As in title, when trying to save a new project the Cube crashes generating a Java error log file.
Seems to happen just before opening a file dialogue window to ask where to save the new project.
Hi GeorgeP ,
Your issue is reported internally and under investigation.
Thanks for looking into it. Please keep me posted.
Can't CubeMX (and other Java-based ST programs) come also in an alternative flavour with the JVM bundled? That would create a standalone executable independent of Java being installed on a given machine, and possibly avoid incompatibility issues such as the one reported here.
Continuing to the original problem, I've clean installed Java and Cube, after removing all previous installations. Now I'm getting the following error. I guess a bundled version might help, as Jan noted.
"java -version" runs normally on the command prompt, so Java and its environmental variables should be correct.
Fixed that error and now I'm still getting the older error (the same as the one generating the log file).
With Windows and the Registry, things became confusing. Uninstallers often fail to remove all remains of the application's trash from the registry, and reliably prevent a successful re-installation.
You can check out one of those various "Registry cleaners", or try it on a clean machine.
Having a virtual machine for such purposes proved to be a good idea as well.
To be honest that's what I did today.
I've made an Ubuntu VM solely for Cube, and it's working pretty well.
That is kind of a different animal.
Unix/Linux uses a different approach to application configuration, and installation/de-installation.
In fact, the Unix community usually references the Windows Registry as "a perfect example of a single-point-of-fail".
They have other issues, though (think OpenJDK vs. Oracle JDK).
I guess a Windows VM would have worked equally well - until tainted ...
Concerning Win, didn't want to run into licensing problems.
In any case, my Cube projects are just a snapshot away if something bad happens, which makes me sleep easier. I should probably do the same with the compilers.
Retrieving data ...