22:02 -!- Irssi: Starting query in freenode with nenolod
22:02 <vollmer> you got a sec?
22:06 <vollmer> ping
22:06 <nenolod> k
22:07 <vollmer> look I'm sorry if I pissed you guys off or acted like an ass hat or whatever
22:07 <vollmer> so can we let bygons be bygons and agree not to be mean to each other :)
22:07 <nenolod> asserting that we're the only player that behaves this way was indeed improper
22:07 <vollmer> you don't have to like me, just don't hate me
22:08 <nenolod> most players behave the way we do
22:08 <vollmer> and I'm sorry for that
22:08 <vollmer> well no
22:08 <vollmer> the seek start, seek end, get file
22:08 <vollmer> is normal
22:08 <nenolod> actually, yes. GStreamer does the exact thing with GPlayBin
22:08 <vollmer> ok I don't know that client
22:08 <nenolod> XMMS2 does the same thing with xform-curl
22:08 <vollmer> xmms, bmpx, winamp, wmp, vlc, foobar don't
22:08 <vollmer> but anyway that doesn't matter
22:08 <nenolod> bmpx cheats by providing their own playbin and making braindead assumptions
22:09 <vollmer> because you've fixed the bug by reading the content-type right?
22:09 <vollmer> so it will do a 'best guess first'
22:09 <nenolod> what i am trying to explain is that it's not a bug
22:09 <vollmer> so that's was what I considered the bug, and my understanding is you've fixed it in 1.4.x?
22:09 <vollmer> ok undesirable behavior then sorry
22:10 <vollmer> either way it's fixed and I'm sorry if I offended you
22:10 <nenolod> undesirable to you, perhaps, but i find emulating Apache's HTTP handling inside a PHP
script to be throughly disgusting
22:10 <vollmer> well as one of the other devs brought up this was also an issue for nfs mounts or slow
drives
22:10 <vollmer> as it was extra reads
22:11 <vollmer> anyway you have correct what I personally saw as undesirable behavior
22:11 <nenolod> what the deal is
22:11 <vollmer> so I'm happy
22:11 <nenolod> is that we do format detection when adding things to a playlist
22:11 <nenolod> so there will be two requests
22:11 <vollmer> yup and I tottaly expect that :)
22:11 <nenolod> one to find the format (which will be a connection regardless, simply because we have to
probe), and then one to play
22:12 <nenolod> the format detection is only done at the first one
22:12 <nenolod> but yes, 1.4 design uses mime first
22:12 <vollmer> so yea what you described is not how 1.3 acts and if that's how 1.4 acts can't wait for it
to be released
22:12 <vollmer> so I'm happy thanks for making that change in behavior
22:13 <vollmer> and I just want to make sure we're cool
22:13 <nenolod> sure
22:13 <vollmer> you're welcome to hang around #ampache :)
22:14 <nenolod> however, your script should be able to handle all circumstances
22:15 <nenolod> and multiple seeks/requests is one of those circumstances that you will need to be able to
properly handle
22:15 <vollmer> ahh yes unless there is a bug I don't know about it does
22:15 <vollmer> and it did still work it just produced some undesirable behavior
22:16 <vollmer> and looking at the traffic I thought that 16+ HTTP requests for a single stream was a bug
(my mistake)
22:16 <nenolod> well those are just seeks
22:16 <nenolod> and it's because it looks for like 5000 different magic bytes
22:17 <vollmer> yea it's the type detection I understand :)
22:17 <vollmer> I had this dicusssion the first time I joined audacious
22:17 <vollmer> and my understanding was that currently it tries each plugin till it finds one that works
rather then trying the content-type plugin first which is what causes the large number of
seeks as it fails on the plugins
22:18 <vollmer> but that doesn't matter anymore because you've fixed it
22:18 <vollmer> so anyway I'm going to go back to coding, lots of work to do and all that
22:18 <vollmer> thanks for your time