 |
Forums |
Admin  Start New Thread
By: Ikkei
 RE: Code cleanup [ reply ]
2004-09-03 03:15
|
> but a daemon would be much more efficient.
Oh, WEBrick can help to make the http based daemon.
I put sample code in my tar ball at (./sample/bt-tracker.rb)
|
By: Tom Copeland
RE: Code cleanup [ reply ]
2004-08-31 20:26
|
> ?0..?9
Good point, even better.
> Trakcer's bee encoder better be
> optimize, because of tracker may
> recieve a lot requests.
Ah, very true - since every HTTP response is bencoded. Hm. My current work on the tracker is running it as a CGI script... but a daemon would be much more efficient. Shouldn't be hard to switch back and forth, though, just a little front end work...
Yours,
Tom
|
By: Ikkei
RE: Code cleanup [ reply ]
2004-08-31 19:48
|
Hello,
Generally, normal recursive way may be slower
but it's more easy to read/write the code.
Now I plan to change it to
< '0'..'9'
> ?0..?9
Because of It is a char.
NOTES: Before local optimize. where should optimize
Download client's bee decoder/encoder does not need optimize. its not called so much time, unless download many files.
Trakcer's bee encoder better be optimize,
because of tracker may recieve a lot requests.
|
By: Tom Copeland
RE: Code cleanup [ reply ]
2004-08-31 17:45
|
Hi Ikkei -
Thanks, very interesting! I like your use of the case statement; I should do some benchmarks to see if it's faster to do a:
=~ /[0-9]
or a:
'0'..'9'
I suspect your way is faster...
Yours,
Tom
|
By: Tom Copeland
Code cleanup [ reply ]
2004-08-30 16:08
|
- the bencoder code isn't very nice; seems like it should use a Tokenizer/Context/Parser rather than the current clutter
- the decoder has some duplicate code
- is there a better way to use optparse?
Onwards and upwards,
Tom
|
|
|
