Distinguish char16_t.

With a new command-line option, this ensures that char16_t
is distinct from uint16_t in generated bindings. On some platforms
these are distinct types, so it can be important for downstream
post processors to spot the difference.

See the documentation on the new command-line option for
expected behavior and usage here.

Part of google/autocxx#124.

Author: adetaylor

bindgen-tests/tests/expectations/tests/char16_t.rs

@@ -0,0 +1,4 @@
+// bindgen-flags: --use-distinct-char16-t --raw-line '#[repr(transparent)] pub struct bindgen_cchar16_t(u16);' -- -x c++ -std=c++14
+
+void receive_char16_t(char16_t input) {
+}

bindgen-tests/tests/headers/char16_t.hpp

@@ -187,6 +187,12 @@ pub(crate) mod ast_ty {
         match ik {
             IntKind::Bool => syn::parse_quote! { bool },
             IntKind::Char { .. } => raw_type(ctx, "c_char"),
+            // The following is used only when an unusual command-line
+            // argument is used. bindgen_cchar16_t is not a real type;
+            // but this allows downstream postprocessors to distinguish
+            // this case and do something special for C++ bindings
+            // containing char16_t.
+            IntKind::Char16 => syn::parse_quote! { bindgen_cchar16_t },
             IntKind::SChar => raw_type(ctx, "c_schar"),
             IntKind::UChar => raw_type(ctx, "c_uchar"),
             IntKind::Short => raw_type(ctx, "c_short"),

bindgen/codegen/helpers.rs

@@ -1980,6 +1980,9 @@ If you encounter an error missing from this list, please file an issue or a PR!"
             CXType_Short => TypeKind::Int(IntKind::Short),
             CXType_UShort => TypeKind::Int(IntKind::UShort),
             CXType_WChar => TypeKind::Int(IntKind::WChar),
+            CXType_Char16 if self.options().use_distinct_char16_t => {
+                TypeKind::Int(IntKind::Char16)
+            }
             CXType_Char16 => TypeKind::Int(IntKind::U16),
             CXType_Char32 => TypeKind::Int(IntKind::U32),
             CXType_Long => TypeKind::Int(IntKind::Long),

bindgen/ir/context.rs

@@ -54,9 +54,12 @@ pub enum IntKind {
     /// A 16-bit signed integer.
     I16,
 
-    /// Either a `char16_t` or a `wchar_t`.
+    /// A 16-bit integer, used only for enum size representation.
     U16,
 
+    /// Either a `char16_t` or a `wchar_t`.
+    Char16,
+
     /// A 32-bit signed integer.
     I32,
 
@@ -94,7 +97,7 @@ impl IntKind {
             // to know whether it is or not right now (unlike char, there's no
             // WChar_S / WChar_U).
             Bool | UChar | UShort | UInt | ULong | ULongLong | U8 | U16 |
-            WChar | U32 | U64 | U128 => false,
+            Char16 | WChar | U32 | U64 | U128 => false,
 
             SChar | Short | Int | Long | LongLong | I8 | I16 | I32 | I64 |
             I128 => true,
@@ -110,7 +113,7 @@ impl IntKind {
         use self::IntKind::*;
         Some(match *self {
             Bool | UChar | SChar | U8 | I8 | Char { .. } => 1,
-            U16 | I16 => 2,
+            U16 | I16 | Char16 => 2,
             U32 | I32 => 4,
             U64 | I64 => 8,
             I128 | U128 => 16,

bindgen/ir/int.rs

@@ -441,6 +441,9 @@ struct BindgenCommand {
     /// Always output explicit padding fields.
     #[arg(long)]
     explicit_padding: bool,
+    /// Use distinct char16_t
+    #[arg(long)]
+    use_distinct_char16_t: bool,
     /// Enables generation of vtable functions.
     #[arg(long)]
     vtable_generation: bool,
@@ -629,6 +632,7 @@ where
         translate_enum_integer_types,
         c_naming,
         explicit_padding,
+        use_distinct_char16_t,
         vtable_generation,
         sort_semantically,
         merge_extern_blocks,
@@ -926,6 +930,7 @@ where
             translate_enum_integer_types,
             c_naming,
             explicit_padding,
+            use_distinct_char16_t,
             vtable_generation,
             sort_semantically,
             merge_extern_blocks,

bindgen/options/cli.rs

@@ -153,6 +153,28 @@ macro_rules! options {
 }
 
 options! {
+    /// Whether we should distinguish between 'char16_t' and 'u16'.
+    /// As standard, bindgen represents `char16_t` as `u16`.
+    /// Rust does not have a `std::os::raw::c_char16_t` type, and thus
+    /// we can't use a built-in Rust type in the generated bindings.
+    /// But for some uses of bindgen, especially when downstream
+    /// post-processing occurs, it's important to distinguish `char16_t`
+    /// from normal `uint16_t`. When this option is enabled, bindgen
+    /// generates a fake type called `bindgen_cchar16_t`. Downstream
+    /// code post-processors should arrange to replace this with a
+    /// real type.
+    use_distinct_char16_t: bool {
+        methods: {
+            /// If this is true, denote 'char16_t' as a separate type from 'u16'
+            /// Disabled by default.
+            pub fn use_distinct_char16_t(mut self, doit: bool) -> Builder {
+                self.options.use_distinct_char16_t = doit;
+                self
+            }
+        },
+        as_args: "--use-distinct-char16-t",
+    },
+
     /// Types that have been blocklisted and should not appear anywhere in the generated code.
     blocklisted_types: RegexSet {
         methods: {