When (check) intervals are too small, app crashes.

[brentlintner]

You get stuff like this:

./bin/torrentfish --db foo.db -i .0001 -e .5 -u "http://www.torrentday.com/torrents/rss?l28;l12;l25;l11;l3;l21;l13;l1;l24;l14;l26;l7;l2;u=138040;tp=bd8d5be5da87a0c920adcfde91ce721c"
db info  loaded -> /Users/brent/src/torrentfish/foo.db
poll info  beginning poll
poll info    check interval: 0.0001h
poll info    email interval: 0.5h
poll info  initially checking feeds

...

poll info  added - Bad Country 2014 720p HDRip x264-KingStoner
poll info  added - Media Watch 2014 03 10 PDTV x264-RTA
poll info  added - The Bag Man 2014 BRRip x264 AAC-SSN
poll info  added - Sunday Night 2014 03 09 PDTV x264-RTA
db info  flushing records to disk -> /Users/brent/src/torrentfish/foo.db

TypeError: Cannot call method 'reduce' of null
    at matches_all (/Users/brent/src/torrentfish/lib/daemon/poll.js:11:18)
    at /Users/brent/src/torrentfish/lib/daemon/poll.js:21:11
    at Array.forEach (native)
    at Function.<anonymous> (/Users/brent/src/torrentfish/lib/daemon/poll.js:20:28)
    at process._tickCallback (node.js:415:13)

[brentlintner]

Why is the watchlist null?..

Is this a bug, or is this the need for check intervals to be queued (sequentially), vs with setInterval (which is probably not the best thing to do).

[brentlintner]

This was a bug with daemon/poll. Fixing.