Backup_Repos/leech
1
0
Fork 0
mirror of https://github.com/kemayo/leech synced 2025-12-06 16:33:16 +01:00
Code Issues Projects Releases Wiki Activity

Improve cloudflare email decoding

Browse source 
New format for the protected emails, wrapping a span in an a.
This commit is contained in:
David Lynch 2024-01-28 13:26:34 -06:00
parent d30e56a518
commit 64d77b62db
1 changed files with 8 additions and 4 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

12
sites/__init__.py
View file

@ -248,12 +248,16 @@ class Site:
# Cloudflare is used on many sites, and mangles things that look like email addresses
# e.g. Point_Me_@_The_Sky becomes
# <a href="/cdn-cgi/l/email-protection" class="__cf_email__" data-cfemail="85d5eaecebf1dac8e0dac5">[email&#160;protected]</a>_The_Sky
for a in contents.find_all('a', class_='__cf_email__', href='/cdn-cgi/l/email-protection'):
# or
# <a href="/cdn-cgi/l/email-protection#85d5eaecebf1dac8e0dac5"><span class="__cf_email__" data-cfemail="85d5eaecebf1dac8e0dac5">[email&#160;protected]</span></a>_The_Sky
for tag in contents.find_all(class_='__cf_email__'):
# See: https://usamaejaz.com/cloudflare-email-decoding/
enc = bytes.fromhex(a['data-cfemail'])
enc = bytes.fromhex(tag['data-cfemail'])
email = bytes([c ^ enc[0] for c in enc[1:]]).decode('utf8')
a.insert_before(email)
a.decompose()
if tag.parent.name == 'a' and tag.parent['href'].startswith('/cdn-cgi/l/email-protection'):
tag = tag.parent
tag.insert_before(email)
tag.decompose()
# strip colors
if self.options['strip_colors']:
for tag in contents.find_all(style=re.compile(r'(?:color|background)\s*:')):