From 12e8343068b906f8b2afddc5569968a8a91fa5b0 Mon Sep 17 00:00:00 2001 From: Daniel Baumann Date: Mon, 29 Apr 2024 06:24:24 +0200 Subject: Adding upstream version 2.1.0. Signed-off-by: Daniel Baumann --- markdown_it/rules_inline/entity.py | 54 ++++++++++++++++++++++++++++++++++++++ 1 file changed, 54 insertions(+) create mode 100644 markdown_it/rules_inline/entity.py (limited to 'markdown_it/rules_inline/entity.py') diff --git a/markdown_it/rules_inline/entity.py b/markdown_it/rules_inline/entity.py new file mode 100644 index 0000000..883a966 --- /dev/null +++ b/markdown_it/rules_inline/entity.py @@ -0,0 +1,54 @@ +# Process html entity - {, ¯, ", ... +import re + +from ..common.entities import entities +from ..common.utils import fromCodePoint, isValidEntityCode +from .state_inline import StateInline + +DIGITAL_RE = re.compile(r"^&#((?:x[a-f0-9]{1,6}|[0-9]{1,7}));", re.IGNORECASE) +NAMED_RE = re.compile(r"^&([a-z][a-z0-9]{1,31});", re.IGNORECASE) + + +def entity(state: StateInline, silent: bool): + + pos = state.pos + maximum = state.posMax + + if state.srcCharCode[pos] != 0x26: # /* & */ + return False + + if (pos + 1) < maximum: + ch = state.srcCharCode[pos + 1] + + if ch == 0x23: # /* # */ + match = DIGITAL_RE.search(state.src[pos:]) + if match: + if not silent: + match1 = match.group(1) + code = ( + int(match1[1:], 16) + if match1[0].lower() == "x" + else int(match1, 10) + ) + state.pending += ( + fromCodePoint(code) + if isValidEntityCode(code) + else fromCodePoint(0xFFFD) + ) + + state.pos += len(match.group(0)) + return True + + else: + match = NAMED_RE.search(state.src[pos:]) + if match: + if match.group(1) in entities: + if not silent: + state.pending += entities[match.group(1)] + state.pos += len(match.group(0)) + return True + + if not silent: + state.pending += "&" + state.pos += 1 + return True -- cgit v1.2.3