From 6fc85f617a5850307fd5b258477070e6ee177796 Mon Sep 17 00:00:00 2001 From: pukkandan Date: Mon, 23 Dec 2024 15:57:25 +0530 Subject: [PATCH] Don't sanitize filename on Unix when `--no-windows-filenames` (#9591) Closes #4547, Closes #8464 Authored by: pukkandan --- README.md | 3 +-- test/test_YoutubeDL.py | 7 +++++++ yt_dlp/YoutubeDL.py | 26 ++++++++++++++++++-------- yt_dlp/options.py | 4 ++-- 4 files changed, 28 insertions(+), 12 deletions(-) diff --git a/README.md b/README.md index 1db4ed2a5d..f8c99ace40 100644 --- a/README.md +++ b/README.md @@ -613,8 +613,7 @@ If you fork the project on GitHub, you can run your fork's [build workflow](.git --no-restrict-filenames Allow Unicode characters, "&" and spaces in filenames (default) --windows-filenames Force filenames to be Windows-compatible - --no-windows-filenames Make filenames Windows-compatible only if - using Windows (default) + --no-windows-filenames Sanitize filenames only minimally --trim-filenames LENGTH Limit the filename length (excluding extension) to the specified number of characters diff --git a/test/test_YoutubeDL.py b/test/test_YoutubeDL.py index 966d27a498..6b022a7eaa 100644 --- a/test/test_YoutubeDL.py +++ b/test/test_YoutubeDL.py @@ -761,6 +761,13 @@ class TestYoutubeDL(unittest.TestCase): test('%(width)06d.%%(ext)s', 'NA.%(ext)s') test('%%(width)06d.%(ext)s', '%(width)06d.mp4') + # Sanitization options + test('%(title3)s', (None, 'foo⧸bar⧹test')) + test('%(title5)s', (None, 'aei_A'), restrictfilenames=True) + test('%(title3)s', (None, 'foo_bar_test'), windowsfilenames=False, restrictfilenames=True) + if sys.platform != 'win32': + test('%(title3)s', (None, 'foo⧸bar\\test'), windowsfilenames=False) + # ID sanitization test('%(id)s', '_abcd', info={'id': '_abcd'}) test('%(some_id)s', '_abcd', info={'some_id': '_abcd'}) diff --git a/yt_dlp/YoutubeDL.py b/yt_dlp/YoutubeDL.py index 65b72e026c..764baf3a00 100644 --- a/yt_dlp/YoutubeDL.py +++ b/yt_dlp/YoutubeDL.py @@ -266,7 +266,9 @@ class YoutubeDL: outtmpl_na_placeholder: Placeholder for unavailable meta fields. restrictfilenames: Do not allow "&" and spaces in file names trim_file_name: Limit length of filename (extension excluded) - windowsfilenames: Force the filenames to be windows compatible + windowsfilenames: True: Force filenames to be Windows compatible + False: Sanitize filenames only minimally + This option has no effect when running on Windows ignoreerrors: Do not stop on download/postprocessing errors. Can be 'only_download' to ignore only download errors. Default is 'only_download' for CLI, but False for API @@ -1192,8 +1194,7 @@ class YoutubeDL: def prepare_outtmpl(self, outtmpl, info_dict, sanitize=False): """ Make the outtmpl and info_dict suitable for substitution: ydl.escape_outtmpl(outtmpl) % info_dict - @param sanitize Whether to sanitize the output as a filename. - For backward compatibility, a function can also be passed + @param sanitize Whether to sanitize the output as a filename """ info_dict.setdefault('epoch', int(time.time())) # keep epoch consistent once set @@ -1309,14 +1310,23 @@ class YoutubeDL: na = self.params.get('outtmpl_na_placeholder', 'NA') - def filename_sanitizer(key, value, restricted=self.params.get('restrictfilenames')): + def filename_sanitizer(key, value, restricted): return sanitize_filename(str(value), restricted=restricted, is_id=( bool(re.search(r'(^|[_.])id(\.|$)', key)) if 'filename-sanitization' in self.params['compat_opts'] else NO_DEFAULT)) - sanitizer = sanitize if callable(sanitize) else filename_sanitizer - sanitize = bool(sanitize) + if callable(sanitize): + self.deprecation_warning('Passing a callable "sanitize" to YoutubeDL.prepare_outtmpl is deprecated') + elif not sanitize: + pass + elif (sys.platform != 'win32' and not self.params.get('restrictfilenames') + and self.params.get('windowsfilenames') is False): + def sanitize(key, value): + return value.replace('/', '\u29F8').replace('\0', '') + else: + def sanitize(key, value): + return filename_sanitizer(key, value, restricted=self.params.get('restrictfilenames')) def _dumpjson_default(obj): if isinstance(obj, (set, LazyList)): @@ -1399,13 +1409,13 @@ class YoutubeDL: if sanitize: # If value is an object, sanitize might convert it to a string - # So we convert it to repr first + # So we manually convert it before sanitizing if fmt[-1] == 'r': value, fmt = repr(value), str_fmt elif fmt[-1] == 'a': value, fmt = ascii(value), str_fmt if fmt[-1] in 'csra': - value = sanitizer(last_field, value) + value = sanitize(last_field, value) key = '{}\0{}'.format(key.replace('%', '%\0'), outer_mobj.group('format')) TMPL_DICT[key] = value diff --git a/yt_dlp/options.py b/yt_dlp/options.py index 930d9d4bef..06b65e0eac 100644 --- a/yt_dlp/options.py +++ b/yt_dlp/options.py @@ -1370,12 +1370,12 @@ def create_parser(): help='Allow Unicode characters, "&" and spaces in filenames (default)') filesystem.add_option( '--windows-filenames', - action='store_true', dest='windowsfilenames', default=False, + action='store_true', dest='windowsfilenames', default=None, help='Force filenames to be Windows-compatible') filesystem.add_option( '--no-windows-filenames', action='store_false', dest='windowsfilenames', - help='Make filenames Windows-compatible only if using Windows (default)') + help='Sanitize filenames only minimally') filesystem.add_option( '--trim-filenames', '--trim-file-names', metavar='LENGTH', dest='trim_file_name', default=0, type=int,