Opened 2 years ago
Closed 2 years ago
#1994 closed defect (fixed)
zzzot scrape returns malformed info_hash
| Reported by: | lazygravy | Owned by: | lazygravy |
|---|---|---|---|
| Priority: | minor | Milestone: | 0.9.31 |
| Component: | apps/plugins | Version: | 0.9.30 |
| Keywords: | zzzot | Cc: | |
| Parent Tickets: |
Description
The zzzot scrape returns malformed info_hashes. According to the BEP spec, they should be 20bytes long.
Here's an example URL of a bad hash: http://utnamrwujeccu7aj7aau7k2ld4evz5t47iogpjo2wgagbihi3jkq.b32.i2p/scrape
Here's a good one: http://5m3pd32zx43xk3uz6hvrdksj6tlg7abnjsc3j5kkd2yzctet4nmq.b32.i2p/scrape
Note how 5m3... has info_hashes of all 20
Subtickets
Attachments (1)
Change History (3)
comment:1 Changed 2 years ago by lazygravy
comment:2 Changed 2 years ago by zzz
- Milestone changed from undecided to 0.9.31
- Resolution set to fixed
- Status changed from new to closed
Bencoder was UTF-8 encoding the hashes. Above fix would have worked for zzzot alone, since it doesn't need UTF-8, but wouldn't for a fix on the snark side. To keep the licensing straight, fixed on the snark side by allowing and correctly encoding byte[]s as keys in maps.
In 584df860d1458f9e535ac31acf8a665664d74ad3 I2P 0.9.30-8 and zzzot 97982a658b8c4e41238ec4b0b2443f0fcea4f733 0.16.0
requires both to work
thanks for the report, initial patch, and testing
zzz did some debugging and found that the issue was in the BEncode'ing in snark. Snark changes the encofing to UTF-8. I've attached a hackish patch to resolve the issue. Some quick tests show it's working. 😊
NOTE: In the patch I commented out some code instead of copying over the necessary classes. Like I said, it's hackish