RP1000: update_modifier_vbox: assertion failed

[Aaron Spike]

I just tried running gdigi for the first time with my RP1000. I used HEAD from github and I get errors very similar to the errors described in this thread:

https://groups.google.com/forum/#!topic/gdigi/JyW97BeZTDE

I tried to look through the change sets from the fix described there, but I can't make heads or tails of what I need to change to fix this.

My debug log is attached.

Thanks for any help. I would love to learn how to fix this.

Aaron Spike

[Tim LaBerge]

Hi Aaron,

Thanks for giving gdigi a spin on the RP1000. It looks like there are
few features that it has that aren't yet supported.

Here's a branch that has

1) Some fixes for unsupported features on the RP1000.

2) Some debugging code that will help us implement the remaining
unsupported features.

https://github.com/tlaberge/gdigi/tree/rp1000

The most important change is that instead of asserting when an
unsupported feature, the code will now emit a debug message and put a
label with the id/position in the gui so we can see where we need to add
support. The gui should work fine except for the missing effects/features.

You can see from the diffs in the first commit how to add support for a
missing feature. Basically, you need a modifier[], xml_settings[], and
possibly some labels.

The positions 28-34 seem to be mostly unsupported. I don't have any
documentation on what effects these correspond to. I'll try to look at
Digitech's website to see if I can get some hints, but what you can do
is fire up "gdigi -D a" with the new code and twist knobs/buttons to
figure out what's going on. If you run the windows gui, you can save a
couple of xml settings which will also give us some hints.

Usually, it takes a couple of rounds of debugging to get everything
figured out. If you can provide the console output and a few xml
settings files, we should be able to make fast progress.

Thanks again,

Tim

> --
> You received this message because you are subscribed to the Google
> Groups "gdigi" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to gdigi+un...@googlegroups.com.
> To post to this group, send email to gd...@googlegroups.com.
> Visit this group at http://groups.google.com/group/gdigi.
> For more options, visit https://groups.google.com/groups/opt_out.

[Tomasz Moń]

On Wed, Dec 25, 2013 at 11:28 PM, Tim LaBerge <tlab...@visi.com> wrote:
> Here's a branch that has

Do you think if this branch needs any more work before merging? Is it
release-ready?

> 1) Some fixes for unsupported features on the RP1000.

In the meantime I have merged Aaron's pull request. Could you please
check if these changes contain all required additions for the (yet)
unsupported features?

> The most important change is that instead of asserting when an unsupported
> feature, the code will now emit a debug message and put a label with the
> id/position in the gui so we can see where we need to add support. The gui
> should work fine except for the missing effects/features.

That's great!

> If you run the windows gui, you can save a couple of xml
> settings which will also give us some hints.

Alternatively, you can use USBPcap with Wireshark on the Windows
machine whilst running X-Edit. Wireshark has a built-in dissector for
X-Midi protocol used in DigiTech products. You can filter on "sysex"
and look up the missing values whilst turning the knobs in X-Edit.

Please notice that Wireshark most likely won't be able to correlate
the ID/Position-pair with effect name. For that you either need to
check the xml file or correlate it on your own during the live
capture.

Tim, the sample file on USBPcap website contains log for RP250.
Please, take a look at this, I think you will find this
Wireshark/USBPcap helpful for this job.

[Tim LaBerge]

Hi Tomasz,

On 12/26/2013 01:12 AM, Tomasz Moń wrote:
> On Wed, Dec 25, 2013 at 11:28 PM, Tim LaBerge <tlab...@visi.com> wrote:
>> Here's a branch that has
> Do you think if this branch needs any more work before merging? Is it
> release-ready?

I'll pick up Aaron's changes before I send a pull request.

>
>> 1) Some fixes for unsupported features on the RP1000.
> In the meantime I have merged Aaron's pull request. Could you please
> check if these changes contain all required additions for the (yet)
> unsupported features?

It looks like everything is there.

>
>> The most important change is that instead of asserting when an unsupported
>> feature, the code will now emit a debug message and put a label with the
>> id/position in the gui so we can see where we need to add support. The gui
>> should work fine except for the missing effects/features.
> That's great!
>
>> If you run the windows gui, you can save a couple of xml
>> settings which will also give us some hints.
> Alternatively, you can use USBPcap with Wireshark on the Windows
> machine whilst running X-Edit. Wireshark has a built-in dissector for
> X-Midi protocol used in DigiTech products. You can filter on "sysex"
> and look up the missing values whilst turning the knobs in X-Edit.
>
> Please notice that Wireshark most likely won't be able to correlate
> the ID/Position-pair with effect name. For that you either need to
> check the xml file or correlate it on your own during the live
> capture.
>
> Tim, the sample file on USBPcap website contains log for RP250.
> Please, take a look at this, I think you will find this
> Wireshark/USBPcap helpful for this job.
>

---tjl

[Aaron Spike]

Those debugging changes are a big help. No more crashes. I've sent a pull request with some additional modifiers to eliminate the debug messages.

Is there any documentation that covers the gdigi program structure or even just the sysex protocol in use?