Quelle  lexer.py

# mako/lexer.py
# Copyright 2006-2025 the Mako authors and contributors <see AUTHORS file>
#
# This module is part of Mako and is released under
# the MIT License: http://www.opensource.org/licenses/mit-license.php

"""provides the Lexer class for parsing template strings into parse trees."""

import codecs
import re

from mako import exceptions
from mako import parsetree
from mako.pygen import adjust_whitespace

_regexp_cache = {}


class Lexer:
    def __init__(
        self, text, filename=None, input_encoding=None, preprocessor=None
    ):
        self.text = text
        self.filename = filename
        self.template = parsetree.TemplateNode(self.filename)
        self.matched_lineno = 1
        self.matched_charpos = 0
        self.lineno = 1
        self.match_position = 0
        self.tag = []
        self.control_line = []
        self.ternary_stack = []
        self.encoding = input_encoding

        if preprocessor is None:
            self.preprocessor = []
        elif not hasattr(preprocessor, "__iter__"):
            self.preprocessor = [preprocessor]
        else:
            self.preprocessor = preprocessor

    @property
    def exception_kwargs(self):
        return {
            "source": self.text,
            "lineno": self.matched_lineno,
            "pos": self.matched_charpos,
            "filename": self.filename,
        }

    def match(self, regexp, flags=None):
        """compile the given regexp, cache the reg, and call match_reg()."""

        try:
            reg = _regexp_cache[(regexp, flags)]
        except KeyError:
            reg = re.compile(regexp, flags) if flags else re.compile(regexp)
            _regexp_cache[(regexp, flags)] = reg

        return self.match_reg(reg)

    def match_reg(self, reg):
        """match the given regular expression object to the current text
        position.

        if a match occurs, update the current text and line position.

        """

        mp = self.match_position

        match = reg.match(self.text, self.match_position)
        if match:
            (start, end) = match.span()
            self.match_position = end + 1 if end == start else end
            self.matched_lineno = self.lineno
            cp = mp - 1
            if cp >= 0 and cp < self.textlength:
                cp = self.text[: cp + 1].rfind("\n")
            self.matched_charpos = mp - cp
            self.lineno += self.text[mp : self.match_position].count("\n")
        return match

    def parse_until_text(self, watch_nesting, *text):
        startpos = self.match_position
        text_re = r"|".join(text)
        brace_level = 0
        paren_level = 0
        bracket_level = 0
        while True:
            match = self.match(r"#.*\n")
            if match:
                continue
            match = self.match(
                r"(\"\"\"|\'\'\'|\"|\')[^\\]*?(\\.[^\\]*?)*\1", re.S
            )
            if match:
                continue
            match = self.match(r"(%s)" % text_re)
            if match and not (
                watch_nesting
                and (brace_level > 0 or paren_level > 0 or bracket_level > 0)
            ):
                return (
                    self.text[
                        startpos : self.match_position - len(match.group(1))
                    ],
                    match.group(1),
                )
            elif not match:
                match = self.match(r"(.*?)(?=\"|\'|#|%s)" % text_re, re.S)
            if match:
                brace_level += match.group(1).count("{")
                brace_level -= match.group(1).count("}")
                paren_level += match.group(1).count("(")
                paren_level -= match.group(1).count(")")
                bracket_level += match.group(1).count("[")
                bracket_level -= match.group(1).count("]")
                continue
            raise exceptions.SyntaxException(
                "Expected: %s" % ",".join(text), **self.exception_kwargs
            )

    def append_node(self, nodecls, *args, **kwargs):
        kwargs.setdefault("source", self.text)
        kwargs.setdefault("lineno", self.matched_lineno)
        kwargs.setdefault("pos", self.matched_charpos)
        kwargs["filename"] = self.filename
        node = nodecls(*args, **kwargs)
        if len(self.tag):
            self.tag[-1].nodes.append(node)
        else:
            self.template.nodes.append(node)
        # build a set of child nodes for the control line
        # (used for loop variable detection)
        # also build a set of child nodes on ternary control lines
        # (used for determining if a pass needs to be auto-inserted
        if self.control_line:
            control_frame = self.control_line[-1]
            control_frame.nodes.append(node)
            if (
                not (
                    isinstance(node, parsetree.ControlLine)
                    and control_frame.is_ternary(node.keyword)
                )
                and self.ternary_stack
                and self.ternary_stack[-1]
            ):
                self.ternary_stack[-1][-1].nodes.append(node)
        if isinstance(node, parsetree.Tag):
            if len(self.tag):
                node.parent = self.tag[-1]
            self.tag.append(node)
        elif isinstance(node, parsetree.ControlLine):
            if node.isend:
                self.control_line.pop()
                self.ternary_stack.pop()
            elif node.is_primary:
                self.control_line.append(node)
                self.ternary_stack.append([])
            elif self.control_line and self.control_line[-1].is_ternary(
                node.keyword
            ):
                self.ternary_stack[-1].append(node)
            elif self.control_line and not self.control_line[-1].is_ternary(
                node.keyword
            ):
                raise exceptions.SyntaxException(
                    "Keyword '%s' not a legal ternary for keyword '%s'"
                    % (node.keyword, self.control_line[-1].keyword),
                    **self.exception_kwargs,
                )

    _coding_re = re.compile(r"#.*coding[:=]\s*([-\w.]+).*\r?\n")

    def decode_raw_stream(self, text, decode_raw, known_encoding, filename):
        """given string/unicode or bytes/string, determine encoding
        from magic encoding comment, return body as unicode
        or raw if decode_raw=False

        """
        if isinstance(text, str):
            m = self._coding_re.match(text)
            encoding = m and m.group(1) or known_encoding or "utf-8"
            return encoding, text

        if text.startswith(codecs.BOM_UTF8):
            text = text[len(codecs.BOM_UTF8) :]
            parsed_encoding = "utf-8"
            m = self._coding_re.match(text.decode("utf-8", "ignore"))
            if m is not None and m.group(1) != "utf-8":
                raise exceptions.CompileException(
                    "Found utf-8 BOM in file, with conflicting "
                    "magic encoding comment of '%s'" % m.group(1),
                    text.decode("utf-8", "ignore"),
                    0,
                    0,
                    filename,
                )
        else:
            m = self._coding_re.match(text.decode("utf-8", "ignore"))
            parsed_encoding = m.group(1) if m else known_encoding or "utf-8"
        if decode_raw:
            try:
                text = text.decode(parsed_encoding)
            except UnicodeDecodeError:
                raise exceptions.CompileException(
                    "Unicode decode operation of encoding '%s' failed"
                    % parsed_encoding,
                    text.decode("utf-8", "ignore"),
                    0,
                    0,
                    filename,
                )

        return parsed_encoding, text

    def parse(self):
        self.encoding, self.text = self.decode_raw_stream(
            self.text, True, self.encoding, self.filename
        )

        for preproc in self.preprocessor:
            self.text = preproc(self.text)

        # push the match marker past the
        # encoding comment.
        self.match_reg(self._coding_re)

        self.textlength = len(self.text)

        while True:
            if self.match_position > self.textlength:
                break

            if self.match_end():
                break
            if self.match_expression():
                continue
            if self.match_control_line():
                continue
            if self.match_comment():
                continue
            if self.match_tag_start():
                continue
            if self.match_tag_end():
                continue
            if self.match_python_block():
                continue
            if self.match_percent():
                continue
            if self.match_text():
                continue

            if self.match_position > self.textlength:
                break
            # TODO: no coverage here
            raise exceptions.MakoException("assertion failed")

        if len(self.tag):
            raise exceptions.SyntaxException(
                "Unclosed tag: <%%%s>" % self.tag[-1].keyword,
                **self.exception_kwargs,
            )
        if len(self.control_line):
            raise exceptions.SyntaxException(
                "Unterminated control keyword: '%s'"
                % self.control_line[-1].keyword,
                self.text,
                self.control_line[-1].lineno,
                self.control_line[-1].pos,
                self.filename,
            )
        return self.template

    def match_tag_start(self):
        reg = r"""
            \<%     # opening tag

            ([\w\.\:]+)   # keyword

            ((?:\s+\w+|\s*=\s*|"[^"]*?"|'[^']*?'|\s*,\s*)*)  # attrname, = \
                                               #        sign, string expression
                                               # comma is for backwards compat
                                               # identified in #366

            \s*     # more whitespace

            (/)?>   # closing

        """

        match = self.match(
            reg,
            re.I | re.S | re.X,
        )

        if not match:
            return False

        keyword, attr, isend = match.groups()
        self.keyword = keyword
        attributes = {}
        if attr:
            for att in re.findall(
                r"\s*(\w+)\s*=\s*(?:'([^']*)'|\"([^\"]*)\")", attr
            ):
                key, val1, val2 = att
                text = val1 or val2
                text = text.replace("\r\n", "\n")
                attributes[key] = text
        self.append_node(parsetree.Tag, keyword, attributes)
        if isend:
            self.tag.pop()
        elif keyword == "text":
            match = self.match(r"(.*?)(?=\</%text>)", re.S)
            if not match:
                raise exceptions.SyntaxException(
                    "Unclosed tag: <%%%s>" % self.tag[-1].keyword,
                    **self.exception_kwargs,
                )
            self.append_node(parsetree.Text, match.group(1))
            return self.match_tag_end()
        return True

    def match_tag_end(self):
        match = self.match(r"\</%[\t ]*([^\t ]+?)[\t ]*>")
        if match:
            if not len(self.tag):
                raise exceptions.SyntaxException(
                    "Closing tag without opening tag: </%%%s>"
                    % match.group(1),
                    **self.exception_kwargs,
                )
            elif self.tag[-1].keyword != match.group(1):
                raise exceptions.SyntaxException(
                    "Closing tag </%%%s> does not match tag: <%%%s>"
                    % (match.group(1), self.tag[-1].keyword),
                    **self.exception_kwargs,
                )
            self.tag.pop()
            return True
        else:
            return False

    def match_end(self):
        match = self.match(r"\Z", re.S)
        if not match:
            return False

        string = match.group()
        if string:
            return string
        else:
            return True

    def match_percent(self):
        match = self.match(r"(?<=^)(\s*)%%(%*)", re.M)
        if match:
            self.append_node(
                parsetree.Text, match.group(1) + "%" + match.group(2)
            )
            return True
        else:
            return False

    def match_text(self):
        match = self.match(
            r"""
                (.*?)         # anything, followed by:
                (
                 (?<=\n)(?=[ \t]*(?=%|\#\#))  # an eval or line-based
                                            # comment, preceded by a
                                            # consumed newline and whitespace
                 |
                 (?=\${)      # an expression
                 |
                 (?=</?%)  # a substitution or block or call start or end
                              # - don't consume
                 |
                 (\\\r?\n)    # an escaped newline  - throw away
                 |
                 \Z           # end of string
                )""",
            re.X | re.S,
        )

        if match:
            text = match.group(1)
            if text:
                self.append_node(parsetree.Text, text)
            return True
        else:
            return False

    def match_python_block(self):
        match = self.match(r"<%(!)?")
        if match:
            line, pos = self.matched_lineno, self.matched_charpos
            text, end = self.parse_until_text(False, r"%>")
            # the trailing newline helps
            # compiler.parse() not complain about indentation
            text = adjust_whitespace(text) + "\n"
            self.append_node(
                parsetree.Code,
                text,
                match.group(1) == "!",
                lineno=line,
                pos=pos,
            )
            return True
        else:
            return False

    def match_expression(self):
        match = self.match(r"\${")
        if not match:
            return False

        line, pos = self.matched_lineno, self.matched_charpos
        text, end = self.parse_until_text(True, r"\|", r"}")
        if end == "|":
            escapes, end = self.parse_until_text(True, r"}")
        else:
            escapes = ""
        text = text.replace("\r\n", "\n")
        self.append_node(
            parsetree.Expression,
            text,
            escapes.strip(),
            lineno=line,
            pos=pos,
        )
        return True

    def match_control_line(self):
        match = self.match(
            r"(?<=^)[\t ]*(%(?!%)|##)[\t ]*((?:(?:\\\r?\n)|[^\r\n])*)"
            r"(?:\r?\n|\Z)",
            re.M,
        )
        if not match:
            return False

        operator = match.group(1)
        text = match.group(2)
        if operator == "%":
            m2 = re.match(r"(end)?(\w+)\s*(.*)", text)
            if not m2:
                raise exceptions.SyntaxException(
                    "Invalid control line: '%s'" % text,
                    **self.exception_kwargs,
                )
            isend, keyword = m2.group(1, 2)
            isend = isend is not None

            if isend:
                if not len(self.control_line):
                    raise exceptions.SyntaxException(
                        "No starting keyword '%s' for '%s'" % (keyword, text),
                        **self.exception_kwargs,
                    )
                elif self.control_line[-1].keyword != keyword:
                    raise exceptions.SyntaxException(
                        "Keyword '%s' doesn't match keyword '%s'"
                        % (text, self.control_line[-1].keyword),
                        **self.exception_kwargs,
                    )
            self.append_node(parsetree.ControlLine, keyword, isend, text)
        else:
            self.append_node(parsetree.Comment, text)
        return True

    def match_comment(self):
        """matches the multiline version of a comment"""
        match = self.match(r"<%doc>(.*?)</%doc>", re.S)
        if match:
            self.append_node(parsetree.Comment, match.group(1))
            return True
        else:
            return False