commit fd35d8cdfdc77ca6ec6d87677fe0d00df0cbb22a
parent 4eec112740910621a7fd9c50158fb2388649d8b7
Author: Jakub Wilk <jwilk@jwilk.net>
Date:   Fri, 10 May 2019 20:42:32 +0200

[utils] Transliterate "þ" as "th" (#20897)

Despite visual similarity "þ" is unrelated to "p".
It is normally transliterated as "th":

    $ echo þ-Þ | iconv -t ASCII//TRANSLIT
    th-TH
Diffstat:
Mtest/test_utils.py | 2+-
Myoutube_dl/utils.py | 4++--
2 files changed, 3 insertions(+), 3 deletions(-)

diff --git a/test/test_utils.py b/test/test_utils.py @@ -183,7 +183,7 @@ class TestUtil(unittest.TestCase): self.assertEqual(sanitize_filename( 'ÂÃÄÀÁÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖŐØŒÙÚÛÜŰÝÞßàáâãäåæçèéêëìíîïðñòóôõöőøœùúûüűýþÿ', restricted=True), - 'AAAAAAAECEEEEIIIIDNOOOOOOOOEUUUUUYPssaaaaaaaeceeeeiiiionooooooooeuuuuuypy') + 'AAAAAAAECEEEEIIIIDNOOOOOOOOEUUUUUYTHssaaaaaaaeceeeeiiiionooooooooeuuuuuythy') def test_sanitize_ids(self): self.assertEqual(sanitize_filename('_n_cd26wFpw', is_id=True), '_n_cd26wFpw') diff --git a/youtube_dl/utils.py b/youtube_dl/utils.py @@ -125,8 +125,8 @@ KNOWN_EXTENSIONS = ( # needed for sanitizing filenames in restricted mode ACCENT_CHARS = dict(zip('ÂÃÄÀÁÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖŐØŒÙÚÛÜŰÝÞßàáâãäåæçèéêëìíîïðñòóôõöőøœùúûüűýþÿ', - itertools.chain('AAAAAA', ['AE'], 'CEEEEIIIIDNOOOOOOO', ['OE'], 'UUUUUYP', ['ss'], - 'aaaaaa', ['ae'], 'ceeeeiiiionooooooo', ['oe'], 'uuuuuypy'))) + itertools.chain('AAAAAA', ['AE'], 'CEEEEIIIIDNOOOOOOO', ['OE'], 'UUUUUY', ['TH', 'ss'], + 'aaaaaa', ['ae'], 'ceeeeiiiionooooooo', ['oe'], 'uuuuuy', ['th'], 'y'))) DATE_FORMATS = ( '%d %B %Y',