aboutsummaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
-rw-r--r--yt_dlp/utils.py7
1 files changed, 3 insertions, 4 deletions
diff --git a/yt_dlp/utils.py b/yt_dlp/utils.py
index 13ad5fd48..fd3912d18 100644
--- a/yt_dlp/utils.py
+++ b/yt_dlp/utils.py
@@ -639,10 +639,9 @@ def clean_html(html):
if html is None: # Convenience for sanitizing descriptions etc.
return html
- # Newline vs <br />
- html = html.replace('\n', ' ')
- html = re.sub(r'(?u)\s*<\s*br\s*/?\s*>\s*', '\n', html)
- html = re.sub(r'(?u)<\s*/\s*p\s*>\s*<\s*p[^>]*>', '\n', html)
+ html = re.sub(r'\s+', ' ', html)
+ html = re.sub(r'(?u)\s?<\s?br\s?/?\s?>\s?', '\n', html)
+ html = re.sub(r'(?u)<\s?/\s?p\s?>\s?<\s?p[^>]*>', '\n', html)
# Strip html tags
html = re.sub('<.*?>', '', html)
# Replace html entities