testers/m3utester2: properly match Mist renditions

[iameli]

This changes the m3u8 parsing logic (with -mist=true set) to properly match Mist's new HLS output.

So given a playlist like this:

#EXTM3U
#EXT-X-INDEPENDENT-SEGMENTS
#EXT-X-STREAM-INF:CODECS="avc1.64002a,mp4a.40.2",RESOLUTION=1920x1080,FRAME-RATE=60,BANDWIDTH=10600418,AVERAGE-BANDWIDTH=8969585
0_1/index.m3u8?mTrack=0&iMsn=2&sessId=2801500359
#EXT-X-STREAM-INF:CODECS="avc1.4d4020,mp4a.40.2",RESOLUTION=1280x720,FRAME-RATE=60,BANDWIDTH=2768303,AVERAGE-BANDWIDTH=2342410
2_1/index.m3u8?mTrack=0&iMsn=2&sessId=2801500359
#EXT-X-STREAM-INF:CODECS="avc1.4d401f,mp4a.40.2",RESOLUTION=854x480,FRAME-RATE=60,BANDWIDTH=1584950,AVERAGE-BANDWIDTH=1341111
3_1/index.m3u8?mTrack=0&iMsn=2&sessId=2801500359
#EXT-X-STREAM-INF:CODECS="avc1.4d401f,mp4a.40.2",RESOLUTION=640x360,FRAME-RATE=60,BANDWIDTH=900141,AVERAGE-BANDWIDTH=761658
4_1/index.m3u8?mTrack=0&iMsn=2&sessId=2801500359
#EXT-X-STREAM-INF:CODECS="avc1.4d401e,mp4a.40.2",RESOLUTION=426x240,FRAME-RATE=60,BANDWIDTH=429998,AVERAGE-BANDWIDTH=363845
5_1/index.m3u8?mTrack=0&iMsn=2&sessId=2801500359

loadtester/recordtester will only look at the 0_1/index.m3u8 piece for deciding whether a new rendition needs to be pulled. But it will still instigate the (initial) pull using the full 0_1/index.m3u8?mTrack=0&iMsn=2&sessId=2801500359 bit, including the session id in the URL and simulating an actual Mist viewer. This fixes the bug where each iMsn and sessId was detected as a new rendition, causing more and more and more to get pulled, aka https://github.com/livepeer/livepeer-infra/issues/943