3.0a7 not always saving playlist (+ other issues)

I'm having trouble which seems to have started around the 3.0a5/a6/a7 stage. Basically it's not always (probably fails 90% of the time) remembering my current playlist when I shut it down in the car. It also happens on AC power. It just reverts to some old playlist.

Weird thing is that it does remember the last playlist I selected so when I used Down, Down.L I get the last playlist I selected. BTW: Down, Down.L is also broken in 3.0a7 - it usually just enqueues immediately rather than just taking you to the playlist and allowing you to select where you want to go. Down, Down.L does work a bit better with 32MB of RAM for some reason but it's by no means 100%

Before anyone asks, yes my accessories and ignition wiring is good and correctly wired.

I've also noticed bookmarks don't seem to work either since I've been trying to use them to save my current playlists.

I'd prefer not to go back to 2.0 since I have a number of .oggs now and I think it will also erase any music metadata on songs played under 2.0. I guess that's the first step to prove it's a 3.0 thing and not hardware though.

Thought I'd ask since I haven't seen anyone else with a similar problem and it's starting to get quite annoying.

Ideas?

Yeah, this has been mentioned a few times. I have a pet theory that if you tweak the order after selecting a new playlist then you don't get the problem. Care to help test that one?

Yeah, this has been mentioned a few times

Hmm, didn't notice it mentioned - perhaps I've just forgotten.

if you tweak the order after selecting a new playlist then you don't get the problem

Haven't noticed that specifically but that's because I don't use tweak order that much these days since it regularly hangs/crashes the player at the end of the current song playing when the tweak was done.

Don't use the oscilloscope visuals.

Don't use the oscilloscope visuals.

I don't - ever. I only use either TexInfo visual or the Track Info display and I'm fairly certain it happens on either screen

Track fade?

Track fade?

Yes for 3 seconds. Another reason I'd like to keep 3.0.

OK replying to my own post but anyway....

The problem surfaced between 3.0a3 and 3.0a5. Something fairly drastic changed between the two with regards to playlists. When going from either a3 to a5 OR a5 to a3, the playlist is completely lost and the empeg starts with [-/-] 0:00.

a3 reliably remembers the playlist over a power down.
a5 reliably doesn't remember the playlist over a power down.

NB: All these tests were done on AC power in a Mark Lord dock.

Which is odd, because I'm running 3.0a7, and it does reliably remember the playlist over a reboot.

The playlist being lost when upgrading (or downgrading) is not necessarily related: there were no guarantees that it would be saved.

Which is odd, because I'm running 3.0a7, and it does reliably remember the playlist over a reboot.

Hmmm.

It also doesn't seem to matter if it's all my tracks (approx 11000) or just a single album of 10 tracks.

The playlist being lost when upgrading (or downgrading) is not necessarily related: there were no guarantees that it would be saved.

I have no issue with that - it's just something I noticed when going up and down versions and I feel must be related somehow.

It's pointing to something on the dynamic data partition I think since there are other anomalies I have noticed:

1) It randomly marks tracks for some reason.
2) Play counts seem to be often incorrect (some tracks I'm sure I've listened to still show 0 plays)
3) Bookmarks are very unreliable (about as reliable as playlist saving)

Roger, do these 3 things all seem to be working OK for you?

Would there possibly be any benefit of wiping my dynamic data partition? I can always back it up and restore it easily enough I guess.

1) It randomly marks tracks for some reason.

I've seen this before, but I'm not sure whether I'm remembering it from an internal release that definitely did have this problem (a piece of code deliberately marked the tracks if they were a different duration than claimed in the *1 file). It's alleged that it was disabled for the alpha.

I'll try resetting all of the marked tracks, and I'll see if any of them get re-marked.

2) Play counts seem to be often incorrect (some tracks I'm sure I've listened to still show 0 plays)

Can't say I've been paying much attention. Play counts seem to be working reliably for me.

3) Bookmarks are very unreliable (about as reliable as playlist saving)

I don't use bookmarks, so I can't comment on this one.

I've got a playlist of 5585 tracks loaded at the moment, and it's been working fine -- I've taken the player in and out of the car several times.

I just went and backed up my dynamic data partition so I could erase it just to see what happens and noticed something strange.

/dev/hdc is the original drive which came in the empeg. It has about 16 megs devoted to /dev/hdc3.

/dev/hda is a 40GB drive I added myself. It has a about 24 megs devoted to /dev/hda3. I built this with the drive upgrade instructions, word for word since I wouldn't have known how to do it otherwise.

i.e. I'm getting something like this

empeg:/empeg/bin# fdisk /dev/hda



Command (m for help): p



Disk /dev/hda: 255 heads, 63 sectors, 4864 cylinders

Units = cylinders of 16065 * 512 bytes



   Device Boot    Start       End    Blocks   Id  System

/dev/hda1             1         5     40131    5  Extended

/dev/hda2             6        10     40162+  83  Linux

/dev/hda3            11        13     24097+  10  OPUS

/dev/hda4            14      4864  38965657+  83  Linux

/dev/hda5             1         3     24034+  83  Linux

/dev/hda6             4         5     16033+  82  Linux swap



Command (m for help): q



empeg:/empeg/bin# fdisk /dev/hdc



Command (m for help): p



Disk /dev/hdc: 16 heads, 63 sectors, 58140 cylinders

Units = cylinders of 1008 * 512 bytes



   Device Boot    Start       End    Blocks   Id  System

/dev/hdc1             1        66     33232+   5  Extended

/dev/hdc2            67       132     33264   83  Linux

/dev/hdc3           133       165     16632   10  OPUS

/dev/hdc4           166     58140  29219400   83  Linux

/dev/hdc5             1        33     16569   83  Linux

/dev/hdc6            34        66     16600+  82  Linux swap

I know that /dev/hda3 is the only dynamic partition in use. Could the fact it's bigger have anything to do with my problems by any chance?

I know I'm clutching at straws here, but this is really starting to annoy me.

Could the fact it's bigger have anything...

It shouldn't do. The reason it's bigger is that, according to this line:

Units = cylinders of 16065 * 512 bytes

You've got a little less than 8Mb available in a single cylinder. Since a cylinder is the smallest unit involved in partitioning, to get the 16Mb required, it's been rounded up to the next higher cylinder (thus taking a little less than 24Mb).

The other drive's geometry translation has resulted in smaller cylinders, but more of them:

Units = cylinders of 1008 * 512 bytes

Thus, the granularity is better (504Kb/cylinder), so that 16Mb minimum requires 33 cylinders, but only 16,632Kb, a little more than 16Mb, but not much.

The technical reasons for the disparity aside, the larger dynamic data partition shouldn't make any difference to the behaviour of the player. OTOH, my dynamic data partition is 16569Kb in size.

Oh, and you can use "fdisk -l" (that's a lower-case L) to list the partitions, like this:

# fdisk -l /dev/hda

I had this type of issue. It seemed to have something to do with Flac files for me. I changed everything back to oggs and it has not lost its place in 2 weeks.

I do have about 20 flac files on there I think, but they definitely aren't in the playlists I'm currently having trouble with - I haven't listened to them for quite a while. I might try removing them anyway to see if it makes a difference.

No flac here.

I assume you're saying you don't have any flac files and the problem is also happening on your machine? I would have expected that to be honest but was going to give a fling tonight - probably not worth bothering now.

Do you also see the other issues I'm seeing genixia?

Don't use bookmarks much, so I haven't seen that one, but yes, I've seen the others.

1) It randomly marks tracks for some reason.

Working with 3a7 and jemplode, I definitely recognize this issue.

Jelle

Random marking on mine too. I get 3 or 4 tracks marked every day (2hrs usage/day on average).

I agree. Defintely am seeing the randomly marked tracks. I'm also noticing (maybe some of you guys can verify?) the following:

1) When a song is paused and then resumed, it starts not where it left off but about 2-3 second later so you loose 2-3 seconds of the song (although curiously enough the time dosen't skip ahead by N seconds, just the song...)

2) Sometimes the track back button dosen't work, although if you start playing a song it suddenly starts working again - like magic!

3) Crossfade support is still shaky, at best. I've already mentioned my beef about not being able to fade more than 5 seconds (2.5 out, 2.5 in on the next song) in other threads. I won't say again that I would really like to have the ability even to do a full 10 second fade...

4) I'm going to confirm that my scope visual is fubared. Not that I think anyone is in doubt of that.

5) Maybe somewhat related to crossfading... Maybe not... But I've had issues where it will just stop for a second or two 5 seconds into a new song... In other threads I've hypothisized that maybe this was a caching issue - but have yet to hear what Roger or anyone else that acctually has their hands and brains in the code for this release has to say... Comments guys?

I would love a new release, btw... I guess I'll just hope that the more time = more stable builds

Ooooh! Maybe a Beta next time...?


Joey