white/blacklist implemented (#44) + docs/tests updated

[sidd]

I've added a simple filter option to the server creation parameters as per issue #44.

If filter() returns false, a warning is emitted and the infoHash is not added to self._torrents. The test checks this by ensuring that the infoHash does not appear in a scrape request.

[feross]

hash is a string and parsedBitlove is an object, so this will never be true

[feross]

Good PR, but the tests need a bit of work. Don't worry about it though - I can fix them up :)

[feross]

Thanks, this is released (with some other breaking changes) as 3.0.0.

[feross]

@S1DD Btw, if you're interested the other breaking changes in 3.0.0 are in this commit: 82e6792 (The fixed filter test is in there)

Thanks again for sending this in!

[sidd]

Yep, took a look at it and things are making a lot more sense now. On a tangential note, I'm trying to figure out the best way to solve the storage adapter issue that was briefly discussed in #44.

[feross]

I believe Astro already solved issue #44 with this PR:
#45 Unless there's some
other functionality that's needed.

On Fri Jan 30 2015 at 4:00:07 PM Sidd Sridharan notifications@github.com
wrote:

Yep, took a look at it and things are making a lot more sense now. On a
tangential note, I'm trying to figure out the best way to solve the storage
adapter issue that was briefly discussed in #44
#44.

—
Reply to this email directly or view it on GitHub
#57 (comment)
.

[sidd]

Ok, I think I'm just misunderstanding some things then. I'll spend some more time tinkering, thanks!

[feross]

Cool, feel free to ask questions if anything's confusing. I think @astro intended the getSwarm method to be overridable by the user, and that's where you'd interface with a custom storage engine.

[sidd]

Alright, I'm unable to see how to override getSwarm and only getSwarm to check if the infohash exists in the database, since in _onAnnounce, we're not using a callback in order to wait until the DB query finishes. I feel like I'm just missing something completely!

[feross]

Pasting from IRC for posterity:

sidd: sup
00:41 <•feross> so, the way to override getSwarm is literally just overwrite it on the tracker server instance
00:41 <•feross> like this:
00:41 <•feross> var server = new TrackerServer({ udp: true, http: true })
00:41 <•feross> server.getSwarm = MySwarm
00:42 <•feross> and then you define a MySwarm class
00:42 <•feross> It needs to have the same API as https://github.com/feross/bittorrent-tracker/blob/master/lib/swarm.js
00:42 <•feross> which is to say, it needs:
00:42 <•feross> - a constructor
00:42 ahhhh
00:42 <•feross> - an announce method
00:42 <•feross> - a scrape method
00:42 thank you
00:43 <•feross> you don't need to implement the methods that start with _, since they're private
00:43 <•feross> but basically, you handle when swarm.announce gets called by storing into a database or whatever backend you want
00:44 <•feross> the implementation in lib/swarm.js just stores everything in a javascript object in memory