Added message on singleton exception. #6
Conversation
…ave some try/catch with std::exception that makes gdb not break when this is thrown, because it is catch but there is no what() so I didn't knew where it came from. This is just an addition of message that state clearly where the exception comes from and why. Maybe the message could be enhanced, obviously. I didn't compile this yet.
|
Actually, I'm not sure why this code throws an exception. The whole point of the name AutoSingleton is that it self-instantiates if you access it before explicitly instantiating it. It looks like 1dfef33 introduced the change, but doesn't have a comment as to why the change was made. @pathorn or @KevinGadd, do you remember why that change was made? |
|
Well it seems that this class follow the right pattern of forcing the user to call the creation function before doing anything with the singleton. I forgot to call it that's why I got the exception, so it's fine for me. The main problem is more that the exception itself wasn't as useful as it should have been. |
|
Yeah, it's not meaningful to use the singleton without having initialized Berkelium first. I don't know why the exception has no message, though. |
|
Laziness? :D I guess it's because there is no clear error strategy defined for the project, so the one who added the exception didn't know what to do exactly and did the simplest thing. Also, shouldn't it be exception_runtime? |
|
I fixed this slightly differently than the submitted patch in d7e6c9f, including changes to get rid of the name AutoSingleton since it is no longer accurate. Thanks for the initial version! |
I've been hunting for this exception with no message for an hour. I have some try/catch with std::exception that makes gdb not break when this is thrown, because it is catch but there is no what() so I didn't knew where it came from.
This is just an addition of message that state clearly where the exception comes from and why.
Maybe the message could be enhanced, obviously.
I didn't compile this yet.