Based on

i added a check if the statement returned is from type TagComponent or not, if not the parser simply goes on

if that fixes this issue or not is unclear and as long i have not the source code causing this, i have no chance to verify.
i know that it must be some code with component<space> , but still valid code, i tried to reproduce the issue with code like this

but that did work just fine

without any further input on this and when all test cases passes after this change, we have to assume that the issue is fixed and close this ticket.
What is certain if their code still fail, they will see a different exception, that maybe gives more inside.

there is something else going on here, i also get a proper exception when i delete an extended component

try/finally is absolutely fine in that context, it simply clean up and then escalates the exception.

your exception completely differs from the exception in the description, yours is a runtime exception, the exception in the description is a compiler exception. something goes wrong when compiling code. compiler errors are easy to reproduce because they do not depend on a specific environment, so we simply need the template that fails to compile and when we have it, we can reproduce the issue.

I tried to reproduce this error, i.e. with an invalid model cfc extends, but the error message was useful for both orm 5.4 and 3.5.5

there’s a try finally without a catch here?

https://github.com/lucee/Lucee/blob/6.2/core/src/main/java/lucee/transformer/cfml/tag/CFMLTransformer.java#L868

ort we could add some context to this exception?
https://github.com/lucee/Lucee/blob/6.2/core/src/main/java/lucee/transformer/cfml/tag/CFMLTransformer.java#L382

so do we have an issue not to get a proper exception message then or is this a non issue?

Via slack:

So, I think this had to do with component file not found due to a directory mapping issue . This resulted in some components that use extends="path.to.basecomponent" not being able to get instantiated. Once I fixed the mappings, seemed to run properly both locally and via Docker