minamotorin

In my environment, the following patch works correctly. ```diff # This patch is in the public domain. diff --git a/yt_dlp/extractor/common.py b/yt_dlp/extractor/common.py --- a/yt_dlp/extractor/common.py +++ b/yt_dlp/extractor/common.py @@ -1005,8 +1005,9 @@ def...

Okay, how about this? ```diff # This patch is in the public domain (CC0). diff --git a/yt_dlp/extractor/common.py b/yt_dlp/extractor/common.py --- a/yt_dlp/extractor/common.py +++ b/yt_dlp/extractor/common.py @@ -957,7 +957,7 @@ def _download_webpage_handle(self, url_or_request, video_id,...

@bashonly Thanks for the suggestion. I created a PR. https://github.com/yt-dlp/yt-dlp/pull/9879

I think the current implementention does not work with old format. In such cases, it may stop at the following point, because old format sometimes (not always) does not have...

This fix may work. ```diff - 'commentsHeaderRenderer' if is_first_continuation else ('commentThreadRenderer', 'commentViewModel'))]] + 'commentsHeaderRenderer' if is_first_continuation else ('commentThreadRenderer', 'commentViewModel', 'commentRenderer'))]] ```

> would you kindly explain how/why adding it improves robustness of data extraction? Okay, the simple reason is > old format sometimes (not always) does not have both `commentThreadRenderer` and...

I was successful in get the old format response with URL given by @shoxie007 and confirmed [my fix](https://github.com/yt-dlp/yt-dlp/pull/9775#issuecomment-2076424457) works correct. I tested it by the following steps. 1. Apply [this...

@jakeogh Here is diff. ```diff diff --git a/yt_dlp/extractor/youtube.py b/yt_dlp/extractor/youtube.py --- a/yt_dlp/extractor/youtube.py +++ b/yt_dlp/extractor/youtube.py @@ -3307,22 +3307,23 @@ class YoutubeIE(YoutubeBaseInfoExtractor): 'value': ('intensityScoreNormalized', {float_or_none}), })) or None - def _extract_comment(self, view_model, entity,...

I found another problem. ```diff - 'like_count': self._search_regex(r'^([\d]+)', try_get(comment_entity_payload, lambda x: x['toolbar']['likeCountA11y'], str), 'like_count', fatal=False) or 0, + 'like_count': str_to_int(self._search_regex(r'^([\d]+)', try_get(comment_entity_payload, lambda x: x['toolbar']['likeCountA11y'], str), 'like_count', fatal=False)) or 0, ```

It seems precision of number different in languages. For example, when I change language of browser to Italiano, I can see full number of like counts. See also [long scale](https://en.wikipedia.org/wiki/Long_and_short_scales).