Torrent paths with invisible left to right mark fail to match

[tomauty]

Caesura expects directory:

 The Byrds- Sweetheart Of The Rodeo (1968) Jpn Sony Records Int'l ‎– MHCP 71 {2003}  [FLAC]

the directory in my content folder:

/The Byrds- Sweetheart Of The Rodeo (1968) Jpn Sony Records Int'l – MHCP 71 {2003}  [FLAC]

It seems even GitHub doesn't notice these characters:

Never seen this before and wondering if it's more common than I think. Ever encountered this? What's the recommended way to go about fixing? It feels kinda weird to rename the directory to add this random unicode character:

https://unicode-explorer.com/c/200E

Thanks for this amazing software btw.

[RogueOneEcho]

I can see the issue but I'm not entirely sure how to solve it because caesura is technically correct.

The file path is retrieved from the tracker API and that includes the mark. It's actually encoded as html entities so it's much clearer:

{
  "filePath": "The Byrds- Sweetheart Of The Rodeo (1968) Jpn Sony Records Int'l ‎– MHCP 71 {2003}  [FLAC]"
}

caesura decodes it here:

caesura/src/utils/source/source_provider.rs

Line 58 in 6995a76

let path = decode_html_entities(&torrent.file_path).to_string();

The issue arises because many torrent clients filter out certain invisible marks (see). They do this because they can be used maliciously to disguise file extensions (see).

I'm going to convert this to an idea discussion because while it's an issue I suspect it's a fairly rare occurrence and I don't have the bandwidth to address it at present but I would welcome anyone else contributing a solution

[tomauty]

Thanks for the attention! I thought for sure this would be a one-off, but surprisingly most of my missing_directory issues are due to this. Good to know the details, never heard of this exploit before. Gonna ponder a bit.

[RogueOneEcho]

Roughly how many are you seeing? I've only got 8 in my queue that have failed due to U+200E so it seems fairly rare

If you could give me a few ID numbers of torrents affected that would also help test.

This is the command I used to count:

cd /path/to/caesura/cache/queue
cat *.yml | yq -y 'map(select(.verify.issues and (.verify.issues | any(.type == "missing_directory" and (.path | test("\u200E")))))) | length'

And to list:

cd /path/to/caesura/cache/queue
cat *.yml | yq -y 'map(select(.verify.issues and (.verify.issues | any(.type == "missing_directory" and (.path | test("\u200E"))))))'

[tomauty]

21 for me. Here's my list:

- name: The Byrds- Sweetheart Of The Rodeo (1968) Jpn Sony Records Int'l ‎– MHCP 71
    {2003}  [FLAC]
  path: /torrents/The Byrds - Sweetheart of the Rodeo - 2003 (CD - FLAC - Lossless)-988284.torrent
  hash: 063a251e085e6d44354ddcd7f0406b6249eca362
  indexer: red
  id: 988284
  verify:
    verified: false
    issues:
      - type: missing_directory
        path: The Byrds- Sweetheart Of The Rodeo (1968) Jpn Sony Records Int'l ‎–
          MHCP 71 {2003}  [FLAC]
    completed: '2025-10-21T03:07:17Z'
- name: Adrianne Lenker ‎- Songs And Instrumentals
  path: /torrents/Adrianne Lenker - songs - 2020 (CD - FLAC - Lossless)-3020953.torrent
  hash: 6c00d3812795356d16ff982750dd8d7102789763
  indexer: red
  id: 3020953
  verify:
    verified: false
    issues:
      - type: missing_directory
        path: Adrianne Lenker ‎- Songs And Instrumentals
    completed: '2025-10-21T01:26:56Z'
- name: 'Brian Eno- More Music For Films (2005) Virgin ‎– VJCP-68747 [FLAC] '
  path: /torrents/Brian Eno- More Music For Films (2005) Virgin – VJCP-68747 [FLAC].torrent
  hash: 6d3d9eea10ba937c9b7d4797a3cbe644dc5b3b7b
  indexer: red
  id: 1326191
  verify:
    verified: false
    issues:
      - type: missing_directory
        path: 'Brian Eno- More Music For Films (2005) Virgin ‎– VJCP-68747 [FLAC] '
    completed: '2025-10-21T01:50:18Z'
- name: Elodie ‎– Musique En Scene (FLAC)
  path: /torrents/Elodie - Musique En Scene - 2015 (Cassette - FLAC - Lossless)-161371.torrent
  hash: 7e3ec3dcb2c7149602fc3634563014cd4c077fde
  indexer: pth
  id: 161371
  verify:
    verified: false
    issues:
      - type: missing_directory
        path: Elodie ‎– Musique En Scene (FLAC)
    completed: '2025-10-21T03:40:53Z'
- name: Abigor ‎- 2004 - Nachthymnen (From The Twilight Kingdom) + Orkblut - The Retaliation
    - IROND CD 04-883
  path: /torrents/Abigor - 2004 - Nachthymnen (From The Twilight Kingdom) + Orkblut
    - The Retaliation - IROND CD 04-883.torrent
  hash: 08eec6151d18d28cff33df2fc8df6039745136ba
  indexer: red
  id: 1322616
  verify:
    verified: false
    issues:
      - type: missing_directory
        path: Abigor ‎- 2004 - Nachthymnen (From The Twilight Kingdom) + Orkblut -
          The Retaliation - IROND CD 04-883
    completed: '2025-10-21T01:25:12Z'
- name: Francis Bebey ‎- Ballades Africaines [1978]
  path: /torrents/Francis Bebey - Ballades Africaines [1978].torrent
  hash: 9ee2b4b2925b001ecdb63eea37baa3b6f68e5eba
  indexer: red
  id: 3738724
  verify:
    verified: false
    issues:
      - type: missing_directory
        path: Francis Bebey ‎- Ballades Africaines [1978]
    completed: '2025-10-22T14:14:19Z'
- name: Boards Of Canada ‎– Telephasic Workshop (1998)  (VINYL-24bit FLAC)
  path: /torrents/Boards Of Canada – Telephasic Workshop (1998)  (VINYL-24bit FLAC).torrent
  hash: 1442bb2c71b88b5618ebc6c0ab1e0a6dcd5a64f4
  indexer: red
  id: 2675429
  verify:
    verified: false
    issues:
      - type: missing_directory
        path: Boards Of Canada ‎– Telephasic Workshop (1998)  (VINYL-24bit FLAC)
    completed: '2025-10-21T01:47:57Z'
- name: Ambient Hawai'i - (1998) Aloha Therapy {FOA Records ‎FRCA-1024 CD} [FLAC]
  path: /torrents/Ambient Hawai'i - Aloha Therapy - 1998 (CD - FLAC - Lossless)-1752741.torrent
  hash: 4301a616287aa24f3322dad872bf054374ffc877
  indexer: red
  id: 1752741
  verify:
    verified: false
    issues:
      - type: missing_directory
        path: Ambient Hawai'i - (1998) Aloha Therapy {FOA Records ‎FRCA-1024 CD} [FLAC]
    completed: '2025-10-21T01:31:13Z'
- name: Tadao Sawai, Kazue Sawai, Hozan Yamamoto - (1973) Jazz Rock 琴 - 日本の民謡 {RCA
    ‎JRS-7234 16-48 Vinyl} [FLAC]
  path: /torrents/Various Artists - Jazz Rock Koto  Nihon no Minyo (Jazz Rock 琴  日本の民謡)
    - 1973 (Vinyl - FLAC - Lossless)-2475740.torrent
  hash: 455ce30606347c135919e291bf28eb5a1a416ead
  indexer: red
  id: 2475740
  verify:
    verified: false
    issues:
      - type: missing_directory
        path: Tadao Sawai, Kazue Sawai, Hozan Yamamoto - (1973) Jazz Rock 琴 - 日本の民謡
          {RCA ‎JRS-7234 16-48 Vinyl} [FLAC]
    completed: '2025-10-21T03:01:18Z'
- name: Francis Bebey - Dibiye (1997) {Pee Wee Music ‎- PW 017} [FLAC]
  path: /torrents/Francis Bebey - Dibiye (1997) {Pee Wee Music - PW 017} [FLAC].torrent
  hash: 4653ccfba1a898cbf0f21576a399b90d33ae99a8
  indexer: red
  id: 1743335
  verify:
    verified: false
    issues:
      - type: missing_directory
        path: Francis Bebey - Dibiye (1997) {Pee Wee Music ‎- PW 017} [FLAC]
    completed: '2025-10-22T14:13:19Z'
- name: 'William Ackerman - It Takes A Year (1977) [FLAC] (Windham Hill Records ‎–
    WD-1003) '
  path: /torrents/William Ackerman - It Takes A Year (1977) [FLAC] (Windham Hill Records
    – WD-1003).torrent
  hash: 567948e90a02e37c1841825eb50deb57736e3a3b
  indexer: red
  id: 2157720
  verify:
    verified: false
    issues:
      - type: missing_directory
        path: 'William Ackerman - It Takes A Year (1977) [FLAC] (Windham Hill Records
          ‎– WD-1003) '
    completed: '2025-10-21T03:16:54Z'
- name: Yoshiaki Ochi (越智義朗) - (1996) Natural Sonic 2 {Ochi Office ‎NSP-1001 CD} [FLAC]
  path: /torrents/Yoshiaki Ochi (越智義朗) - Natural Sonic 2 - 1996 (CD - FLAC - Lossless)-1649643.torrent
  hash: 893fbc06ff3e47b5a5360f4f8806e7407252eac1
  indexer: red
  id: 1649643
  verify:
    verified: false
    issues:
      - type: missing_directory
        path: Yoshiaki Ochi (越智義朗) - (1996) Natural Sonic 2 {Ochi Office ‎NSP-1001
          CD} [FLAC]
    completed: '2025-10-21T03:19:43Z'
- name: Harmonia & Roger Eno ‎– Classical Music For Those With No Memory (1994) flac
  path: /torrents/Harmonia & Roger Eno - Classical Music for Those With No Memory
    - 1994 (CD - FLAC - Lossless)-342674.torrent
  hash: 97c33da4ea909fa846bd0c489f88b638662b4ead
  indexer: pth
  id: 342674
  verify:
    verified: false
    issues:
      - type: missing_directory
        path: Harmonia & Roger Eno ‎– Classical Music For Those With No Memory (1994)
          flac
    completed: '2025-10-21T03:42:39Z'
- name: Millencolin ‎– The Melancholy Collection (1999) [FLAC]
  path: /torrents/Millencolin – The Melancholy Collection (1999) [FLAC].torrent
  hash: adb33a47447b8001308a98797aed6b374d52e764
  indexer: pth
  id: 653314
  verify:
    verified: false
    issues:
      - type: missing_directory
        path: Millencolin ‎– The Melancholy Collection (1999) [FLAC]
    completed: '2025-10-21T03:46:09Z'
- name: The Byrds- Mr. Tambourine Man (1965) Jpn Sony Records Int'l ‎– MHCP 66 (Jpn}
    {2003}  [FLAC]
  path: /torrents/The Byrds - Mr. Tambourine Man - 2003 (CD - FLAC - Lossless)-1028401.torrent
  hash: b9efdbd52513d74f3186cf92b82a49104bd2dbc4
  indexer: red
  id: 1028401
  verify:
    verified: false
    issues:
      - type: missing_directory
        path: The Byrds- Mr. Tambourine Man (1965) Jpn Sony Records Int'l ‎– MHCP
          66 (Jpn} {2003}  [FLAC]
    completed: '2025-10-21T03:07:10Z'
- name: Gastr del Sol ‎– Crookt, Crackt, Or Fly (1994) FLAC
  path: /torrents/Gastr del Sol - Crookt, Crackt, or Fly - 1994 (CD - FLAC - Lossless)-1788582.torrent
  hash: be6aa802fe92f4f6568136984c98f01a8a81a794
  indexer: red
  id: 1788582
  verify:
    verified: false
    issues:
      - type: missing_directory
        path: Gastr del Sol ‎– Crookt, Crackt, Or Fly (1994) FLAC
    completed: '2025-10-21T02:12:30Z'
- name: Francis Bebey ‎- Mbira Dance  (2000) {Gravity ‎- GRAVITY 20015} [FLAC]
  path: /torrents/Francis Bebey - Mbira Dance  (2000) {Gravity - GRAVITY 20015} [FLAC].torrent
  hash: c18101cdc4b464b128f44edb7fce9f497cc3ff7b
  indexer: red
  id: 1743322
  verify:
    verified: false
    issues:
      - type: missing_directory
        path: Francis Bebey ‎- Mbira Dance  (2000) {Gravity ‎- GRAVITY 20015} [FLAC]
    completed: '2025-10-22T14:14:27Z'
- name: Christoph Heemann & Andrew Chalk ‎– Mirror of the Sea (2000) [FLAC] {Robot
    Records, Vinyl Rip 24bit 44.1kHz}
  path: /torrents/Andrew Chalk & Christoph Heemann - Mirror of the Sea - 2000 (Vinyl
    - FLAC - 24bit Lossless)-132198.torrent
  hash: d510c3da5ec4824d168dd2b246995d3edee900d8
  indexer: pth
  id: 132198
  verify:
    verified: false
    issues:
      - type: missing_directory
        path: Christoph Heemann & Andrew Chalk ‎– Mirror of the Sea (2000) [FLAC]
          {Robot Records, Vinyl Rip 24bit 44.1kHz}
    completed: '2025-10-21T03:38:04Z'
- name: Hiroshi Yoshimura (吉村弘) ‎- (1988) 静けさの本 (Static) {Misawa Home C32SD-4 CD}
    [FLAC]
  path: /torrents/Hiroshi Yoshimura (吉村弘) performed by Satsuki Shibano (柴野さつき) - 静けさの本
    Static for Piano - 1988 (CD - FLAC - Lossless)-1464932.torrent
  hash: e14703d33dd15a3549d6523186f83ee4293a486f
  indexer: red
  id: 1464932
  verify:
    verified: false
    issues:
      - type: missing_directory
        path: Hiroshi Yoshimura (吉村弘) ‎- (1988) 静けさの本 (Static) {Misawa Home C32SD-4
          CD} [FLAC]
    completed: '2025-10-21T02:18:56Z'
- name: Yoshiaki Ochi (越智義朗) - (1996) Aqua {FOA Records ‎FRCA-1013 CD} [FLAC]
  path: /torrents/Yoshiaki Ochi (越智義朗) - Aqua - 1996 (CD - FLAC - Lossless)-1649650.torrent
  hash: e6b01fdacdfb7e823fd4537c353f63c4ad760bf4
  indexer: red
  id: 1649650
  verify:
    verified: false
    issues:
      - type: missing_directory
        path: Yoshiaki Ochi (越智義朗) - (1996) Aqua {FOA Records ‎FRCA-1013 CD} [FLAC]
    completed: '2025-10-21T03:19:37Z'
- name: The Byrds- Turn Turn Turn (1965) Sony Records Int'l ‎– MHCP 67 (Jpn){2003}  [FLAC]
  path: /torrents/The Byrds - Turn! Turn! Turn! - 2003 (CD - FLAC - Lossless)-1029656.torrent
  hash: f015de6ec75aec8b0521db6b781b53a5dc0747f2
  indexer: red
  id: 1029656
  verify:
    verified: false
    issues:
      - type: missing_directory
        path: The Byrds- Turn Turn Turn (1965) Sony Records Int'l ‎– MHCP 67 (Jpn){2003}  [FLAC]
    completed: '2025-10-21T03:07:20Z'

[vilohuhu]

Hi,
FYI I had similar issue a while back and came up with this solution (based on libtorrent's logic): #169

[RogueOneEcho]

Fixed in #169