Commit Graph

673 Commits (a08f2b7e4567cdc50c0614ee0a4ffdff49b8b6e6)

Author SHA1 Message Date
Ori Avtalion 4ea59c6107
[utils] Fix crash in _report_ignoring_subs from c58b655 (#32762)
Align `utils.bug_reports_message()` with yt-dlp https://github.com/yt-dlp/yt-dlp/commit/5873d4ccdd, thanks fstirlitz

---------

Co-authored-by: dirkf <fieldhouse@gmx.net>
2024-04-05 15:25:29 +01:00
dirkf 19dc10b986 [utils] Apply compat_contextlib_suppress 2024-03-27 13:11:17 +00:00
dirkf acc383b9e3 [utils] Let int_or_none() accept a base, like int() 2024-03-08 13:03:42 +00:00
Hubert Hirtz f0812d7848
[utils] Handle user:pass in URLs (#28801)
* Handle user:pass in URLs

Fixes "nonnumeric port" errors when youtube-dl is given URLs with
usernames and passwords such as:

    http://username:password@example.com/myvideo.mp4

Refs:
- https://en.wikipedia.org/wiki/Basic_access_authentication
- https://tools.ietf.org/html/rfc1738#section-3.1
- https://docs.python.org/3.8/library/urllib.parse.html#urllib.parse.urlsplit

Fixes #18276 (point 4)
Fixes #20258
Fixes #26211 (see comment)

* Align code with yt-dlp

---------

Co-authored-by: dirkf <fieldhouse@gmx.net>
2024-03-04 01:27:55 +00:00
dirkf 427472351c [utils] Make restricted filenames ignore characters in Unicode categories Mark, Other
Resolves #32629
2023-11-29 22:08:01 +00:00
dirkf 66ab0814c4 [utils] Revert bbd3e7e, updating docstring, test instead 2023-09-03 23:15:19 +01:00
dirkf bbd3e7e999 [utils] Properly handle list values in update_url()
An actual list value in a query update could have been treated
as a list of values because of the key:list parse_qs format.
2023-09-03 01:18:22 +01:00
dirkf 2efc8de4d2 [utils] Advertise optional supported `Content-Encoding`s 2023-08-01 01:05:09 +01:00
dirkf e4178b5af3 [utils] Add and use `filter_dict()` from yt-dlp 2023-08-01 01:05:09 +01:00
dirkf 2d2a4bc832 [utils] Revise `isinstance()` tests (especially for str/unicode/bytes) to complete Linter fix 2023-08-01 01:05:09 +01:00
dirkf 7d965e6b65 [utils] Avoid comparing `type(var)`, etc, to pass new Linter rules 2023-08-01 01:05:09 +01:00
dirkf abef53466d [utils] Rework URL path munging for ., .. components
* move processing to YoutubeDLHandler
* also process `Location` header for redirect
* use tests from https://github.com/yt-dlp/yt-dlp/pull/7662
2023-07-29 14:27:26 +01:00
dirkf e7926ae9f4 [utils] Rework decoding of `Content-Encoding`s
* support nested encodings
* support optional `br` encoding, if brotli package is installed
* support optional 'compress' encoding, if ncompress package is installed
* response `Content-Encoding` has only unprocessed encodings, or removed
* response `Content-Length` is decoded length (usable for filesize metadata)
* use zlib for both deflate and gzip decompression
* some elements taken from yt-dlp: thx especially coletdjnz
2023-07-29 14:27:26 +01:00
dirkf 2b7dd3b2a2 [utils] Fix update_Request() with empty data (not None) 2023-07-25 13:19:43 +01:00
dirkf 1fa8b86f0b [utils] Remove stray undocumented Host header in redirect (fix 46fde7c) 2023-07-20 05:29:59 +01:00
dirkf a190b55964 [utils] Fix broken Py 3.11+ compat in `traverse_obj()`
* inspect.getargspec is missing despite doc claiming backward compat
* replace with emulation of `Signature.bind()`
2023-07-19 22:14:50 +01:00
dirkf cb9366eda5 [utils] Minor updates (merge_dicts, T)
A couple of mods to ease yt-dlp back-ports:
* add kwargs to merge_dicts:
  `unblank=True` (disallow empty string), `rev=False` (reverse the merge list)
* add `T(x)` shortcut for `{x}`, unsupported in Py2.6
2023-07-19 22:14:50 +01:00
dirkf d9d07a9581 [utils] Improve js_to_json, align with yt-dlp
* support variable substitution, from https://github.com/yt-dlp/yt-dlp/pull/#521 etc,
  thanks ChillingPepper, Grub4k, pukkandan
* improve escape handling, from https://github.com/yt-dlp/yt-dlp/pull/#521
  thanks Grub4k
* support template strings from https://github.com/yt-dlp/yt-dlp/pull/6623
  thanks Grub4k
* add limited `!` evaluation (eg, !!0 -> false, see tests)
2023-07-19 22:14:50 +01:00
dirkf 825a40744b [utils] Align traverse_obj() with yt-dlp
Thanks Grub4k for these:
* traverse `Iterable`s, from https://github.com/yt-dlp/yt-dlp/pull/6902, etc
* traverse `set` key for transformations/filters, `re.Match` group names, from
  776995bc10, etc
* traverse `re.Match`es, from https://github.com/yt-dlp/yt-dlp/pull/5174
* always return list when branching, from https://github.com/yt-dlp/yt-dlp/pull/5170
2023-07-19 22:14:50 +01:00
dirkf 1d8d5a93f7 [test] Fixes for old Pythons 2023-07-18 10:50:46 +01:00
bashonly 3801d36416 [utils] `YoutubeDLCookieJar`: Add `get_cookie_header` and `get_cookies_for_url` methods 2023-07-18 10:50:46 +01:00
dirkf b383be9887 [core] Remove `Cookie` header on redirect to prevent leaks
Adated from yt-dlp/yt-dlp-ghsa-v8mc-9377-rwjj/pull/1/commits/101caac
Thx coletdjnz
2023-07-18 10:50:46 +01:00
dirkf 46fde7caee [core] Update redirect handling from yt-dlp
* Thx coletdjnz: https://github.com/yt-dlp/yt-dlp/pull/7094
* add test that redirected `POST` loses its `Content-Type`
2023-07-18 10:50:46 +01:00
dirkf f47fdb9564 [utils] Add {expected_type} and Iterable support to traverse_obj() 2023-07-18 10:50:46 +01:00
dirkf f24bc9272e [Misc] Fixes for 2.6 compatibility 2023-07-05 22:58:54 +01:00
dirkf 11cc3f3ad0 [utils] Fix `compiled_regex_type` in 249f2b6 2023-05-11 20:53:07 +01:00
dirkf 64d6dd64c8 [YouTube] Support Releases tab 2023-04-23 22:58:35 +01:00
dirkf 25124bd640 [devscripts] Improve hack to convert command-line options to API options
* define equality for DateRange
* don't show default DateRange
2023-04-05 19:05:16 +01:00
dirkf f35b757c82 [utils] Ensure `allow_types` for `variadic()` is a tuple 2023-03-19 02:29:00 +00:00
pukkandan 1d3751c3fe Escape URLs in `sanitized_Request`, not `sanitize_url` d2558234cf5dd12d6896eed5427b7dcdb3ab7b5a added escaping of URLs while sanitizing. However, `sanitize_url` may not always receive an actual URL. Eg: When using `youtube-dl "search query" --default-search ytsearch`, `search query` gets escaped to `search%20query` before being prefixed with `ytsearch:` which is not the intended behavior. So the escaping is moved to `sanitized_Request` instead. 2023-02-20 20:27:25 +00:00
dirkf 90c9f789d9 [utils] Add parse_qs, update_url
[skip ci]
2023-02-13 03:54:51 +00:00
dirkf 58988c1421 [YouTube] Bypass age-gating for certain restricted videos
* Use TVHTML5_SIMPLY_EMBEDDED_PLAYER client

* Also add and fix tests

* Introduce and use new utility function `update_url()`
2023-02-13 03:54:51 +00:00
Andrei Lebedev 27ed77aabb
[utils] Backport traverse_obj (etc) from yt-dlp (#31156)
* Backport traverse_obj and closely related function from yt-dlp (code by pukkandan)
* Backport LazyList, variadic(), try_call (code by pukkandan)
* Recast using yt-dlp's newer traverse_obj() implementation and tests (code by grub4k)
* Add tests for Unicode case folding support matching Py3.5+ (requires f102e3d)
* Improve/add tests for variadic, try_call, join_nonempty

Co-authored-by: dirkf <fieldhouse@gmx.net>
2022-11-03 10:09:37 +00:00
dirkf c94a459a24
[utils] Sanitize look-alike Unicode glyphs in non-ID filename fields when --restrict-filenames
Implements https://github.com/ytdl-org/youtube-dl/issues/31216#issuecomment-1236102822, which has a test.
2022-10-11 12:18:12 +00:00
dirkf 556862bc91 [utils] Ensure RFC3986 encoding result is unicode 2022-08-21 00:45:06 +01:00
dirkf d231b56717
[jsinterp] Overhaul JSInterp to handle new YT players 4c3f79c5, 324f67b9 (#31170)
* back-port from yt-dlp 8f53dc44a0cc1c2d98c35740b9293462c080f5d0, thanks pukkandan
* also support void, improve <</>> precedence, improve expressions in comma-list
* add more tests
2022-08-14 18:45:45 +01:00
pukkandan 0700fde640 [utils, etc] Kill child processes when yt-dl is killed
* derived from PR #26592, closes #26592

Authored by: Unrud
2022-06-10 19:57:46 +01:00
pukkandan 1baa0f5f66 [utils] Escape URL while sanitizing
Closes #31008, #yt-dlp/263

While this fixes the issue in question, it does not try to address the root-cause of the problem
Refer: 915f911e365736227e134ad654601443dbfd7ccb, f5fa042c82300218a2d07b95dd6b9c0756745db3
2022-06-06 16:03:04 +01:00
dirkf 52c3751df7 [utils] Enable ALPN in HTTPS to satisfy broken servers
See https://github.com/yt-dlp/yt-dlp/issues/3878
2022-05-28 13:52:51 +01:00
Sergey M․ cfee2dfe83
[utils] PEP 8 2021-04-17 03:32:04 +07:00
Sergey M․ a00a7e0cad
[utils] Add support for support for experimental HTTP response status code 308 Permanent Redirect (refs #27877, refs #28768) 2021-04-17 03:22:13 +07:00
Remita Amine e88c9ef62a [utils] add a function to clean podcast URLs 2021-01-04 01:14:25 +01:00
Remita Amine 9dd674e1d2 [utils] accept only supported protocols in url_or_none 2020-12-30 09:22:30 +01:00
Josh Soref 71ddc222ad
Fix typos (#27084)
* spelling: authorization

Signed-off-by: Josh Soref <jsoref@users.noreply.github.com>

* spelling: brightcove

Signed-off-by: Josh Soref <jsoref@users.noreply.github.com>

* spelling: creation

Signed-off-by: Josh Soref <jsoref@users.noreply.github.com>

* spelling: exceeded

Signed-off-by: Josh Soref <jsoref@users.noreply.github.com>

* spelling: exception

Signed-off-by: Josh Soref <jsoref@users.noreply.github.com>

* spelling: extension

Signed-off-by: Josh Soref <jsoref@users.noreply.github.com>

* spelling: extracting

Signed-off-by: Josh Soref <jsoref@users.noreply.github.com>

* spelling: extraction

Signed-off-by: Josh Soref <jsoref@users.noreply.github.com>

* spelling: frontline

Signed-off-by: Josh Soref <jsoref@users.noreply.github.com>

* spelling: improve

Signed-off-by: Josh Soref <jsoref@users.noreply.github.com>

* spelling: length

Signed-off-by: Josh Soref <jsoref@users.noreply.github.com>

* spelling: listsubtitles

Signed-off-by: Josh Soref <jsoref@users.noreply.github.com>

* spelling: multimedia

Signed-off-by: Josh Soref <jsoref@users.noreply.github.com>

* spelling: obfuscated

Signed-off-by: Josh Soref <jsoref@users.noreply.github.com>

* spelling: partitioning

Signed-off-by: Josh Soref <jsoref@users.noreply.github.com>

* spelling: playlist

Signed-off-by: Josh Soref <jsoref@users.noreply.github.com>

* spelling: playlists

Signed-off-by: Josh Soref <jsoref@users.noreply.github.com>

* spelling: restriction

Signed-off-by: Josh Soref <jsoref@users.noreply.github.com>

* spelling: services

Signed-off-by: Josh Soref <jsoref@users.noreply.github.com>

* spelling: split

Signed-off-by: Josh Soref <jsoref@users.noreply.github.com>

* spelling: srmediathek

Signed-off-by: Josh Soref <jsoref@users.noreply.github.com>

* spelling: support

Signed-off-by: Josh Soref <jsoref@users.noreply.github.com>

* spelling: thumbnail

Signed-off-by: Josh Soref <jsoref@users.noreply.github.com>

* spelling: verification

Signed-off-by: Josh Soref <jsoref@users.noreply.github.com>

* spelling: whitespaces

Signed-off-by: Josh Soref <jsoref@users.noreply.github.com>
2020-11-21 22:00:05 +07:00
Sergey M․ fe07e788bf
[utils] Skip ! prefixed code in js_to_json 2020-11-17 01:30:43 +07:00
Kevin O'Connor 4eda10499e
[utils] Don't attempt to coerce JS strings to numbers in js_to_json (#26851)
The current logic in `js_to_json` tries to rewrite octal/hex numbers to
decimal. However, when the logic actually happens the `"` or `'` have
already been trimmed off. This causes what were originally strings, that
happen to look like octal/hex numbers, to get rewritten to decimal and
returned as a number rather than a string.

In practive something like:

```js
{
  "0x40": "foo",
  "040": "bar",
}
```

would get rewritten as:

```json
{
  64: "foo",
  32: "bar
}
```

This is problematic since this isn't valid JSON as you cannot have
non-string keys.
2020-10-18 00:10:41 +07:00
Sergey M․ 1d9bf655e6
[utils] Recognize wav mimetype (closes #26463) 2020-09-06 11:19:53 +07:00
Rob 9cd5f54e31
[utils] Fix file permissions in write_json_file (closes #12471) (#25122) 2020-05-20 03:21:52 +07:00
Sergey M․ c380cc28c4
[utils] Improve cookie files support
+ Add support for UTF-8 in cookie files
* Skip malformed cookie file entries instead of crashing (invalid entry len, invalid expires at)
2020-05-05 04:21:25 +07:00
Sergey M․ f1a8511f7b
[utils] Add reference to cookie file format 2020-03-10 04:59:02 +07:00