You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
107 lines
3.7 KiB
107 lines
3.7 KiB
9 months ago
|
"""
|
||
|
pygments.lexers.dns
|
||
|
~~~~~~~~~~~~~~~~~~~
|
||
|
|
||
|
Pygments lexers for DNS
|
||
|
|
||
|
:copyright: Copyright 2006-2023 by the Pygments team, see AUTHORS.
|
||
|
:license: BSD, see LICENSE for details.
|
||
|
"""
|
||
|
|
||
|
import re
|
||
|
|
||
|
from pygments.token import Comment, Operator, Keyword, Name, String, \
|
||
|
Number, Punctuation, Whitespace, Literal
|
||
|
from pygments.lexer import RegexLexer, bygroups, include
|
||
|
|
||
|
__all__ = ['DnsZoneLexer']
|
||
|
|
||
|
|
||
|
CLASSES = [
|
||
|
"IN",
|
||
|
"CS",
|
||
|
"CH",
|
||
|
"HS",
|
||
|
]
|
||
|
|
||
|
CLASSES_RE = "(" + "|".join(CLASSES) + ')'
|
||
|
|
||
|
|
||
|
class DnsZoneLexer(RegexLexer):
|
||
|
|
||
|
"""
|
||
|
Lexer for DNS zone file
|
||
|
|
||
|
.. versionadded:: 2.16
|
||
|
"""
|
||
|
|
||
|
flags = re.MULTILINE
|
||
|
|
||
|
name = 'Zone'
|
||
|
aliases = ['zone']
|
||
|
filenames = [ "*.zone" ]
|
||
|
url = "https://datatracker.ietf.org/doc/html/rfc1035"
|
||
|
mimetypes = ['text/dns']
|
||
|
|
||
|
tokens = {
|
||
|
'root': [
|
||
|
# Empty/comment line:
|
||
|
(r'([ \t]*)(;.*)(\n)', bygroups(Whitespace, Comment.Single, Whitespace)),
|
||
|
# Special directives:
|
||
|
(r'^\$ORIGIN\b', Keyword, 'values'),
|
||
|
(r'^\$TTL\b', Keyword, 'values'),
|
||
|
(r'^\$INCLUDE\b', Comment.Preproc, 'include'),
|
||
|
# TODO, $GENERATE https://bind9.readthedocs.io/en/v9.18.14/chapter3.html#soa-rr
|
||
|
(r'^\$[A-Z]+\b', Keyword, 'values'),
|
||
|
# Records:
|
||
|
# <domain-name> [<TTL>] [<class>] <type> <RDATA> [<comment>]
|
||
|
(r'^(@)([ \t]+)(?:([0-9]+[smhdw]?)([ \t]+))?(?:' + CLASSES_RE + "([ \t]+))?([A-Z]+)([ \t]+)",
|
||
|
bygroups(Operator, Whitespace, Number.Integer, Whitespace, Name.Class, Whitespace, Keyword.Type, Whitespace),
|
||
|
"values"),
|
||
|
(r'^([^ \t\n]*)([ \t]+)(?:([0-9]+[smhdw]?)([ \t]+))?(?:' + CLASSES_RE + "([ \t]+))?([A-Z]+)([ \t]+)",
|
||
|
bygroups(Name, Whitespace, Number.Integer, Whitespace, Name.Class, Whitespace, Keyword.Type, Whitespace),
|
||
|
"values"),
|
||
|
# <domain-name> [<class>] [<TTL>] <type> <RDATA> [<comment>]
|
||
|
(r'^(Operator)([ \t]+)(?:' + CLASSES_RE + "([ \t]+))?(?:([0-9]+[smhdw]?)([ \t]+))?([A-Z]+)([ \t]+)",
|
||
|
bygroups(Name, Whitespace, Number.Integer, Whitespace, Name.Class, Whitespace, Keyword.Type, Whitespace),
|
||
|
"values"),
|
||
|
(r'^([^ \t\n]*)([ \t]+)(?:' + CLASSES_RE + "([ \t]+))?(?:([0-9]+[smhdw]?)([ \t]+))?([A-Z]+)([ \t]+)",
|
||
|
bygroups(Name, Whitespace, Number.Integer, Whitespace, Name.Class, Whitespace, Keyword.Type, Whitespace),
|
||
|
"values"),
|
||
|
],
|
||
|
# Parsing values:
|
||
|
'values': [
|
||
|
(r'\n', Whitespace, "#pop"),
|
||
|
(r'\(', Punctuation, 'nested'),
|
||
|
include('simple-values'),
|
||
|
],
|
||
|
# Parsing nested values (...):
|
||
|
'nested': [
|
||
|
(r'\)', Punctuation, "#pop"),
|
||
|
include('simple-values'),
|
||
|
],
|
||
|
# Parsing values:
|
||
|
'simple-values': [
|
||
|
(r'(;.*)(\n)', bygroups(Comment.Single, Whitespace)),
|
||
|
(r'[ \t]+', Whitespace),
|
||
|
(r"@\b", Operator),
|
||
|
('"', String, 'string'),
|
||
|
(r'[0-9]+[smhdw]?$', Number.Integer),
|
||
|
(r'([0-9]+[smhdw]?)([ \t]+)', bygroups(Number.Integer, Whitespace)),
|
||
|
(r'\S+', Literal),
|
||
|
],
|
||
|
'include': [
|
||
|
(r'([ \t]+)([^ \t\n]+)([ \t]+)([-\._a-zA-Z]+)([ \t]+)(;.*)?$',
|
||
|
bygroups(Whitespace, Comment.PreprocFile, Whitespace, Name, Whitespace, Comment.Single), '#pop'),
|
||
|
(r'([ \t]+)([^ \t\n]+)([ \t\n]+)$', bygroups(Whitespace, Comment.PreprocFile, Whitespace), '#pop'),
|
||
|
],
|
||
|
"string": [
|
||
|
(r'\\"', String),
|
||
|
(r'"', String, "#pop"),
|
||
|
(r'[^"]+', String),
|
||
|
]
|
||
|
}
|
||
|
|
||
|
def analyse_text(text):
|
||
|
return text.startswith("$ORIGIN")
|