DSP Robotics Support

RJHollins wrote:
RUBY hangs on to all this data internally ... I cannot figure out how to null everything.
I tried:

Code: Select all

container = nil   # see if container & file needs to be cleared before load ?!?
file = nil

tricky for me ...

I told you I'm just dangerous with this stuff

i added a mkdir in the ruby code which creates the directory when it is not existing, so you dont need the makefolder prim, further you can set the location pretty easy in the properties panel for fast testing of the different locations,

i also added a reload file 1.5 second after saving it to see if it loads succesfully (like trog mentioned)

and you`ll find a module inside with an arrayclass methode to set nil items in an array to a default value
which is per default "0"
so if you want to set all nils to zero you need no argument for the methode but you can use a specified default value for each array, there is a destructive and a nondestructive version of it!
but i`n not sure if you mean that?
anyway its there...

if you mean that the container from the save methode still exists when you call the load methode you are wrong because there is no @ in front of it so it only exists inside the methode, so in the load methode you create a new container which is filled from the file you try to load

if you mean that the container from the save methode still exists when you call the load methode you are wrong because there is no @ in front of it so it only exists inside the methode, so in the load methode you create a new container which is filled from the file you try to load

ahhh ... thanks for helping to understand better. In the effort to solve my 'BT' issue, I was going in circles inside my original Ruby code. Me and Ruby are at the stage that when something I've coded 'seems' to work, no one is more surprised than ME
Something else that catches me out is the 'retention' of data. For 'testing', I was looking to 'zero out' all arrays so that I can trace if code & logic were correct. From your explanation, I have a better understanding that hopefully more practice will solidify.

BTW ... this module you are creating is getting really cool The link to the 'IDs' ... very nice
I'm waiting on some saved files from 'BT', as we are having to baby-step through all this to figure out why he had problems ...

Once again ... Thank-you NuBeat !


**** edit ***
I'm going to use this module [slightly modified] for the 'control' test on 'BT' system. This way I can compare files saved.

RJHollins wrote:The link to the 'IDs' ... very nice

because i like it comfortable

Since I was made aware of this Marshal technique by TROG ... then your design fixes/enhancements ... for the type of data that I always seem to deal with ... this is a phenominal concept, and a powerful addition to FS and the Community.

I just hope that the issue my BT'r was having was due to MY error [whatever it is], or if this is strickly a Win 7/8 issue, that this can be corrected. I'm waiting for that file to be returned and compared.

I'm already preparing the next test [if all is well], using this module to load up the 'Selectors' with the lists.

The final verdict will be results from the VST form of this. This is the baby-step by step approach in troubleshooting.

Test file results are in.

I'm doing a visual compare between the file I save and the file my tester saved.

I see a difference Compared to what I get. He has 2 separate computers, a main Win 8.1, and a laptop Win7.

Both of the saved on his system LOOK the same [format wise - viewed in NOTEPAD] ... however, It is not the same look as mine. It is like there is an extra C/R return or NewLine character.

I'm attaching the 2 TXT files here. Maybe something can be spotted. These files are straight out of the latest routine posted. BTW ... these differences is what I saw when I worked up the routines in my VST ... so this issue is still there ???



It appears we can't upload TXT files, so I RAR'd up both files. Mine is in the root dir, Beta-tester in folder.

Why would there be a difference ... and importantly ... what impact does this have ???

******* edit ***

Houston ... we do have a problem ...

With the included 'error report' routines now implemented by NuBeat, I replaced my saved file with the file from my beta-tester.

I see the same error. "dump format error(0xa)" when attempting to load the file.

HELP

Anyone on Win 8 that can test this and post the saved file

Here is my cryptic one...

looks like that the marshalling of a file is done a little different in xp, but the question is if your tester can reload it - you dont need to reload his file!

Walter Sommerfeld wrote:Here is my cryptic one...

Thanks Walter ... Please, What OS's did you test this on ???

Nubeat7 » Sun Oct 27, 2013 9:38 am

looks like that the marshalling of a file is done a little different in xp, but the question is if your tester can reload it - you dont need to reload his file!

I have sent BT'r yet another test EXE to get additional confirmation. This test now includes 2 versions [SM & FS] of the Selector pull down menu. If the data that is Save and Auto-Loaded is indeed correct, these menus will show that.

To your point, NuBeat ... that thought had crossed my mind, but I've tried to dismiss it. [Wishful/hopefull thinking] Operationally, though ... if it works within each system ONLY, well ... that should be good but I don't really understand how this could be. Maybe I don't need to understand ... but its hard to be confident about this.

I hope to know more later in the day when results come back.

I am so grateful that you Guys have been so helpful with all of this. This single 'issue' has completely stop all my projects. If we find the EXE form to work ... we then have the ultimate 'VST format' to verify if working. I honestly never anticipated an issue such as this. It does seem we've had some positive results [I think this has more to do with NuBeats coding of this routine ... although I can't find what might be wrong with my attempted version ].

The wait is on. ....

As promised ... back with the test report.

The 'Test' module was expanded to include feeding Selector pull-down Menus with the results returned from the Auto-Load function. I used both the FS and SM modules of the Selectors. These were modified to handle feeds from a specific array.

The OS in question was both Win 7 & 8 ... but the focus centered more on the Win 8.1 system [BT'r main computer].

We began using the EXE format to eliminate any 'Host' influence. The test module was successful in all functions, including the feeding of both Selectors. Each performed as expected.
The Saved 'test file' was sent for comparison with my save. There is a definite difference in the layout format starting at the end of the first line [it appears]. The returned file is NOT compatible and generates a 'dump error' in my Ruby module. However, staying on our own systems this did not cause any internal errors and the Menus displayed and operated properly.

The 'ultimate' test followed as we then went to the VST format. Same code as before, slightly mod'd to allow VST export.

The results [using Reaper DAW].

Everything performed the same as the EXE version YEAH

I had insured that the Selector Menus were completely BLANK in the test. If you clicked on them, neither did anything ... they acted frozen.

It wasn't till the Saved data Auto-Loaded that they Menus displayed, and actually performed as designed. Unloading and reloading the VST in the same Reaper session confirmed that Reaper did NOT alter the initial BLANK status. The data had to be Save to re-load ... then both Selectors worked.

So ... why the test is working ?? Good question

I had hope that something would have jumped out in my original coding of the Marshal routine. As I compared it to NuBeats, which was organized different than mine ... but if we looked just to the specific, basic code ... I could not find an error in my code. In fact, the frustration throughout all this was each design and flavor used worked without issue on my computer [WinXP].

Possible significance difference ... NuBeat had a 1.5 sec delay add to the Re-load command after the Save. I had a 'done' trigger in my Ruby code that triggered a Delay Prim of 500ms. Being 'Green' I wonder if a combination of things were messing the execution order or something ... this is a mystery ...

Regardless ... I am deeply grateful to NuBeat for not only rewriting the code, but then to provide a series of enhancements to also include a, type of, front-end replacement. It is the FS replacement to the SM core module, that easily ties into this fantastic Marshal routine. Which alone has provided a great solution to the way I was having to store and load multiple arrays of data that had dynamic allocation in entry size. For a NewB as me, this has drastically reduced the internal management I had concocted prior.

Of course, this is not the first time that NuBeat has taken something to a much refined level. Really ... thank you very much for all this ! And to be very fair ... there are the other GURU's who have been so generous with their time, knowledge, and examples that have help me convert a idea into a working tool. Now I'm feeling better that others may be able to use my project ! Well, this is the next step ...

I'm looking to transplant/replace my original section with this tested module into my project, hoping that we can finally get out of the gate and actually test the program in a working state. All my BT'r has been able to do is see the GUI, parse the XML file, see the patch numbers ... but then nada ... just another pretty face that only a mudder would lov.

I also need to rest that last brain cell ... it took quite another frying these last couple weeks.

Even though I'm exhausted, I can appreciate some of the new things I have learned during this. I've also expanded my awareness of new things I need to explore. One of which was the XML parsing that I looked at. Only a very minor play, but I think this is an area I need to look at very closely. I only had a hint from trying to read in my XML file, but I saw some very powerful possibilities. Again, something I never really knew about ... and something I'll probably explore. But FIRST !!! Got to get this project running on the BT'r system