x)) == x.
 @[inline]
 pub fn f64_bits(f f64) u64 {
-	p := *unsafe { &u64(&f) }
-	return p
+	$if tinyc {
+		return *unsafe { &u64(&f) }
+	}
+	return unsafe {
+		U64_F64{
+			f: f
+		}.u
+	}
 }
 
 // f64_from_bits returns the floating-point number corresponding
@@ -37,6 +65,12 @@ pub fn f64_bits(f f64) u64 {
 // f64_from_bits(f64_bits(x)) == x.
 @[inline]
 pub fn f64_from_bits(b u64) f64 {
-	p := *unsafe { &f64(&b) }
-	return p
+	$if tinyc {
+		return *unsafe { &f64(&b) }
+	}
+	return unsafe {
+		U64_F64{
+			u: b
+		}.f
+	}
 }