I refactored my example and got same result as above. I hope this narrows down the problem. Also used wireshark to look traffic between test client and engine, but didn't find any problem there (so, most probably my test client script and visionappster.js is working fine).
Nothing to be sorry about. This is really valuable feedback to us. I cannot directly tell what's going on, but we are trying to reproduce the problem in a controlled setting. The two issues could indeed be related. Thanks for reporting.
@mt Nothing. But when I start it from command line:
jocc@gep1deb:~$ /usr/bin/flatpak run --branch=master --arch=x86_64 --command=va-builder com.visionappster.Builder
Qt: Session management error: None of the authentication protocols specified are supported
qml: Could not start web server.
Model size of -784 is less than 0
Model size of -112 is less than 0
The first 2 is immediately at start, the last 2 (Model size...) are written at the moment of "Add images" !
@dantcamui This error generally means that the version of the engine that you have installed does not recognize some of the tools used in your app. Running it in the same (or newer) VisionAppster release as the one the app was created with should fix it.
A bug fix for this popup problem amongst quite a few other fixes will be available in a couple of days in beta 6 release, stay tuned! Should anything similar ever happen, there is also a shortcut CTRL+S for saving the project, which may be worth a try.
Thanks for the reports. To be able to verify and fix issues we need to know the exact steps you took to trigger a bug. Now, I'm not quite sure if I understood your situation correctly. However, here's how I see it:
The sensitivity parameter is documented to have an effect only in the dynamic thresholding mode here: https://doc.visionappster.com/3.0-beta6/tools/detectblobstool/#detect-blobs. Some tools have input parameters that have an optional effect. They don't however disappear to avoid breaking connections.
invert in Find Threshold is an output parameter (function return value). It is calculated from input data and cannot be manually changed.
You mean the Mask Image tool? RGB triplets can represent gray levels, and the tool does the RGB->gray conversion automatically. Did you try it? Is the output incorrect?
This looks like a genuine bug. We were however unable to reproduce it. Can you send us the exact steps?