buddy | A cross platform app

Not sure how much of an answer this will be, but the issue is now (seemingly) self-fixed... I'll describe what I've done and what happened in case it helps other users experiencing similar issues.

After unsuccessfully trying all the things described in the question, I had to restart the laptop (again) for an unrelated problem (Zoom wouldn't start up). Upon restart the laptop simply crashed and restarted again on its own. Then it did the same a second time. Upon the third automatic restart, I was no longer experiencing issues building my Java applications.

My guess is that the NVRAM got somehow corrupted and this was affecting the JVM's ability to accept incoming requests for attachment. On a MBP M1 one cannot reset the NVRAM as with Intel models (pressing Cmd + Opt + R + P), instead the NVRAM is supposed to self-heal when it detects something wrong. I suppose after crashing the MBP finally decided to repair the NVRAM and that's why the build starting working again.

UPDATE: I have now encountered this issue a few more times and every time the "solution" was to restart the laptop a few times until it crashed. After that the issue would disappear. This adds weight to the idea that there is some inherent instability with the combination of hardware and software specified in the question. I'll leave this as an "answered" question in case anybody else has the same issue.

UPDATE 2: This may be pure coincidence, but the issue seems to be exacerbated by the usage of an additional screen, particularly using an iPad via Sidecar.

The problem with this approach is that the field accessor considers the input type prior to the modification. Byte Buddy prohibits this as it does not consider the mutation to be legal, not knowing about the removed modifier. As a result, the transformation fails in its entirety and you get the error you are seeing. (Register a listener to see this error.)

To avoid this, you can implement a custom Implementation using FieldAccess (without or). You can have a look at the more convenient FieldAccessor to see how this is implemented, only that you need to drop the validity checks.

As you speak about the PROD database - of course, a better option is to make backup/restore local copy and play with it. But regarding the first part of your question you can use ssh port forwarding e.g.:

Use setjmp() & longjmp() pair with alarm().

You need to add setters to your properties, otherwise the deserializer can't assign values to those properties

Acording to this answer: https://stackoverflow.com/a/51236918/16651073 tomcat falls back to default logging if it can resolve the location

Can you try to save the properties without the spaces.

Like this: logging.file.name=application.logs

You can create such a class. Have a look at the DSL and FieldAccessor for creating (overriding) methods that access fields.

On the other hand, it seems to me that you should rather look at MethodHandles which can access field values as fast as any Java byte code once generated. And I argue that resolving a method handle takes less resources then creating a class.

You are loading the class into a child loader of the boot loader by using

I just accessed your database with the following URL: