OK.. I now have displayserver successfully installed and running on my empeg. Now the big question.

How do you create an m3u file so that I can stream MP3's? I know how to create m3u files using winamp, but then what?

Thanks,
Rob

How do you create an m3u file so that I can stream MP3's? I know how to create m3u files using winamp, but then what?


In displayserver klick mp3-list; select a playlist which contains some songs; right click on the bottom link called "m3u" and choose "save as".
Save the file on your local pc and edit it with notepad to include your favorite songs

the only thing in the URL that's used by displayserver is the last part : "?file=<fid>"
You can use emplode to generate a CSV-file to find fids for specific songs...

Frank van Gestel

Is exporting mp3's and database pretty easy to do?

12gig Mk. II BLUE
Detroit, MI USA
www.PfeifferBeer.com

displayserver has an option which will save one big tar file of the first harddisk to your PC, but I heard there were problems when exceeding 4Gb. It's probably windows that has this limit.
Another option is using some kind of software which allows you to copy a complete webserver. Set the starting point to http://<ip-address of empeg>/mp3list.html and it will duplicate the playlist structure including mp3's to your local PC.
The filenames of the MP3's will be : "artist - album - title.mp3"

Frank van Gestel

So... if I understand this right... do a big ftp with something like GoZilla that lets you copy an entire website? That seems VERY easy! Is this with ethernet only?

12gig Mk. II BLUE
Detroit, MI USA
www.PfeifferBeer.com

Great idea. I never thought of that. What a great way to get a backup of your empeg. Just WebWhack displayserver. Brilliant.

Still doesn't retrieve emplode-entered title/artist name fixes to the MP3 files themselves, but it's a great start.

Question for Frank: When Displayserver shows the names of the MP3s, is it pulling the tag data to get that, or is it pulling the empeg database to get that? If the latter, then someone could use this technique to retrieve the MP3s, then use Tag Studio to batch-apply filenames to the tags. Done.

(Except you'd need to have displayserver add the year to the filenames, too, in order for it to be fully effective).

That seems VERY easy! Is this with ethernet only?

Yes, that would work with ethernet only because displayserver only works over ethernet.

___________
Tony Fabris

Great idea. I never thought of that. What a great way to get a backup of your empeg. Just WebWhack displayserver. Brilliant.

That's the power of an open protocol. Every programming-language can interface with a network-service with minimal effort.
I like it. If you like it too, have a look at plan9, an OS which was developed at Bell Labs which takes this kind of interfacing to the maximum.
to give you an idea of how it works, look at the /proc under linux, which was borrowed from plan9. Another great design is the ability to mount remote processors in your filesystem instead of traditional telnet/rsh functionality. All services have a file interface; to resolve a hostname using DNS you would have to cat a request to a file and read the result back from the same file. Anyway, read the papers there if you think it's interesting


Question for Frank: When Displayserver shows the names of the MP3s, is it pulling the tag data to get that, or is it pulling the empeg database to get
that? If the latter, then someone could use this technique to retrieve the MP3s, then use Tag Studio to batch-apply filenames to the tags. Done.

mp3list.html ( the page in displayserver which shows the main playlist ) is generated by reading the main fid, reading the titles from the corresponding *1 fids; when a *1 fid is clicked, it's *0 fid is opened for reading the next level of fids, and so on... (Is this confusing?)
Can Tag Studio be set up to retrieve tags from the file name? If so, I should also take care of a unique seperator between tag-data in the filenames. The year and maybe other tags(comment,genre) should be added for full functionality, I agree...


Yes, that would work with ethernet only because displayserver only works over ethernet.

I've just downloaded Borland C++ 5.5 from their website to try to compile the connection classes with USB support from emptool on win32. I have a working link allready, but it is only one way; from PC to empeg

Frank van Gestel

Can Tag Studio be set up to retrieve tags from the file name?

Yes. You can, as a batch operation, tell it to write tags based on interpreting a filename or folder format. If I recall correctly, it is very flexible, allowing you to specify formats and such.

If so, I should also take care of a unique seperator between tag-data in the filenames.

The only thing you'll need to do is make sure that the separator doesn't appear in the valid tag data. For instance, using a hyphen as the separator but having a hypen in the song title. Is that what you meant?

I don't know how to handle that, it seems that it'd be Tag Studio's problem in that case.

The year and maybe other tags(comment,genre) should be added for full functionality, I agree...

Would it look too ugly to have all that stuff on the web display page? If so, perhaps you could code it so that you've got two links: One with the simple title and one with the full filename as a little asterisk link next to it or something.

Unless that'd confuse a WebWhacker program and cause the file to be downloaded twice...

___________
Tony Fabris

I am having some problems editing the init file, I wish to run displayserver in noauth mode, however, I cannot for the life of me remember how to chmod the init file after I am done editing it off-empeg, so that it is run apon startup (Or do I have to chmod it at all? Once again, by lack of unix experience bite me in the arse)

So this is what I have done
On Empeg:
q
cd /sbin
rw
rw
sz init

On System:
Using Programmers File Editor, mentioned on this bbs
open INIT
edit ./displayserver to ./displayserver noauth
save
rz to emepg

On Empeg:
cp init init.old (I do know enough to make a backup, duh!)
cp INIT init
reboot
get kernel panic, cannot mount root fs"
reboot
boots into command prompt without player
cd /sbin
no init to be seen

Any help would be great thanks!

Smoker_Man
#080000449 MkII - 36Gb Blue

My script to do the work looks like this....

rwm
cd /drive0/var
rw
cd /bin
mv init init.old
cat /drive0/var/testinit > init
chmod 755 init
ro
rom
exit

This assumes that my edited version of the init script is sitting in drive0/var and is called testinit. Does that help?

___________
Tony Fabris

On Empeg:
cp init init.old (I do know enough to make a backup, duh!)
cp INIT init
reboot

should be :

# cp init init.old

# cp INIT init

# chmod +x init

# ro

reboot

Remember to allways set the disk to readonly, as otherwise the data can be gone ( as it was in this case )

Frank van Gestel

both tips help, and now mounting the drives vack to ro makes sense, however, I think I have poochied things a bit, so I am reapplying 1.02dev, and re-installing displayserver.
I will let you know how things proceed.

Thanks!

Smoker_Man
#080000449 MkII - 36Gb Blue

OK, well, all seems well, I do have an init script that works, and displayserver runs in noauth mode.
However, I am getting a rather ugly looking error on remount of partitions;

Tried to mount /proc but got error 16
Error mounting partitions (possibly already mounted)
Remounting first music partition read-only
Remounting second music partition read-only
No secondary hard disk

But I do have this:

Probing primary interface...
hda: IBM-DARA-218000, ATA DISK drive
hdb: IBM-DARA-218000, ATA DISK drive
ide0 at 0x000-0x007,0x038 on irq 6
hda: IBM-DARA-218000, 17301MB w/418kB Cache, CHS=35152/16/63
hdb: IBM-DARA-218000, 17301MB w/418kB Cache, CHS=35152/16/63

So I am thinking something is incorrect in the init, btw it is a dual drive mk2.
And lo and behold, drive1 isnt mounted.
but the new init has this stuff in it:

bin/mount -n -o nocheck,rw /proc
/bin/mount -n -o nocheck,ro /dev/hda4 /drive0
[ -e /proc/ide/hdc ] && /bin/mount -n -o nocheck,ro /dev/hdc4 /drive1

cd /usr/local/displayserver
./displayserver -noauth
exec /sbin/init.before_displayserver

any particular reason why drive1 isnt being mounted?

Thanks

Smoker_Man
#080000449 MkII - 36Gb Blue

[ -e /proc/ide/hdc ] && /bin/mount -n -o nocheck,ro /dev/hdc4 /drive1


should be :
[ -e /proc/ide/hdb ] && /bin/mount -n -o nocheck,ro /dev/hdb4 /drive1

I looked in the version on the website and it is indeed faulty. I've replaced it.


Frank van Gestel

I have since edited the /etc/fstab and /sbin/init files to what is below, and it does seem to work.

init
#!/bin/bash
/bin/mount -n -o nocheck,rw /proc
/bin/mount -n -o nocheck,ro /dev/hda4 /drive0
/bin/mount -n -o nocheck,ro /dev/hdc4 /drive1
[ -e /proc/ide/hdc ]

cd /usr/local/displayserver
./displayserver -noauth
exec /sbin/init.before_displayserver


fstab
#
# /etc/fstab
/dev/hda5 / ext2 defaults,ro 1 1
/dev/hda4 /drive0 ext2 defaults,ro 1 1
none /proc proc defaults
/dev/hdc4 /drive1 ext2 defaults,ro 1 1

All this seems to work, so I am testing it, by uploading a mp3.
Well, other than this showing up e2fsck 1.19, 13-Jul-2000 for EXT2 FS 0.5b, 95/08/09 in hyperterm, all seems to have gone well.

Can anyone see any problems with what I have put into init and fstab ??

Thanks


Smoker_Man
#080000449 MkII - 36Gb Blue

Thanks smoker, you've solved the puzzle. Apparently the empeg has /proc/ide/hdb, but the device to mount should be /dev/hdc4.
I understood from the empeg-guys there were some compatibility issues between MK1 and MK2.
Mk1 has /proc/ide/hda, /proc/ide/hdc, /dev/hda and /dev/hdc
MK2 has /proc/ide/hda, /proc/ide/hdb, dev/hda and /dev/hdc

Can anyone with a dual drive system verify this?

Frank van Gestel

What puzzle?

All I did was try to decipher the boot messages, and the partition info they contained, back to this later. Then I looked at fstab and init, and couldnt figure out why a process was mounted as a volume (/proc/ide/hdb), so I though Hugo must be smarter than I (And he is!), so I figured drive1 must be hdc4 (drive0 is hda4, so why not). Then, remembering the boot messages, I figured, why the heck not!
So I added to fstab, and modified the line in init to move the proc mounting lower, and edited the drive1 mount string to hdc4.
From what I gathered Frank's original init had this in it:
[ -e /proc/ide/hdc ] && /bin/mount -n -o nocheck,ro /dev/hdc4 /drive1

So I replaced it with this:
/bin/mount -n -o nocheck,ro /dev/hdc4 /drive1
[ -e /proc/ide/hdc ]

As per my previous posts.

I have now no problems mounting/remounting/unmounting volumes, and synching generates no extraordinary error messages (in hyperterm).

Dare I say I licked it? The problem, my empeg has already given me a good shock! ;-)

Is this set in stone with drive0 being hda4, drive1 being hdc4, and /proc/ide... being hdb4 ?? (And why?)

Enlighten me!


Smoker_Man
#080000449 MkII - 36Gb Blue

The MK1 has two seperate IDE controllers, so following the linux device-naming convention, the name of the first harddrive on the first controller would be /dev/hda
the name of the first harddisk on the secondary controller should be /dev/hdc
On the MKII there is only one controller, so the first drive on the controller is /dev/hda and the second /dev/hdb
For software compatiblity, they probably renamed the device files /dev/hdb* to /dev/hdc* on the MKII, this way software will access the second drive always as /dev/hdc, independently of which hardware is used (MK1 or MKII)

It was a puzzle for me as people said displayserver wasn't working on their empeg, and I hadn't thought about the possibility to check for /dev/hdb and use /dev/hdc

Frank van Gestel

Makes sense to me, seeing as Empeg/Rio wouldnt want completely different software running on the two generations of Emepg.

Does all this help you in constructing DS2.0 ??
I would also be more than willing to help test is out. (With remote access through my firewall to the gui for you to test)

Cheers


Smoker_Man
#080000449 MkII - 36Gb Blue

Yes, that was my "oh dear, quick hack" decision.

I'm sure Mike will look all superior and tell me we should have done that *properly* (ok, he's done this already...). It was getting very close to mk2 shipping when we worked this out...

Hugo

I'm sure Mike will look all superior and tell me we should have done that *properly* (ok, he's done this already...).

I'm sure he does that a couple of times a week, so that's no surprise.

___________
Tony Fabris