RB 3.8 Sansa clip+ doesn't terminate generating database

[lwizard]

I have a Sansa clip+ with a 32GB uSD.
All was right with 3.7 version, also database worked very well.

Now I have updated to 3.8.
When I click on database it ask me to initialize.
I find about 6000 items then it stuck there.
After a lot of minutes it power off.
When I turn on again the database is still to be initialized.
I have tried keeping it alive by pressing buttons, but database doesn't terminate to compile.

Once I have also played music while compiling but RB hanged and I had to force power off.

Any idea?
Anyway.. great great job! I am a very fan of RB! Sadly I haven't time to code but I support RB with donation!

[saratoga]

Under the debug menu there is an option to enable database logging.  This will generate a list of all the files parsed when you build the database.  Could you enable that, find the file it crashes on decoding, and upload it here?

[lwizard]

I'll do as soon as possible!
Post Merge: March 01, 2011, 11:57:13 AMIt does stop at a certain file in the internal memory, but if I delete it, it stops at another file and so on.
Analyzing the internal memory in windows, it was corrupted.
So I reformatted the Sansa from the original firmware.
I reinstalled rockbox 3.8 with no mp3 on the internal memory.
I insrted the 32GB uSD card and initialized the database: all went right!
So I had few al lot of mp3 on internal memory, updated database: allright.
I added some mp3 again in internal memory: again database hang on refreshing or reinitializing and the internal memory is corrupted in windows.

I always do safe removal.

I think something corrupt internal memory.. maybe the new version of rock box?
II'll try for the last time, than turn back to 3.7.1 that was stable.


NEW:turned back to 3.7.1, the problem disappear.
It seems a 3.8 new bug!

[lebellium]

I confirm this issue with Rockbox 3.8 on my 8GB Clip+

[vroetman]

I also had problems with 3.8. It would hang on database generation. Furthermore, something got corrupted and I could not even access the filesystem with USB using the OF. The Sansa OF also took an unusually long time to boot. Other than those three problems, things seemed to work fine (with no database). I finally had to reformat from the OF (and will revert to 3.7.1).

[Progweed]

I've updated the Flyspray task concerning this bug:
http://www.rockbox.org/tracker/task/11965

This is what I did:

I tried r29511 again, turned on metadata logging and found that this is the offending file:
http://www.mediafire.com/?76ei0xwou83duel

Other things I noticed:
* foobar's integrity checker does not show any warnings
* I can't enable dircache any longer, it's always turned off after a reset.

[aargl]

I've got a recent 2Gb clip+ and went on the same problem.
After spending hours, updating, reinstalling, moving music and so on, I reinstalled the OF and RB 3.8, and initialized the Database from the menu.
It was still buggy and I decided to let it that way, using the File menu instead of the Database menu and switched it off.
Surprisingly, now, a few hours later, I started it and everything works fine! Is it possible that it only needs a loooooooong time to update the database and did it while switched off? 
The only issue still present is an <Untagged> artist and inside it an <Untagged> album: when I play the track in it, there's a digital flushing sound and then the display goes incoherent.

It's interresting to notice that when I loaded it back with music, I first tried with a single album and everything was ok, so the bug probably deals with the quantity of space occupied.

Perhapps it has nothing to do with it but before updating to 3.8 yesterday, I cancelled the preloaded songs and folders from within the Clip (as I'm on Mac).
Post Merge: March 06, 2011, 07:24:52 PMI have to mention that the last build I installed was r29533, it probably sorted out something.
I've just installed r29534 and there's still the "Untagged" thing...

[lwizard]

I can confirm that till I put one album or two, it was right. When I filled internal memory with albums, the bug appear.
Database generation hang at a very hi-degree of completion.. when almost all files are done..
Anyway now I'm back to 3.7.1 and everything is fine.

Another bug is that if you set as default browsing directory a location on external memory and than you power on rockbox without external memory, pressing the central button does nothing and you can't access the browser. You have to put back in the ext memory

[aargl]

It might help someone, maybe in a future bug  : I used a utility to visualize the content of the .rockbox directory and erased all the database_*.tcd files (not the database.ignore file).
It's more likely that the last build I installed cured the problem, but just in case...

[njd]

I got the same thing after I updated my Clip+ from 3.7.1 to 3.8 but now the FAT32 partition is unreadable in Windows.

I updated the Clip+ to 3.8 using RockboxUtility 1.2.8. On restarting, I accepted the prompt to initialize the database.
The database initialization got stuck after about 500 files, though still displaying the "disk write" (lightning) icon.

I guessed the update freeze was caused by lack of free space (about 100MB), and set about deleting some music folders, but everything was moving very slowly, so I canceled the deletions.

The disk write icon was displayed permanently, and I couldn't see how to cancel the db update, so I shut down the device, restarted, declined the update prompt, deleted about 1GB, and was able to complete the db initialization via the menus.

Now I can start up in RB3.8 or in the original firmware, and play everything just fine, but a USB connection to Windows XP (via Sansa f/w, MSC) either is not recognized (WinXP doesn't display the drive icons) or is recognized, but Windows reports the drive as "Not ready" or unreadable.

It seems the internal memory's FAT filesystem is corrupted in some way which doesn't affect normal operation, but it bothers Windows. Clip+ works fine, but I can't update the internal memory from Windows.

Anyone else seen this? And got out of it?

I've considered refomatting using the Sansa OF, but I'm not convinced it'll fix things; the Clip+ still kinda works at the moment, but if I wipe all my music/audiobooks/etc and still can't use USB, then I'll have a (nearly) useless player.

(I have a μSD card, but RB can't read it)

*** Update ***
Yes, I have tried another USB cable ( I have three )
Yes, I have tried other USB ports.

[aargl]

Hi!
I'm not sure if I can help, as I'm on Mac, but being an ooooold technician... 

I suppose you also tried to connect it in MTP mode? (I don't know how it works 'cos we Mac users can't do that)

Before formatting, I'd try it on another pc, if you can.
And if the problem persist, I'd try to check the volume structure of the clip before anything else with Windows utility (when I used pc, there was the CHKDSK DOS command, if I recall well).

If all works fine on another pc, then the problem may be in some file corruption on your pc, but there my skills end. 

[vroetman]

Anyone else seen this? And got out of it?

I've considered refomatting using the Sansa OF, but I'm not convinced it'll fix things; the Clip+ still kinda works at the moment, but if I wipe all my music/audiobooks/etc and still can't use USB, then I'll have a (nearly) useless player.

njd,

I had exactly the same problem (as I posted above). I was not using Windows so don't know what it would have looked like there. But a reformat with the Sansa Firmware brought everything back to normal, and then I installed 3.7.1. I would recommend you go ahead and format with the Sansa Firmware.

[njd]

Thanks for the update @lwizard. All I can say for you is that in my case. clearing our enough space (around 1GB) and choosing "Initialize database" from the menu (not the prompt) worked.

So there may be a minimum of free space required.


Post Merge: March 16, 2011, 03:15:07 AMUpdate: I've fixed the Clip+ now!

Running the Windows XP Computer Management control panel (compmgmt.msc) showed the Clip+ memory as a 3.66GB removable disk, and "Healthy".

Explorer's "Properties -> Tools -> Error Checking" GUI didn't do anything, but running "CHKDSK /F" on the drive worked.

Chkdsk warned me that the "disk" was not a valid FAT-formatted file system, and prompted whether to continue. I allowed it to carry on, and it reported lots of incorrectly linked files, but I now have the disk back.

[lwizard]

It's strange because I had a lot of free space on internal memory.. but it could also be that it need A LOT of free space.

Anyway.. since 3.7.1 works very well and we don't need album art on a sansa clip+ I'll wait for the next release to try again.

Keep on with the great work!!

[funman]

Please try a current build to see if r29625 fixes the problem