That keybinding is Ctrl-\ on Linux .. maybe the same on OS X. But I would jump the memory up first .. you might be seeing weird problems related to the outofmemory situation.

That must be it. I run 47 from the "app" the installer blatted down, and just ran the jar directly to try this one. I have several soups. Let's find out.

With the 128 setting I don't get out of memory errors. I still get the following, but it doesn't seem to make jEmplode unusable.

Of course, popping up "Options" and trying to scroll results in the options window moving down the screen while it's scrolling, which is bizarre, but presumably isn't related.


Unknown fid type for fid 272 titled ""
java.lang.NullPointerException
at com.inzyme.container.ContainerSortableContainer.getSize(ContainerSortableContainer.java:33)
at com.inzyme.sort.ContainerSortUtils.linearSearch(ContainerSortUtils.java:172)
at com.inzyme.container.ContainerUtils.getChildContainer(ContainerUtils.java:92)
at org.jempeg.nodestore.soup.AbstractTagSoupUpdater.getContainer(AbstractTagSoupUpdater.java:87)
at org.jempeg.nodestore.soup.AbstractTagSoupUpdater.getContainer(AbstractTagSoupUpdater.java:89)
at org.jempeg.nodestore.soup.AbstractTagSoupUpdater.initialize(AbstractTagSoupUpdater.java:127)
at org.jempeg.nodestore.soup.AbstractTagSoupUpdater.initialize(AbstractTagSoupUpdater.java:111)
at org.jempeg.nodestore.soup.SoupUtils$SoupInitializeRunnable.run(SoupUtils.java:324)
at java.lang.Thread.run(Thread.java:554)

On Linux-x86, I get totally silent output when running it from a terminal window.

Cheers

How many soups?

Dunno. About 5600 tracks, though.

What's in your fid 272 that it is complaining about? 272 = 0x110, so look at the 111 file and also try to view it with Hijack:

http://your.players.ip.addr/?FID=111&EXT=.htm


Cheers

Note: 110/111 are the "Unattached Items" soup list.

Cheers

In reply to:

---

Of course, popping up "Options" and trying to scroll results in the options window moving down the screen while it's scrolling, which is bizarre, but presumably isn't related.

---

I just fixed this a couple minutes ago for RMML ... It appears to be a bug in Apple's Java impl when I turn on the brushed metal look. The next build will have that turned off which fixes the problem.

The fast database rebuild works for me (didn't try flac).

tried flac, and it loads great. gonna see if it crashes v3a5 in my car. what i'd like to see is seeing the speed improvement on a database rebuild for pgrzelak.

Perhaps I did something wrong here as I could still not get a database to work with alpha 5. I downloaded the jemplode.jar file and pointed by shortcut to it instead of the jemplode.exe created with the installer. Well, I did the manual rebuild from the command line with having alpha 3 on the player. It completed with no real errors. Then upgraded to alpha 5 and it went into the databse rebuild loop as all have seen. No database found. Tried to then do the command line rebuild and it would do nothing... Got a java error and no database. Then went back to alpha 3 and tried a simple upload and was having huge problems such as described by others with out of memory errors. I have a Mark2A model with 80GB on 2 hard drives and I've got close to 20,000 songs on it. I've been unable to do any sort of uploading of songs using Jemplode 48 and had to go back to emplode to even sync it all after failure and lockup of Jemplode 48. Did I do something horrible wrong or is it the size of my database or why can I not seem to get alpha 5 on like others have?

Jemmi

You shouldn't need to do anything with alpha3 anymore ... I just installed alpha5 and ran the commandline rebuild. When you say you got a java error, can you include the full stack trace of that error?

I am at work now and cannot do anything with this until the earliest tomorrow night as far as reflashing to alpha 5 and trying again. I am wondering if my huge database is causing problems


Jemmi

If you have a huge database, memory problems definitely aren't helping the situation ... I need to profile jEmplode again now that things have changed so drastically behind the scenes. Maybe I'm doing something silly.

Hello Mike,

Does this "Commandline full database rebuild" also work on a Windows PC?

Thanks a lot in advance...

Yep .. it's all pure java. I actually develop on Windows.

strange....

I get the following message:

C:\>java -classpath jemplode.jar org.jempeg.empeg.protocol.RebuildDatabase 192.168.0.51
Exception in thread "main" java.lang.NoClassDefFoundError: org/jempeg/empeg/protocol/RebuildDatabase

Same message, when I change to the jEmplode directory..

bye

unless your jemplode.jar is located at c:\, you're gonna need to change the directory to its location.

Sounds like you don't have the latest jemplode.jar (http://www.inzyme.com/rio/empeg)

it might not be able to find jemplode.jar either -- you should fully qualify the path to jemplode.jar in the commandline to be on the safe side too

yeep...That was the problem...Thanks..

It haves to look like this:
C:\Program Files\jEmplode>java -classpath lib\jemplode.jar org.jempeg.empeg.protocol.RebuildDatabase 192.168.0.51

And Hijack must be installed....

Thanks a lot

There is a new jEmplode at http://www.inzyme.com/rio/empeg

ACK, DANGER! From the debug output it looks like you have the early version of my ftp upload code in there. That version will cause strange, bizarre, and hard to recover from things to happen on dual drive players if "Use hijack when possible" is enabled in the options (it doesn't properly reuse fids). I sent another patch at some point, but it must not have made it in to this build.

The attached patch should be applied on top of what is already in there. In the meantime please don't use this version on a dual drive player with hijack enabled!

-Mike

Forgot the attachment...

cool, thanks. now, how do i apply it?

cool, thanks. now, how do i apply it?

We need mschrag to do another release with that patch applied unless you feel like building jemplode from source. If he doesn't poke his head in soon I'll but up a temp version somewhere after I get dinner.

-Mike

So, I have "Use hijack" disabled, and I'm using the given command line to rebuild my player database, but all the playlists vanish. When I re-launch jEmplode, it forces a media check, and then reports "There isn't a file with FID 2 on this device."
Yes, I have a dual-drive Mk2a. =(

Ughh, I just looked at the source and that patch may have been applied, but then Mike rewrote some of my code and duplicated one of my earlier bugs. Mike, take a look at HijackUtils.uploadFid(). That function needs to check both drives for the fid number before uploading and either reuse or delete existing fids otherwise you can end up with two copies of the same fid on a dual drive player (one on each drive). At one point I had a findFid() function in there that was intended to do that, but it isn't in this version.

-Mike

So, I have "Use hijack" disabled, and I'm using the given command line to rebuild my player database, but all the playlists vanish. When I re-launch jEmplode, it forces a media check, and then reports "There isn't a file with FID 2 on this device."
Yes, I have a dual-drive Mk2a. =(

That is not going to be related to the bug I mentioned, but for what it is worth I had the same problem. Does the new on PC database rebuild work with version 2 of the player software?

-Mike

I'm not sure if the rebuild process works with 2.0. I don't have any MP3s on my player at the moment, but I could roll back to 2.0, encode some, and try it. Would that be valuable test data for you, or should I just wait for a fix?
Speaking of patches.. I tried to apply the patch you linked earlier, but my Windows copy of "patch" behaves very oddly.
I tried "patch -p3 < whatever_the_patch_was_called.txt", and it just created a subdirectory called "-p3"
I was too lazy to send the files over to a Linux box with a 'real' patch tool.

Ack .. Sorry about that.. I shuffled things around a bit to do the database rebuild stuff. I'll fix this tonight when I get home. I don't remember seeing the findFid function, but I'll restore from CVS ... Now I have to try and remember what problem I was trying to solve by /removing/ it ... I'm terribly lazy with cvs messages when I'm not at work.

Now you see why I didn't post a top level thread about 48 being out I knew this one would be sketchy ..