1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
|
# -*- coding: utf-8 -*-
"""
pygments.lexers.whiley
~~~~~~~~~~~~~~~~~~~~~~
Lexers for the Whiley language.
:copyright: Copyright 2006-2019 by the Pygments team, see AUTHORS.
:license: BSD, see LICENSE for details.
"""
from pygments.lexer import RegexLexer, bygroups, words
from pygments.token import Comment, Keyword, Name, Number, Operator, \
Punctuation, String, Text
__all__ = ['WhileyLexer']
class WhileyLexer(RegexLexer):
"""
Lexer for the Whiley programming language.
.. versionadded:: 2.2
"""
name = 'Whiley'
filenames = ['*.whiley']
aliases = ['whiley']
mimetypes = ['text/x-whiley']
# See the language specification:
# http://whiley.org/download/WhileyLanguageSpec.pdf
tokens = {
'root': [
# Whitespace
(r'\s+', Text),
# Comments
(r'//.*', Comment.Single),
# don't parse empty comment as doc comment
(r'/\*\*/', Comment.Multiline),
(r'(?s)/\*\*.*?\*/', String.Doc),
(r'(?s)/\*.*?\*/', Comment.Multiline),
# Keywords
(words((
'if', 'else', 'while', 'for', 'do', 'return',
'switch', 'case', 'default', 'break', 'continue',
'requires', 'ensures', 'where', 'assert', 'assume',
'all', 'no', 'some', 'in', 'is', 'new',
'throw', 'try', 'catch', 'debug', 'skip', 'fail',
'finite', 'total'), suffix=r'\b'), Keyword.Reserved),
(words((
'function', 'method', 'public', 'private', 'protected',
'export', 'native'), suffix=r'\b'), Keyword.Declaration),
# "constant" & "type" are not keywords unless used in declarations
(r'(constant|type)(\s+)([a-zA-Z_]\w*)(\s+)(is)\b',
bygroups(Keyword.Declaration, Text, Name, Text, Keyword.Reserved)),
(r'(true|false|null)\b', Keyword.Constant),
(r'(bool|byte|int|real|any|void)\b', Keyword.Type),
# "from" is not a keyword unless used with import
(r'(import)(\s+)(\*)([^\S\n]+)(from)\b',
bygroups(Keyword.Namespace, Text, Punctuation, Text, Keyword.Namespace)),
(r'(import)(\s+)([a-zA-Z_]\w*)([^\S\n]+)(from)\b',
bygroups(Keyword.Namespace, Text, Name, Text, Keyword.Namespace)),
(r'(package|import)\b', Keyword.Namespace),
# standard library: https://github.com/Whiley/WhileyLibs/
(words((
# types defined in whiley.lang.Int
'i8', 'i16', 'i32', 'i64',
'u8', 'u16', 'u32', 'u64',
'uint', 'nat',
# whiley.lang.Any
'toString'), suffix=r'\b'), Name.Builtin),
# byte literal
(r'[01]+b', Number.Bin),
# decimal literal
(r'[0-9]+\.[0-9]+', Number.Float),
# match "1." but not ranges like "3..5"
(r'[0-9]+\.(?!\.)', Number.Float),
# integer literal
(r'0x[0-9a-fA-F]+', Number.Hex),
(r'[0-9]+', Number.Integer),
# character literal
(r"""'[^\\]'""", String.Char),
(r"""(')(\\['"\\btnfr])(')""",
bygroups(String.Char, String.Escape, String.Char)),
# string literal
(r'"', String, 'string'),
# operators and punctuation
(r'[{}()\[\],.;]', Punctuation),
(u'[+\\-*/%&|<>^!~@=:?'
# unicode operators
u'\u2200\u2203\u2205\u2282\u2286\u2283\u2287'
u'\u222A\u2229\u2264\u2265\u2208\u2227\u2228'
u']', Operator),
# identifier
(r'[a-zA-Z_]\w*', Name),
],
'string': [
(r'"', String, '#pop'),
(r'\\[btnfr]', String.Escape),
(r'\\u[0-9a-fA-F]{4}', String.Escape),
(r'\\.', String),
(r'[^\\"]+', String),
],
}
|