Scrape doesn't work, but must

[Nafania]

Hello. I'm developer of tracker software for bittorrent and i found some strange bug in scrape.

Here we go. My tracker allow scrape and some clients, Azureus for example, working correctly with them, but no utorrent.

Here is request that utorrent send to tracker

GET /tracker/scrape.php?passkey=here_goes_my_passkey&info_hash=%0bKva%d4%3cQ%80%be%193%b0p%5e%d5%9a%e1X%96%ef HTTP/1.1
Host: ***
User-Agent: uTorrent/1610
Accept-Encoding: gzip

And tracker answer

HTTP/1.1 200 OK
Date: Thu, 22 Mar 2007 20:19:34 GMT
Server: Apache/1.3.33 (Win32) PHP/4.4.4
X-Powered-By: PHP/4.4.4
Content-Encoding: gzip
Transfer-Encoding: chunked
Content-Type: text/plain

57 
d5:filesd20:KvaФ<QЂѕ3°p^ХљбX–пd8:completei0e10:downloadedi0e10:incompletei0eeee

And utorrent don't working with this answer.

Firstly i think that here's need Content-Length header, because it's HTTP/1.1.

I added this header - not working.

Than i think - may be something with gzip? I disable gzip and header "Content-Encoding: gzip" - not working.

It's utorrent bug or my error? If my error - what i'm doing wrong?

Thanks for answers and sorry for my bad english.

[DreadWingKnight]

the "57" line in your response is probably breaking it.

[Nafania]

Thanks for answer, but you're not right, "57" - it's a gzip feature.

When i disable gzip answer from tracker like this

HTTP/1.1 200 OK
Date: Thu, 22 Mar 2007 20:19:34 GMT
Server: Apache/1.3.33 (Win32) PHP/4.4.4
X-Powered-By: PHP/4.4.4
Transfer-Encoding: chunked
Content-Type: text/plain

d5:filesd20:KvaФ<QЂѕ3°p^ХљбX–пd8:completei0e10:downloadedi0e10:incompletei0eeee

But scrape not working

[DreadWingKnight]

Have you compared it to the output scrape from an existing tracker such as CBTT?

[Nafania]

I compared answer with tracker where utorrent's scrape work.

Here it is

GET /scrape.php?passkey=here_goes_my_passkey&info_hash=%3d%fc%90%a9%10bl%2e%c5%1cj%b7%b9h%18%8d%c4%e8EN HTTP/1.1
Host: ***
User-Agent: uTorrent/1610
Accept-Encoding: gzip

HTTP/1.1 200 OK
Date: Thu, 22 Mar 2007 22:10:34 GMT
Server: Apache/2.2.3 (Unix) DAV/2 mod_ssl/2.2.3 OpenSSL/0.9.8d PHP/4.4.4 mod_apreq2-20051231/2.5.7 mod_perl/2.0.2 Perl/v5.8.7
X-Powered-By: PHP/4.4.4
Content-Encoding: gzip
Vary: Accept-Encoding
Content-Length: 100
Content-Type: text/plain

d5:filesd20:=ьђ©bl.Еj·№hЌДиENd8:completei38e10:downloadedi40e10:incompletei28eeee

As you see, answers are similiar, excepts some headers.

But, on my opinion, headers can't break utorrent's work, but i try to add headers at my tracker which are in this answer (Vary, Content-Length) - it doesn't work.

I can't understand what's wrong.

Does utorrent has some expert settings or logging where i can see reasons why utorrent don't allow answers of my scrape?

PS i don't know what is CBTT

[DreadWingKnight]

Are you getting "Scrape OK" messages in uT on the one you're claiming isn't working?

[Nafania]

No, i have no it message or message 'scraping' in status.

Do you think - if i got this messages, will i start this topic?