blatyo / bencodr Goto Github PK

In ruby 1.9 length of utf-8 string is equal to the number of characters, not the number of bytes. Non-ASCII characters are encoded 2-6 bytes. For fix replace "length" to "bytesize" in file "bencodr/lib/bencodr/string.rb" line 14:

[string.length, ':', string].join

replace to:

[string.bytesize, ':', string].join

String#bytesize is works in ruby 1.8 and 1.9.

Hello,
I've been working with bencodr for some weeks parsing and re-encoding 15.000 torrent files for a specific project.

However, one torrent file - just one - gives me a
BEncode::DecodeError: BEncode::DecodeError
decode.rb:46:in `bdecode'

It is a correctly encoded torrent file because I can open it with other torrent applications, such as Transmission.

I could not fix the bencodr by myself but I think this file is a good case for make this lib 100%, after being considered stable for years.

The "special" torrent file is uploaded to http://extrapolo.com/download/7305.torrent

I appreciate any help.

Sometimes with torrents I get this error

too big number for repeat range: /.{133400}/m
~/.rvm/gems/ruby-1.9.3-p125/gems/bencodr-3.0.1/lib/bencodr/parser.rb:40:in `parse_string'

How would I be able to solve this?

Bencdr.decode Bencodr.bencode(3.1415) == 3

This conversion happens silently, which makes it a bit dangerous, because people will not know that it happens, and with some data this will create issues.
It would be great if not having float support could be added to the documentation. Also, maybe throwing an error instead of silently rounding is better behaviour in cases like this.