skorotkiewicz/rDrama
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

nicer and more efficient sanitisation

Browse source
This commit is contained in:
Jack Byrne 2022-09-10 18:56:13 +01:00 • committed by Ben Rog-Wilhelm
parent 8c7c76feb6
commit 1b9f7860c5
3 changed files with 14 additions and 4 deletions
Download patch file Download diff file Expand all files Collapse all files

7
files/helpers/sanitize.py
View file

@ -1,4 +1,5 @@
import functools
import html
import bleach
from bs4 import BeautifulSoup
from bleach.linkifier import LinkifyFilter, build_url_re
@ -166,7 +167,7 @@ def sanitize(sanitized, alert=False, comment=False, edit=False):
sanitized = strikethrough_regex.sub(r'<del>\1</del>', sanitized)
# remove left-to-right mark; remove zero width space; remove zero width no-break space; remove Cuneiform Numeric Sign Eight;
sanitized = sanitized.replace('‎','').replace('​','').replace("\ufeff", "").replace("𒐪","")
sanitized = unwanted_bytes_regex.sub('', sanitized)
if alert:
matches = { g.group(1):g for g in mention_regex2.finditer(sanitized) if g }
@ -340,7 +341,9 @@ def allowed_attributes_emojis(tag, name, value):
@with_sigalrm_timeout(1)
def filter_emojis_only(title, edit=False, graceful=False):
title = title.replace('‎','').replace('​','').replace("\ufeff", "").replace("𒐪","").replace("\n", "").replace("\r", "").replace("\t", "").replace("&", "&amp;").replace('<','&lt;').replace('>','&gt;').replace('"', '&quot;').replace("'", "&#039;").strip()
title = unwanted_bytes_regex.sub('', title)
title = whitespace_regex.sub(' ', title)
title = html.escape(title, quote=True)
# title = render_emoji(title, emoji_regex3, edit)