CSY/reowolf Changeset - a52ba09256ad · Centrum Wiskunde & Informatica (CWI)

Changeset - a52ba09256ad

Parent rev.

Child rev.

[Not reviewed]

0 2 0

MH - 4 years ago 2021-04-23 10:26:53
henger@cwi.nl

WIP on compiler rearchitecting

2 files changed with 102 insertions and 8 deletions:

src/protocol/parser/pass_definitions.rs

src/protocol/parser/token_parsing.rs

0 comments (0 inline, 0 general)

src/protocol/parser/pass_definitions.rs

➞

Show inline comments

@@ @@ -692,51 +692,68 @@ impl PassDefinitions { @@
             let mut expressions = Vec::new();
             consume_comma_separated(
                 TokenKind::OpenCurly, TokenKind::CloseCurly, &module.source, iter,
                 |source, iter| self.consume_expression(module, iter, ctx),
                 &mut expressions, "an expression", "a list of expressions", Some(&mut end_pos)
             )?;
             // TODO: Turn into literal
             result = ctx.heap.alloc_array_expression(|this| ArrayExpression{
                 this,
                 span: InputSpan::from_positions(start_pos, end_pos),
                 elements: expressions,
                 parent: ExpressionParent::None,
                 concrete_type: ConcreteType::default(),
             }).upcast();
         } else if next == Some(TokenKind::Integer) {
             let (literal, span) = consume_integer_literal(&module.source, iter, &mut self.buffer)?;
             result = ctx.heap.alloc_literal_expression(|this| LiteralExpression{
                 this, span,
                 value: Literal::Integer(LiteralInteger{ unsigned_value: literal, negated: false }),
                 parent: ExpressionParent::None,
                 concrete_type: ConcreteType::default(),
             }).upcast();
         } else if next == Some(TokenKind::String) {
             let (text, span) = consume_string_literal(&module.source, iter, &mut self.buffer)?;
             let span = consume_string_literal(&module.source, iter, &mut self.buffer)?;
             let interned = ctx.pool.intern(self.buffer.as_bytes());
             result = ctx.heap.alloc_literal_expression(|this| LiteralExpression{
                 this, span,
                 value: Literal::String(interned),
                 parent: ExpressionParent::None,
                 concrete_type: ConcreteType::default(),
             }).upcast();
         } else if next == Some(TokenKind::Character) {
             let (character, span) = consume_character_literal(&module.source, iter)?;
             result = ctx.heap.alloc_literal_expression(|this| LiteralExpression{
                 this, span,
                 value: Literal::Character(character),
                 parent: ExpressionParent::None,
                 concrete_type: ConcreteType::default(),
             }).upcast();
         } else if next == Some(TokenKind::Ident) {
             // May be a variable, a type instantiation or a function call. If we
             // have a single identifier that we cannot find in the type table
             // then we're going to assume that we're dealign with a variable.
+        }
         Ok(result)
+    }
     //--------------------------------------------------------------------------
     // Expression Utilities
     //--------------------------------------------------------------------------
     #[inline]
     fn consume_generic_binary_expression<
         M: Fn(Option<TokenKind>) -> Option<BinaryOperator>,
         F: Fn(&mut PassDefinitions, &Module, &mut TokenIter, &mut PassCtx) -> Result<ExpressionId, ParseError>
     >(
         &mut self, module: &Module, iter: &mut TokenIter, ctx: &mut PassCtx, match_fn: M, higher_precedence_fn: F
     ) -> Result<ExpressionId, ParseError> {
         let mut result = higher_precedence_fn(self, module, iter, ctx)?;
         while let Some(operation) = match_fn(iter.next()) {
             let span = iter.next_span();
             iter.consume();
             let left = result;
             let right = higher_precedence_fn(self, module, iter, ctx)?;

src/protocol/parser/token_parsing.rs

➞

Show inline comments

         if char == b'_' {
             continue;
+        }
         if !char.is_ascii_digit() {
             return Err(ParseError::new_error_at_span(
                 source, integer_span,
                 format!("incorrectly formatted {} number", radix_name)
             ));
+        }
         buffer.push(char::from(char));
+    }
     // Use the cleaned up string to convert to integer
     match u64::from_str_radix(&buffer, radix) {
         Ok(number) => Ok((number, integer_span)),
         Err(_) => Err(ParseError::new_error_at_span(
             source, integer_span,
             format!("incorrectly formatted {} number", radix_name)
         )),
+    }
+}
 /// Consumes a character literal. We currently support a limited number of
 /// backslash-escaped characters
 pub(crate) fn consume_character_literal(source: &InputSource, iter: &mut TokenIter, buffer: &mut String) -> Result<char, ParseError> {
 pub(crate) fn consume_character_literal(
     source: &InputSource, iter: &mut TokenIter
 ) -> Result<(char, InputSpan), ParseError> {
     if Some(TokenKind::Character) != iter.next() {
         return Err(ParseError::new_error_str_at_pos(source, iter.last_valid_pos(), "expected a character literal"));
+    }
-    let char_span = iter.next_span();
+    let span = iter.next_span();
     iter.consume();
     let char_text = source.section_at_span(char_span);
     let char_text = source.section_at_span(span);
     if !char_text.is_ascii() {
         return Err(ParseError::new_error_str_at_span(
             source, span, "expected an ASCII character literal"
         ));
+    }
     match char_text.len() {
 => return Err(ParseError::new_error_str_at_span(source, span, "too little characters in character literal")),
 => {
             // We already know the text is ascii, so just throw an error if we have the escape
             // character.
             if char_text[0] == b'\\' {
                 return Err(ParseError::new_error_str_at_span(source, span, "escape character without subsequent character"));
+            }
             return Ok((char_text[0] as char, span));
         },
 => {
             if char_text[0] == b'\\' {
                 let result = parse_escaped_character(char_text[1])?;
                 return Ok((result, span))
+            }
         },
         _ => {}
+    }
-    //
+    return Err(ParseError::new_error_str_at_span(source, span, "too many characters in character literal"))
+}
 /// Consumes a string literal. We currently support a limited number of
 /// backslash-escaped characters.
 pub(crate) fn consume_string_literal(source: &InputSource, iter: &mut TokenIter, buffer: &mut String) -> Result<(>
 /// backslash-escaped characters. Note that the result is stored in the
 /// buffer.
 pub(crate) fn consume_string_literal(
     source: &InputSource, iter: &mut TokenIter, buffer: &mut String
 ) -> Result<InputSpan, ParseError> {
     if Some(TokenKind::String) != iter.next() {
         return Err(ParseError::new_error_str_at_pos(source, iter.last_valid_pos(), "expected a string literal"));
+    }
     buffer.clear();
     let span = iter.next_span();
     iter.consume();
     let text = source.section_at_span(span);
     if !text.is_ascii() {
         return Err(ParseError::new_error_str_at_span(source, span, "expected an ASCII string literal"));
+    }
     buffer.reserve(text.len());
     let mut was_escape = false;
     for idx in 0..text.len() {
         let cur = text[idx];
         if cur != b'\\' {
             if was_escape {
                 let to_push = parse_escaped_character(cur)?;
                 buffer.push(to_push);
             } else {
                 buffer.push(cur as char);
+            }
             was_escape = false;
         } else {
             was_escape = true;
+        }
+    }
     Ok(span)
+}
 fn parse_escaped_character(v: u8) -> Result<char, ParseError> {
     let result = match v {
         b'r' => '\r',
         b'n' => '\n',
         b't' => '\t',
         b'0' => '\0',
         b'\\' => '\\',
         b'\'' => '\'',
         b'"' => '"',
         v => return Err(ParseError::new_error_at_span(
             source, span, format!("unexpected escaped character '{}'", v)
         )),
     };
     Ok(result)
+}
 pub(crate) fn consume_pragma<'a>(source: &'a InputSource, iter: &mut TokenIter) -> Result<(&'a [u8], InputPosition, InputPosition), ParseError> {
     if Some(TokenKind::Pragma) != iter.next() {
         return Err(ParseError::new_error_str_at_pos(source, iter.last_valid_pos(), "expected a pragma"));
+    }
     let (pragma_start, pragma_end) = iter.next_positions();
     iter.consume();
     Ok((source.section_at_pos(pragma_start, pragma_end), pragma_start, pragma_end))
+}
 pub(crate) fn has_ident(source: &InputSource, iter: &mut TokenIter, expected: &[u8]) -> bool {
     peek_ident(source, iter).map_or(false, |section| section == expected)
+}
 pub(crate) fn peek_ident<'a>(source: &'a InputSource, iter: &mut TokenIter) -> Option<&'a [u8]> {
     if Some(TokenKind::Ident) == iter.next() {
         let (start, end) = iter.next_positions();
         return Some(source.section_at_pos(start, end))
+    }
     None
+}
 /// Consumes any identifier and returns it together with its span. Does not

0 comments (0 inline, 0 general)