130 lines
4.0 KiB
Python
130 lines
4.0 KiB
Python
# -*- coding: utf-8 -*-
|
|
"""
|
|
pygments.lexers.zig
|
|
~~~~~~~~~~~~~~~~~~~
|
|
|
|
Lexers for Zig.
|
|
|
|
:copyright: Copyright 2006-2019 by the Pygments team, see AUTHORS.
|
|
:license: BSD, see LICENSE for details.
|
|
"""
|
|
|
|
from pygments.lexer import RegexLexer, words
|
|
from pygments.token import Comment, Operator, Keyword, Name, String, \
|
|
Number, Punctuation, Whitespace
|
|
|
|
__all__ = ['ZigLexer']
|
|
|
|
|
|
class ZigLexer(RegexLexer):
|
|
"""
|
|
For `Zig <http://www.ziglang.org>`_ source code.
|
|
|
|
grammar: https://ziglang.org/documentation/master/#Grammar
|
|
"""
|
|
name = 'Zig'
|
|
aliases = ['zig']
|
|
filenames = ['*.zig']
|
|
mimetypes = ['text/zig']
|
|
|
|
type_keywords = (
|
|
words(('bool', 'f16', 'f32', 'f64', 'f128', 'void', 'noreturn', 'type',
|
|
'anyerror', 'promise', 'i0', 'u0', 'isize', 'usize', 'comptime_int',
|
|
'comptime_float', 'c_short', 'c_ushort', 'c_int', 'c_uint', 'c_long',
|
|
'c_ulong', 'c_longlong', 'c_ulonglong', 'c_longdouble', 'c_void'
|
|
'i8', 'u8', 'i16', 'u16', 'i32', 'u32', 'i64', 'u64', 'i128',
|
|
'u128'), suffix=r'\b'),
|
|
Keyword.Type)
|
|
|
|
storage_keywords = (
|
|
words(('const', 'var', 'extern', 'packed', 'export', 'pub', 'noalias',
|
|
'inline', 'comptime', 'nakedcc', 'stdcallcc', 'volatile', 'allowzero',
|
|
'align', 'linksection', 'threadlocal'), suffix=r'\b'),
|
|
Keyword.Reserved)
|
|
|
|
structure_keywords = (
|
|
words(('struct', 'enum', 'union', 'error'), suffix=r'\b'),
|
|
Keyword)
|
|
|
|
statement_keywords = (
|
|
words(('break', 'return', 'continue', 'asm', 'defer', 'errdefer',
|
|
'unreachable', 'try', 'catch', 'async', 'await', 'suspend',
|
|
'resume', 'cancel'), suffix=r'\b'),
|
|
Keyword)
|
|
|
|
conditional_keywords = (
|
|
words(('if', 'else', 'switch', 'and', 'or', 'orelse'), suffix=r'\b'),
|
|
Keyword)
|
|
|
|
repeat_keywords = (
|
|
words(('while', 'for'), suffix=r'\b'),
|
|
Keyword)
|
|
|
|
other_keywords = (
|
|
words(('fn', 'usingnamespace', 'test'), suffix=r'\b'),
|
|
Keyword)
|
|
|
|
constant_keywords = (
|
|
words(('true', 'false', 'null', 'undefined'), suffix=r'\b'),
|
|
Keyword.Constant)
|
|
|
|
tokens = {
|
|
'root': [
|
|
(r'\n', Whitespace),
|
|
(r'\s+', Whitespace),
|
|
(r'//.*?\n', Comment.Single),
|
|
|
|
# Keywords
|
|
statement_keywords,
|
|
storage_keywords,
|
|
structure_keywords,
|
|
repeat_keywords,
|
|
type_keywords,
|
|
constant_keywords,
|
|
conditional_keywords,
|
|
other_keywords,
|
|
|
|
# Floats
|
|
(r'0x[0-9a-fA-F]+\.[0-9a-fA-F]+([pP][\-+]?[0-9a-fA-F]+)?', Number.Float),
|
|
(r'0x[0-9a-fA-F]+\.?[pP][\-+]?[0-9a-fA-F]+', Number.Float),
|
|
(r'[0-9]+\.[0-9]+([eE][-+]?[0-9]+)?', Number.Float),
|
|
(r'[0-9]+\.?[eE][-+]?[0-9]+', Number.Float),
|
|
|
|
# Integers
|
|
(r'0b[01]+', Number.Bin),
|
|
(r'0o[0-7]+', Number.Oct),
|
|
(r'0x[0-9a-fA-F]+', Number.Hex),
|
|
(r'[0-9]+', Number.Integer),
|
|
|
|
# Identifier
|
|
(r'@[a-zA-Z_]\w*', Name.Builtin),
|
|
(r'[a-zA-Z_]\w*', Name),
|
|
|
|
# Characters
|
|
(r'\'\\\'\'', String.Escape),
|
|
(r'\'\\(|x[a-fA-F0-9]{2}|u[a-fA-F0-9]{4}|U[a-fA-F0-9]{6}|[nr\\t\'"])\'',
|
|
String.Escape),
|
|
(r'\'[^\\\']\'', String),
|
|
|
|
# Strings
|
|
(r'\\\\[^\n]*', String.Heredoc),
|
|
(r'c\\\\[^\n]*', String.Heredoc),
|
|
(r'c?"', String, 'string'),
|
|
|
|
# Operators, Punctuation
|
|
(r'[+%=><|^!?/\-*&~:]', Operator),
|
|
(r'[{}()\[\],.;]', Punctuation)
|
|
],
|
|
'string': [
|
|
(r'\\(x[a-fA-F0-9]{2}|u[a-fA-F0-9]{4}|U[a-fA-F0-9]{6}|[nr\\t\'"])',
|
|
String.Escape),
|
|
(r'[^\\"\n]+', String),
|
|
(r'"', String, '#pop')
|
|
]
|
|
}
|
|
|
|
def get_tokens_unprocessed(self, text):
|
|
for index, token, value in \
|
|
RegexLexer.get_tokens_unprocessed(self, text):
|
|
yield index, token, value
|