LonamiWebs/Telethon
1
1
Fork 0
mirror of https://github.com/LonamiWebs/Telethon.git synced 2025-06-24 23:43:10 +03:00
Code Issues Packages Projects Releases Wiki Activity

Fix unparsing text with malformed message entities

Browse Source
This commit is contained in:
Lonami Exo 2019-12-19 15:48:27 +01:00
parent ccbc1c669c
commit f3111f93b2
2 changed files with 27 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

22
telethon/extensions/html.py
View File

@ -168,9 +168,21 @@ def unparse(text: str, entities: Iterable[TypeMessageEntity], _offset: int = 0,
continue continue
skip_entity = False skip_entity = False
entity_text = unparse(text=text[relative_offset:relative_offset + entity.length], length = entity.length
# If we are in the middle of a surrogate nudge the position by +1.
# Otherwise we would end up with malformed text and fail to encode.
# For example of bad input: "Hi \ud83d\ude1c"
# https://en.wikipedia.org/wiki/UTF-16#U+010000_to_U+10FFFF
if '\ud800' <= text[relative_offset] <= '\udfff':
relative_offset += 1
if '\ud800' <= text[relative_offset + length] <= '\udfff':
length += 1
entity_text = unparse(text=text[relative_offset:relative_offset + length],
entities=entities[i + 1:], entities=entities[i + 1:],
_offset=entity.offset, _length=entity.length) _offset=entity.offset, _length=length)
entity_type = type(entity) entity_type = type(entity)
if entity_type == MessageEntityBold: if entity_type == MessageEntityBold:
@ -208,6 +220,10 @@ def unparse(text: str, entities: Iterable[TypeMessageEntity], _offset: int = 0,
.format(entity.user_id, entity_text)) .format(entity.user_id, entity_text))
else: else:
skip_entity = True skip_entity = True
last_offset = relative_offset + (0 if skip_entity else entity.length) last_offset = relative_offset + (0 if skip_entity else length)
if last_offset < len(text) and '\ud800' <= text[last_offset] <= '\udfff':
last_offset += 1
html.append(escape(text[last_offset:])) html.append(escape(text[last_offset:]))
return _del_surrogate(''.join(html)) return _del_surrogate(''.join(html))

8
telethon/extensions/markdown.py
View File

@ -184,6 +184,14 @@ def unparse(text, entities, delimiters=None, url_fmt=None):
insert_at.sort(key=lambda t: t[0]) insert_at.sort(key=lambda t: t[0])
while insert_at: while insert_at:
at, what = insert_at.pop() at, what = insert_at.pop()
# If we are in the middle of a surrogate nudge the position by +1.
# Otherwise we would end up with malformed text and fail to encode.
# For example of bad input: "Hi \ud83d\ude1c"
# https://en.wikipedia.org/wiki/UTF-16#U+010000_to_U+10FFFF
if '\ud800' <= text[at] <= '\udfff':
at += 1
text = text[:at] + what + text[at:] text = text[:at] + what + text[at:]
return del_surrogate(text) return del_surrogate(text)