why use result to rather than panic to imp RuntimeException

[578223592]

hi， I'm sorry to disturb you. I would like to ask what considerations led to using Result to implement runtime exceptions instead of maintaining similar behavior to the Java version, i.e., directly causing a panic. For example：
java version：

                    throw new RuntimeException("only add attribute is supported while creating topic. key: " + realKey);

rust version

                    return Err(AttributeError::CreateOnlySupportsAdd(real_key.to_string()));

The reason for my confusion is that if an Err is returned instead of directly panic, it may lead to inconsistent behavior between the two versions. In fact, I noticed that after an error occurs, the Rust version continues to run, which indeed leads to some differences from the Java version：
link:
https://github.com/578223592/rocketmq-rust/blob/43e0d24fd356b81eae55cf2213acd75ddc3a43bc/rocketmq-broker/src/topic/manager/topic_config_manager.rs#L495

[578223592]

@mxsm

[mxsm]

@578223592 In Java, Exceptions are usually passed along and then caught somewhere to handle the errors. In Rust, I also adopt a similar approach. The error handling in Rust and Java has both differences and similarities. I'm still thinking about how to handle errors better in this project. If you have any better ideas, we can discuss them together.

[578223592]

@mxsm thanks for your reply.
My point is: If an unexpected error occurs (I haven't thought of a particularly suitable example in rust), use panic. If an expected error occurs (for example, the file does not exist or the network rate limiting policy is triggered), use result.

In this case, perhaps it is more appropriate to use panic.

of course,the baseline is keep the same behavior between java version and rust version ,we should stop the function