MIDI getting killed by Jack?

MIDI getting killed by Jack?

[Mark Knecht]

Hi,
   I posted this last night on LAU but received no responses. Either it
was of no interest or just no one had anything to say about it.

   Is anyone else experiencing this problem? I set up a bunch of
aconnect commands to route my keyboard controller (64:0) to a number of
synths on different MIDI interfaces, and also to route the output of Pro
Tools (on another computer - 80:0) to pretty much the same places:

aconnect 64:0 64:32
aconnect 64:0 80:1
aconnect 80:0 64:0
aconnect 80:0 64:32
aconnect 80:0 80:1

I also route Pro Tools output back to my keyboard (line 3) in case I
want to plug another hardware synth in there, but nothing is hooked up
to that right now.

At this point I set up my mixer and play. Everything works and is
perfectly fine for long periods of time.

I then start Jack and within a minute or two I lose MIDI connectivity. I
can run alsaplayer or other audio apps and I still get audio, but there
is no more MIDI connectivity. I look at the synths and they are not
receiving MIDI input. (No event lights.)

I then stop Jack and I immediately hear a bunch of MIDI notes get
played. These notes were stored from when I wasn't getting connectivity.
At this point I have MIDI back for live playing until the next time I
start Jack when it happens again.

I do this 4 or 5 times and at this point I cannot get MIDI back at all.
I have to completely power down as far as I can tell. I tried rebooting
and had no luck the one time I did that.

I don't know what to look at to figure out what could be causing this.
Hardware and modules shown below.

In thinking about this overnight it dawned on me that I am only seeing 
this problem on MIDI generated by my controller hooked to the HDSP
MIDI input. I do not see this problem on MIDI generated by Pro Tools
hooked to the MidiSport. I will likely rewire what is hooked to what
this evening and see if I can say that it is only the HDSP input, or
something else.

Thanks,
Mark


64:X - HDSP 9652
80:X - MidiSport 2x2

Wizard root # lspci
00:00.0 Host bridge: VIA Technologies, Inc. VT8366/A/7 [Apollo
KT266/A/333]
00:01.0 PCI bridge: VIA Technologies, Inc. VT8366/A/7 [Apollo
KT266/A/333 AGP]
00:09.0 Ethernet controller: Broadcom Corporation BCM4401 100Base-T (rev
01)
00:0a.0 FireWire (IEEE 1394): Texas Instruments TSB12LV23 IEEE-1394
Controller
00:0e.0 Multimedia audio controller: Xilinx Corporation RME Hammerfall
DSP (rev 65)
00:10.0 USB Controller: VIA Technologies, Inc. USB (rev 80)
00:10.1 USB Controller: VIA Technologies, Inc. USB (rev 80)
00:10.2 USB Controller: VIA Technologies, Inc. USB (rev 80)
00:10.3 USB Controller: VIA Technologies, Inc. USB 2.0 (rev 82)
00:11.0 ISA bridge: VIA Technologies, Inc. VT8235 ISA Bridge
00:11.1 IDE interface: VIA Technologies, Inc.
VT82C586A/B/VT82C686/A/B/VT8233/A/
C/VT8235 PIPC Bus Master IDE (rev 06)
00:11.5 Multimedia audio controller: VIA Technologies, Inc.
VT8233/A/8235 AC97 Audio Controller (rev 50)
01:00.0 VGA compatible controller: ATI Technologies Inc Radeon R250 If
[Radeon 9000] (rev 01)
01:00.1 Display controller: ATI Technologies Inc Radeon R250 [Radeon
9000] (Secondary) (rev 01)

Wizard root # lsmod
Module                  Size  Used by    Not tainted
radeon                102460   1 
snd-seq-midi            4096  10  (autoclean)
snd-seq-midi-event      3840   0  (autoclean) [snd-seq-midi]
snd-usb-audio          45056   2 
snd-via82xx            13792   0 
snd-ac97-codec         41400   0  [snd-via82xx]
snd-mpu401-uart         3680   0  [snd-via82xx]
snd-hdsp               34436   3 
snd-hwdep               5152   0  [snd-hdsp]
snd-rawmidi            14976   0  [snd-seq-midi snd-usb-audio
snd-mpu401-uart snd-hdsp]
snd-pcm-oss            39588   0  (unused)
snd-pcm                64676   0  [snd-usb-audio snd-via82xx snd-hdsp
snd-pcm-oss]
snd-page-alloc          6708   0  [snd-via82xx snd-hdsp snd-pcm]
snd-mixer-oss          13624   0  [snd-pcm-oss]
hid                    15300   1 
snd-seq                39312   0  [snd-seq-midi snd-seq-midi-event]
snd-timer              15620   0  [snd-pcm snd-seq]
snd-seq-device          4384   0  [snd-seq-midi snd-rawmidi snd-seq]
snd                    32676   0  [snd-seq-midi snd-seq-midi-event
snd-usb-audio snd-via82xx snd-ac97-codec snd-mpu401-uart snd-hdsp
 snd-hwdep snd-rawmidi snd-pcm-oss snd-pcm snd-mixer-oss snd-seq
snd-timer snd-seq-device]
sbp2                   18112   0  (unused)
raw1394                 8056   0  (unused)
ohci1394               18312   0  (unused)
ieee1394               35692   0  [sbp2 raw1394 ohci1394]
Wizard root # 





-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf

RE: MIDI getting killed by Jack?

[Mark Knecht]

Hi. Does anyone else experience very slow responses from Alsa-Devel? I sent
this email this morning at 9:16AM this morning and it just arrived back here
and at my home account at 1:30PM. Most other reflectors I'm on normally have
round trip times of just a minute or two. Alsa-devel is always so slow,
usually taking hours.

At lunch today I rewired my MIDI input from the keyboard controller to use
the MidiSport input instead of the HDSP 9652 input. In this configuration,
with only about 15 minutes of testing, I had no problems. I will do more
testing this evening, but possibly this data says it's an HDSP 9652 MIDI
problem, or the HDSP 9652 MIDI combined with Jack-type audio, and not
specifically 'Alsa MIDI'.

With this set of connections:

aconnect 80:0 64:0
aconnect 80:0 64:32
aconnect 80:0 80:0
aconnect 80:0 80:1
aconnect 80:1 64:0
aconnect 80:1 64:32
aconnect 80:1 80:0
aconnect 80:1 80:1
aconnect 80:1 130:0

I had no immediate problems driving zynaddsubfx on 130:0. (I think it was
130:0)

Cheers,
Mark

> -----Original Message-----
> From: alsa-devel-admin@lists.sourceforge.net
> [mailto:alsa-devel-admin@lists.sourceforge.net]On Behalf Of Mark Knecht
> Sent: Wednesday, September 10, 2003 9:16 AM
> To: Alsa-Devel
> Subject: [Alsa-devel] MIDI getting killed by Jack?
>
>
> Hi,
>    I posted this last night on LAU but received no responses. Either it
> was of no interest or just no one had anything to say about it.
>
>    Is anyone else experiencing this problem? I set up a bunch of
> aconnect commands to route my keyboard controller (64:0) to a number of
> synths on different MIDI interfaces, and also to route the output of Pro
> Tools (on another computer - 80:0) to pretty much the same places:
>
> aconnect 64:0 64:32
> aconnect 64:0 80:1
> aconnect 80:0 64:0
> aconnect 80:0 64:32
> aconnect 80:0 80:1
>
> I also route Pro Tools output back to my keyboard (line 3) in case I
> want to plug another hardware synth in there, but nothing is hooked up
> to that right now.
>
> At this point I set up my mixer and play. Everything works and is
> perfectly fine for long periods of time.
>
> I then start Jack and within a minute or two I lose MIDI connectivity. I
> can run alsaplayer or other audio apps and I still get audio, but there
> is no more MIDI connectivity. I look at the synths and they are not
> receiving MIDI input. (No event lights.)
>
> I then stop Jack and I immediately hear a bunch of MIDI notes get
> played. These notes were stored from when I wasn't getting connectivity.
> At this point I have MIDI back for live playing until the next time I
> start Jack when it happens again.
>
> I do this 4 or 5 times and at this point I cannot get MIDI back at all.
> I have to completely power down as far as I can tell. I tried rebooting
> and had no luck the one time I did that.
>
> I don't know what to look at to figure out what could be causing this.
> Hardware and modules shown below.
>
> In thinking about this overnight it dawned on me that I am only seeing
> this problem on MIDI generated by my controller hooked to the HDSP
> MIDI input. I do not see this problem on MIDI generated by Pro Tools
> hooked to the MidiSport. I will likely rewire what is hooked to what
> this evening and see if I can say that it is only the HDSP input, or
> something else.
>
> Thanks,
> Mark
>
>
> 64:X - HDSP 9652
> 80:X - MidiSport 2x2
>
> Wizard root # lspci
> 00:00.0 Host bridge: VIA Technologies, Inc. VT8366/A/7 [Apollo
> KT266/A/333]
> 00:01.0 PCI bridge: VIA Technologies, Inc. VT8366/A/7 [Apollo
> KT266/A/333 AGP]
> 00:09.0 Ethernet controller: Broadcom Corporation BCM4401 100Base-T (rev
> 01)
> 00:0a.0 FireWire (IEEE 1394): Texas Instruments TSB12LV23 IEEE-1394
> Controller
> 00:0e.0 Multimedia audio controller: Xilinx Corporation RME Hammerfall
> DSP (rev 65)
> 00:10.0 USB Controller: VIA Technologies, Inc. USB (rev 80)
> 00:10.1 USB Controller: VIA Technologies, Inc. USB (rev 80)
> 00:10.2 USB Controller: VIA Technologies, Inc. USB (rev 80)
> 00:10.3 USB Controller: VIA Technologies, Inc. USB 2.0 (rev 82)
> 00:11.0 ISA bridge: VIA Technologies, Inc. VT8235 ISA Bridge
> 00:11.1 IDE interface: VIA Technologies, Inc.
> VT82C586A/B/VT82C686/A/B/VT8233/A/
> C/VT8235 PIPC Bus Master IDE (rev 06)
> 00:11.5 Multimedia audio controller: VIA Technologies, Inc.
> VT8233/A/8235 AC97 Audio Controller (rev 50)
> 01:00.0 VGA compatible controller: ATI Technologies Inc Radeon R250 If
> [Radeon 9000] (rev 01)
> 01:00.1 Display controller: ATI Technologies Inc Radeon R250 [Radeon
> 9000] (Secondary) (rev 01)
>
> Wizard root # lsmod
> Module                  Size  Used by    Not tainted
> radeon                102460   1
> snd-seq-midi            4096  10  (autoclean)
> snd-seq-midi-event      3840   0  (autoclean) [snd-seq-midi]
> snd-usb-audio          45056   2
> snd-via82xx            13792   0
> snd-ac97-codec         41400   0  [snd-via82xx]
> snd-mpu401-uart         3680   0  [snd-via82xx]
> snd-hdsp               34436   3
> snd-hwdep               5152   0  [snd-hdsp]
> snd-rawmidi            14976   0  [snd-seq-midi snd-usb-audio
> snd-mpu401-uart snd-hdsp]
> snd-pcm-oss            39588   0  (unused)
> snd-pcm                64676   0  [snd-usb-audio snd-via82xx snd-hdsp
> snd-pcm-oss]
> snd-page-alloc          6708   0  [snd-via82xx snd-hdsp snd-pcm]
> snd-mixer-oss          13624   0  [snd-pcm-oss]
> hid                    15300   1
> snd-seq                39312   0  [snd-seq-midi snd-seq-midi-event]
> snd-timer              15620   0  [snd-pcm snd-seq]
> snd-seq-device          4384   0  [snd-seq-midi snd-rawmidi snd-seq]
> snd                    32676   0  [snd-seq-midi snd-seq-midi-event
> snd-usb-audio snd-via82xx snd-ac97-codec snd-mpu401-uart snd-hdsp
>  snd-hwdep snd-rawmidi snd-pcm-oss snd-pcm snd-mixer-oss snd-seq
> snd-timer snd-seq-device]
> sbp2                   18112   0  (unused)
> raw1394                 8056   0  (unused)
> ohci1394               18312   0  (unused)
> ieee1394               35692   0  [sbp2 raw1394 ohci1394]
> Wizard root #
>
>
>
>
>
> -------------------------------------------------------
> This sf.net email is sponsored by:ThinkGeek
> Welcome to geek heaven.
> http://thinkgeek.com/sf
> _______________________________________________
> Alsa-devel mailing list
> Alsa-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/alsa-devel
>
>




-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf

Re: MIDI getting killed by Jack?

[Clemens Ladisch]

Mark Knecht wrote:
> Hi. Does anyone else experience very slow responses from Alsa-Devel? I sent
> this email this morning at 9:16AM this morning and it just arrived back here
> and at my home account at 1:30PM. Most other reflectors I'm on normally have
> round trip times of just a minute or two. Alsa-devel is always so slow,
> usually taking hours.

The "Received" headers of your mails show the following arrival times:

                               first mail                   second mail
ladisch.de                     10 Sep 2003 20:23:30 -0000   10 Sep 2003 20:47:18 -0000
sc8-sf-list2.sourceforge.net   10 Sep 2003 12:56:20 -0700   10 Sep 2003 13:28:43 -0700
sc8-sf-list1.sourceforge.net   10 Sep 2003 09:16:28 -0700   10 Sep 2003 13:13:40 -0700
sc8-sf-mx1.sourceforge.net     10 Sep 2003 09:16:27 -0700   10 Sep 2003 13:13:39 -0700
vivaldi.controlnet.com         10 Sep 2003 09:48:12 -0700   10 Sep 2003 13:45:27 -0700

This translates to the following mail processing durations:

sc8-sf-list2.sourceforge.net   0:27:10   0:18:35
sc8-sf-list1.sourceforge.net   3:39:52   0:15:03
sc8-sf-mx1.sourceforge.net     0:00:01   0:00:01
vivaldi.controlnet.com        -0:41:45  -0:31:48

If we assume that all clocks are set correctly, we can conclude:

1) SourceForge's list servers are overloaded at times.

2) Your ISP (controlnet.com) has a time machine.  :-)

> At lunch today I rewired my MIDI input from the keyboard controller to use
> the MidiSport input instead of the HDSP 9652 input. In this configuration,
> with only about 15 minutes of testing, I had no problems. I will do more
> testing this evening, but possibly this data says it's an HDSP 9652 MIDI
> problem, or the HDSP 9652 MIDI combined with Jack-type audio, and not
> specifically 'Alsa MIDI'.

When an interrupt for MIDI input occurs, the hdsp driver disables all
further MIDI interrupts until the current input data has been read. I
don't know why it does this, but you may try to remove/disable lines
3181, 3182, 3188, and 3189 in hdsp.c.


HTH
Clemens




-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf

Re: MIDI getting killed by Jack?

[Paul Davis]

>When an interrupt for MIDI input occurs, the hdsp driver disables all
>further MIDI interrupts until the current input data has been read. I
>don't know why it does this, but you may try to remove/disable lines
>3181, 3182, 3188, and 3189 in hdsp.c.

the hdsp driver is one of the first alsa drivers to use tasklets to
handle MIDI without impacting audio latency (that is, we can continue
processing audio i/o even while MIDI i/o is pending). once we have
launched a tasklet to handle MIDI input, there is no point handling
more interrupts that tell us the same thing.

the tasklet stuff relies on the kernel scheduler, but AFAIK, the
tasklet effectively runs like the kernel equivalent of a SCHED_FIFO
thread. i don't know at what priority, however.

--p


-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf

RE: MIDI getting killed by Jack?

[Mark Knecht]

Paul,
   I'm pretty sure that I was not the intended audience for this level of
technical data, but it does raise the question (for me) as to whether this
problem might be created by using the 'wrong' (I.e. - undocumented
requirements) kernel?

   The machine exhibiting the problem is running a pretty much vanilla
Gentoo 2.4.20 kernel with preemptive multitasking and low latency scheduling
turned on. This kernel does not have the capabilities patch. The problem
occurs when I run Jack as a user.

   I have not investigated what happens as root.

   However, as I stated in my second message yesterday, none of the above
seems to impact the system if I get my MIDI input through the MidiSport vs.
the HDSP 9652, so my presumption is that the first-order problem lies with
that driver and not the kernel.

   Please let me know specifically if there is anything you would like me to
try.

Thanks,
Mark

> -----Original Message-----
> From: Paul Davis [mailto:paul@linuxaudiosystems.com]
> Sent: Thursday, September 11, 2003 6:06 AM
> To: Clemens Ladisch
> Cc: Mark Knecht; Alsa-Devel
> Subject: Re: [Alsa-devel] MIDI getting killed by Jack?
>
>
> >When an interrupt for MIDI input occurs, the hdsp driver disables all
> >further MIDI interrupts until the current input data has been read. I
> >don't know why it does this, but you may try to remove/disable lines
> >3181, 3182, 3188, and 3189 in hdsp.c.
>
> the hdsp driver is one of the first alsa drivers to use tasklets to
> handle MIDI without impacting audio latency (that is, we can continue
> processing audio i/o even while MIDI i/o is pending). once we have
> launched a tasklet to handle MIDI input, there is no point handling
> more interrupts that tell us the same thing.
>
> the tasklet stuff relies on the kernel scheduler, but AFAIK, the
> tasklet effectively runs like the kernel equivalent of a SCHED_FIFO
> thread. i don't know at what priority, however.
>
> --p
>
>




-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf

Re: MIDI getting killed by Jack?

[Clemens Ladisch]

Paul Davis wrote:
> >When an interrupt for MIDI input occurs, the hdsp driver disables all
> >further MIDI interrupts until the current input data has been read.
>
> the hdsp driver is one of the first alsa drivers to use tasklets to
> handle MIDI without impacting audio latency (that is, we can continue
> processing audio i/o even while MIDI i/o is pending). once we have
> launched a tasklet to handle MIDI input, there is no point handling
> more interrupts that tell us the same thing.
>
> the tasklet stuff relies on the kernel scheduler, but AFAIK, the
> tasklet effectively runs like the kernel equivalent of a SCHED_FIFO
> thread. i don't know at what priority, however.

AFAIK tasklets have higher priority than any user thread.

>From the error description I diagnose that either MIDI interrupts or
the MIDI tasklet are suppressed while jack is running. It shouldn't be
possible to prevent a tasklet from running unless the CPU usage is
100%, so I think it's somehow related to interrupts.

Maybe the hardware doesn't set (or clears) the MIDI bit when too many
PCM interrupts occur, but I'm just guessing here.


Regards,
Clemens




-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf

RE: MIDI getting killed by Jack?

[Mark Knecht]

>
> AFAIK tasklets have higher priority than any user thread.
>
> From the error description I diagnose that either MIDI interrupts or
> the MIDI tasklet are suppressed while jack is running. It shouldn't be
> possible to prevent a tasklet from running unless the CPU usage is
> 100%, so I think it's somehow related to interrupts.
>
> Maybe the hardware doesn't set (or clears) the MIDI bit when too many
> PCM interrupts occur, but I'm just guessing here.
>
>
> Regards,
> Clemens

Hi,
   I wanted to remind people that the version of Alsa that I am using is
0.9.6, but the hdsp 9652 driver is not the one in CVS. It's the one Thomas
developed remotely at his place and has been tested on my machine and
possibly a few others. I think a patch was supplied to asla-devel but I'm
not sure about that. To the best of my knowledge Thomas didn't do any
editing on the MIDI side of things, but we didn't talk about that level of
detail much.

   I just didn't want anyone to be surprised later if it turns out it has
something to do with the audio side of the hdsp 9652 driver.

- Mark




-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf

Re: MIDI getting killed by Jack?

[Clemens Ladisch]

Paul Davis wrote:
> >When an interrupt for MIDI input occurs, the hdsp driver disables all
> >further MIDI interrupts until the current input data has been read.

... and midi_input_trigger flushes the input FIFO if the interrupt bit
isn't set.

Unfortunately, this does not seem to be the bug we're searching for.
:)


Regards,
Clemens




-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf

RE: MIDI getting killed by Jack?

[Mark Knecht]

> > At lunch today I rewired my MIDI input from the keyboard
> controller to use
> > the MidiSport input instead of the HDSP 9652 input. In this
> configuration,
> > with only about 15 minutes of testing, I had no problems. I will do more
> > testing this evening, but possibly this data says it's an HDSP 9652 MIDI
> > problem, or the HDSP 9652 MIDI combined with Jack-type audio, and not
> > specifically 'Alsa MIDI'.
>
> When an interrupt for MIDI input occurs, the hdsp driver disables all
> further MIDI interrupts until the current input data has been read. I
> don't know why it does this, but you may try to remove/disable lines
> 3181, 3182, 3188, and 3189 in hdsp.c.

Thanks Clemens. I'll take a look at this over the weekend probably, although
I wonder about your response to Paul's message on this subject?

Cheers,
Mark





-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf

RE: MIDI getting killed by Jack?

[alsa-devel]

Hello,

Sorry I didn't follow-up on this sooner; I must have been
distrcted by some shiny object.  I haven't had any problems
with MIDI and Jack.  I've tested routing MIDI in the following
ways:
    HDSP port 0 loopback
    HDSP port 1 -> port 0
    HDSP port 0 -> aseqview -> HDSP port 0
    HDSP port 1 -> aseqview -> HDSP port 0
    All of the above plus playback on port 0 using pmidi
I don't have another card with MIDI I/O. I tried starting Jack
both before and after doing the MIDI routing (not for every
possible combination).  I ran Jack overnight with the
port 1 -> port 0 setup.  No problem.

I am running Jack 0.80.0, compiled locally, kernel 2.4.21,
unpatched, compiled locally, and ALSA as described below.  I
wonder if the problem you found has something to do with
low-latency or preemptibility patches.

--Rob


Quoting Mark Knecht <mknecht@controlnet.com>:

> Rob,
>    I wondered if you tried finding the MIDI/Jack problem? I actually
> haven't
> revisited the issue as I'm using a MidiSport to get around the
> issue.
> 
>    I just wanted to drop you a note before I lost your email. ;-)
> 
> Cheers,
> Mark
> 
> > -----Original Message-----
> > From: alsa-devel@robdye.net [mailto:alsa-devel@robdye.net]
> > Sent: Thursday, September 11, 2003 3:40 PM
> > To: Mark Knecht
> > Subject: RE: [Alsa-devel] MIDI getting killed by Jack?
> >
> >
> > Quoting Mark Knecht <mknecht@controlnet.com>:
> >
> > > Hi,
> > >    I wanted to remind people that the version of Alsa that I am
> using is
> > > 0.9.6, but the hdsp 9652 driver is not the one in CVS. It's the
> > one Thomas
> > > developed remotely at his place and has been tested on my machine
> and
> > > possibly a few others. I think a patch was supplied to
> > asla-devel but I'm
> > > not sure about that.
> >
> > Yes.  The patch was posted.  I applied it to what was mostly alsa
> 0.9.5
> > but with the latest CVS hdsp.c at the time.  There was one conflict
> in the
> > patch (an extra firmware version case that was already in there),
> and the
> > hdsp.h posted was missing a macro for one of the ioctls (I have a
> very
> > small patch sitting somewhere, but it is probably irrelevant by
> > now anyway).
> >
> > The upshot is that audio I/O seems to be working great on my HDSP
> 9652
> > (Thanks!).  I haven't done any extensive multichannel testing;
> maybe I'll
> > get a chance this weekend.  However, hdspmixer still causes the
> > same message
> > from the driver via syslogd about hardware metering not supported.
> >
> > MIDI also works, at least on the first IN/OUT.  I haven't tried
> using JACK
> > and MIDI at the same time.  I'll give that a try tonight.
> >
> > --Rob Dye
> >
> >
> >
> 
> 
> 





-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf