Merge pull request #39 from yeslogic/brendanzab/rework-builtins

Rework primitive types and built-ins

examples/ddl/cmap.ddl

@@ -1,21 +1,19 @@
-//! endian: be
-
-Offset32 = u32;
+Offset32 = u32be;
 
 /// https://www.microsoft.com/typography/otspec/cmap.htm
 CMap = struct {
     /// Table version number (0).
-    version: u16,
+    version: u16be,
     /// Number of encoding tables that follow.
-    numTables: u16,
+    numTables: u16be,
     encodingRecords: [EncodingRecord; numTables],
 };
 
 EncodingRecord = struct {
     /// Platform ID.
-    platformID: u16,
+    platformID: u16be,
     /// Platform-specific encoding ID.
-    encodingID: u16,
+    encodingID: u16be,
     /// Byte offset from beginning of table to the subtable for this encoding.
     offset: Offset32,
 };
@@ -35,11 +33,11 @@ CMapSubtable = union {
 /// Format 0: Byte encoding table
 Format0 = struct {
     /// Format number is set to 0.
-    format: u16 where format => format == 0,
+    format: u16be where format => format == 0,
     /// This is the length in bytes of the subtable.
-    length: u16,
+    length: u16be,
     /// Please see “Note on the language field in 'cmap' subtables“ in this document.
-    language: u16,
+    language: u16be,
     /// An array that maps character codes to glyph index values.
     glyphIdArray: [u8; 256],
 };
@@ -51,36 +49,36 @@ Format0 = struct {
 /// Format 6: Trimmed table mapping
 Format6 = struct {
     /// Format number is set to 6.
-    format: u16 where format => format == 6,
+    format: u16be where format => format == 6,
     /// This is the length in bytes of the subtable.
-    length: u16,
+    length: u16be,
     /// Please see “Note on the language field in 'cmap' subtables“ in this
     /// document.
-    language: u16,
+    language: u16be,
     /// First character code of subrange.
-    first_code: u16,
+    first_code: u16be,
     /// Number of character codes in subrange.
-    entry_count: u16,
+    entry_count: u16be,
     /// Array of glyph index values for character codes in the range.
-    glyph_id_array: [u16; entry_count],
+    glyph_id_array: [u16be; entry_count],
 };
 
 /// Format 8: mixed 16-bit and 32-bit coverage
 Format8 = struct {
     /// Subtable format; set to 8.
-    format: u16 where format => format == 8,
+    format: u16be where format => format == 8,
     /// Reserved; set to 0
-    reserved: u16,
+    reserved: u16be,
     /// Byte length of this subtable (including the header)
-    length: u32,
+    length: u32be,
     /// Please see “Note on the language field in 'cmap' subtables“ in this
     /// document.
-    language: u32,
+    language: u32be,
     /// Tightly packed array of bits (8K bytes total) indicating whether the
     /// particular 16-bit (index) value is the start of a 32-bit character code
     is32: [u8; 8192],
     /// Number of groupings which follow
-    num_groups: u32,
+    num_groups: u32be,
     /// Array of SequentialMapGroup records.
     groups: [Format8SequentialMapGroup; num_groups],
 };
@@ -89,83 +87,83 @@ Format8SequentialMapGroup = struct {
     /// First character code in this group; note that if this group is for one
     /// or more 16-bit character codes (which is determined from the is32
     /// array), this 32-bit value will have the high 16-bits set to zero
-    start_char_code: u32,
+    start_char_code: u32be,
     /// Last character code in this group; same condition as listed above for
     /// the `start_char_code`
-    end_char_code: u32,
+    end_char_code: u32be,
     /// Glyph index corresponding to the starting character code
-    start_glyph_id: u32,
+    start_glyph_id: u32be,
 };
 
 // TODO: Format10
 
 // Format 12: Segmented coverage
 Format12 = struct {
     /// Subtable format; set to 12.
-    format: u16 where format => format == 12,
+    format: u16be where format => format == 12,
     /// Reserved; set to 0
-    reserved: u16,
+    reserved: u16be,
     /// Byte length of this subtable (including the header)
-    length: u32,
+    length: u32be,
     /// Please see “Note on the language field in 'cmap' subtables“ in this
     /// document.
-    language: u32,
+    language: u32be,
     /// Number of groupings which follow
-    num_groups: u32,
+    num_groups: u32be,
     /// Array of SequentialMapGroup records.
     groups: [Format12SequentialMapGroup; num_groups],
 };
 
 Format12SequentialMapGroup = struct {
     /// First character code in this group
-    start_char_code: u32,
+    start_char_code: u32be,
     /// Last character code in this group
-    end_char_code: u32,
+    end_char_code: u32be,
     /// Glyph index corresponding to the starting character code
-    start_glyph_id: u32,
+    start_glyph_id: u32be,
 };
 
 // Format 13: Many-to-one range mappings
 Format13 = struct {
     /// Subtable format; set to 13.
-    format: u16 where format => format == 13,
+    format: u16be where format => format == 13,
     /// Reserved; set to 0
-    reserved: u16,
+    reserved: u16be,
     /// Byte length of this subtable (including the header)
-    length: u32,
+    length: u32be,
     /// Please see “Note on the language field in 'cmap' subtables“ in this
     /// document.
-    language: u32,
+    language: u32be,
     /// Number of groupings which follow
-    num_groups: u32,
+    num_groups: u32be,
     /// Array of ConstantMapGroup records.
     groups: [ConstantMapGroup; num_groups],
 };
 
 ConstantMapGroup = struct {
     /// First character code in this group
-    start_char_code: u32,
+    start_char_code: u32be,
     /// Last character code in this group
-    end_char_code: u32,
+    end_char_code: u32be,
     /// Glyph index to be used for all the characters in the group's range.
-    start_glyph_id: u32,
+    start_glyph_id: u32be,
 };
 
 /// Format 14: Unicode Variation Sequences
 Format14 = struct {
     /// Subtable format. Set to 14.
-    format: u16 where format => format == 14,
+    format: u16be where format => format == 14,
     /// Byte length of this subtable (including this header)
-    length: u32,
+    length: u32be,
     /// Number of variation Selector Records
-    num_var_selector_records: u32,
+    num_var_selector_records: u32be,
     /// Array of VariationSelector records.
     var_selector: [VariationSelector; num_var_selector_records],
 };
 
 VariationSelector = struct {
     /// Variation selector
-    var_selector: [u8; 3], // FIXME: should be u24
+    var_selector: [u8; 3], // FIXME: should be u24be
     /// Offset from the start of the format 14 subtable to Default UVS Table. May be 0.
     default_uvs_offset: Offset32,
     /// Offset from the start of the format 14 subtable to Non-Default UVS Table. May be 0.
@@ -175,29 +173,29 @@ VariationSelector = struct {
 /// Default UVS table
 DefaultUVS = struct {
     /// Number of Unicode character ranges.
-    num_unicode_value_ranges: u32,
+    num_unicode_value_ranges: u32be,
     /// Array of UnicodeRange records.
     ranges: [UnicodeRange; num_unicode_value_ranges],
 };
 
 UnicodeRange = struct {
     /// First value in this range
-    start_unicode_value: [u8; 3], // FIXME: should be u24
+    start_unicode_value: [u8; 3], // FIXME: should be u24be
     /// Number of additional values in this range
     additional_count: u8,
 };
 
 /// NonDefaultUVS Table
 NonDefaultUVS = struct {
     /// Number of UVS Mappings that follow
-    num_uvs_mappings: u32,
+    num_uvs_mappings: u32be,
     /// Array of UVSMapping records.
     uvs_mappings: [UVSMapping; num_uvs_mappings]
 };
 
 UVSMapping = struct {
     /// Base Unicode value of the UVS
-    unicode_value: [u8; 3], // FIXME: should be u24
+    unicode_value: [u8; 3], // FIXME: should be u24be
     /// Glyph ID of the UVS
-    glyph_id: u16,
+    glyph_id: u16be,
 };

examples/ddl/edid.ddl

@@ -1,14 +1,12 @@
-//! endian: le
-
 Header = struct {
     /// Fixed header pattern
     magic: [u8; 8], // FIXME: constrain to [0x00, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0x00]
     /// Manufacturer ID
-    mfg_bytes: u16,
+    mfg_bytes: u16le,
     /// Manufacturer product code
-    product_code: u16,
+    product_code: u16le,
     /// Serial number
-    serial: u32,
+    serial: u32le,
     /// Week of manufacture. Week numbering is not consistent between manufacturers.
     mfg_week: u8,
     /// Year of manufacture, less 1990. (1990–2245). If week=255, it is the model year instead.

examples/ddl/heroes_of_might_and_magic_bmp.ddl

@@ -1,9 +1,7 @@
-//! endian: le
-
 /// http://formats.kaitai.io/heroes_of_might_and_magic_bmp/index.html
 HerosOfMightAndMagicBmp = struct {
-    magic: u16,
-    width: u16,
-    height: u16,
+    magic: u16le,
+    width: u16le,
+    height: u16le,
     data: [u8; width * height],
 };

examples/ddl/object_id.ddl

@@ -1,9 +1,7 @@
-//! endian: le
-
 /// https://docs.mongodb.com/manual/reference/method/ObjectId/
 ObjectId = struct {
-    epoch_time: u32,
-    machine_id: [u8; 3], // FIXME: should be u24
-    process_id: u16,
-    counter: [u8; 3], // FIXME: should be u24
+    epoch_time: u32le,
+    machine_id: [u8; 3], // FIXME: should be u24le
+    process_id: u16le,
+    counter: [u8; 3], // FIXME: should be u24le
 };

examples/ddl/stl.ddl

@@ -1,10 +1,9 @@
-//! endian: le
 //! file-extension: stl
 
 Vec3d = struct {
-    x: f32,
-    y: f32,
-    z: f32,
+    x: f32le,
+    y: f32le,
+    z: f32le,
 };
 
 Triangle = struct {
@@ -19,7 +18,7 @@ Stl = struct {
     /// Generally ignored
     header: [u8; 80],
     /// Number of triangles that follow
-    num_triangles: u32,
+    num_triangles: u32le,
     /// The triangle data
     triangles: [Triangle; num_triangles],
 };

src/ir/owned/ast.rs

@@ -49,8 +49,8 @@ pub enum RepeatBound<N> {
 pub enum ParseExpr<N> {
     /// A reference to another parser
     Var(Var<N, u32>),
-    /// Parse a bit
-    Bit,
+    /// Parse a byte
+    U8,
     /// The name of another parsable type
     Ident(N),
     /// Parse that is repeated for the given bound
@@ -150,7 +150,7 @@ impl<'a, N: Name + for<'b> From<&'b str>> From<&'a binary::Type<N>> for ParseExp
 
         match *src {
             Type::Var(_, ref var) => ParseExpr::Var(var.clone()),
-            Type::Const(TypeConst::Bit) => ParseExpr::Bit,
+            Type::Const(TypeConst::U8) => ParseExpr::U8,
             Type::Array(_, ref elem_ty, ref size_expr) => {
                 let elem_parser = ParseExpr::from(&**elem_ty);
                 ParseExpr::repeat(elem_parser, RepeatBound::Exact(size_expr.clone()))

src/syntax/ast/binary.rs

@@ -32,7 +32,7 @@ impl Kind {
 
 #[derive(Debug, Clone, PartialEq, Eq)]
 pub enum TypeConst {
-    Bit,
+    U8,
 }
 
 /// A binary type
@@ -71,9 +71,9 @@ impl<N: Name> Type<N> {
         Type::Var(span, Var::Bound(Named(x.into(), i)))
     }
 
-    /// Bit type constant
-    pub fn bit() -> Type<N> {
-        Type::Const(TypeConst::Bit)
+    /// Byte type constant
+    pub fn u8() -> Type<N> {
+        Type::Const(TypeConst::U8)
     }
 
     /// An array of the specified type, with a size: eg. `[T; n]`
@@ -307,7 +307,7 @@ impl<N: Name> Type<N> {
     pub fn repr(&self) -> host::RcType<N> {
         match *self {
             Type::Var(_, ref v) => Rc::new(host::Type::Var(v.clone())),
-            Type::Const(TypeConst::Bit) => Rc::new(host::Type::Const(host::TypeConst::Bit)),
+            Type::Const(TypeConst::U8) => Rc::new(host::Type::Const(host::TypeConst::U8)),
             Type::Array(_, ref elem_ty, _) => Rc::new(host::Type::Array(elem_ty.repr())),
             Type::Assert(_, ref ty, _) => ty.repr(),
             Type::Interp(_, _, _, ref repr_ty) => repr_ty.clone(),