Join GitHub today
GitHub is home to over 28 million developers working together to host and review code, manage projects, and build software together.Sign up
Handle objects that overload stringification #7
I wasn't ignoring your PR, I just have to work myself into the right frame of mind before diving back into the internals of XML::Simple. It's not a quick process.
Did you end up deploying code that relied on the change in your pull request? Or did it turn out to be a terrible idea for other reasons, causing you to give up and abandon it? If you ended up using it, then I'm happy to merge the PR. If you didn't, then I'd just as soon close it and pretend this never happened.
Your PR didn't include a test, so I wrote one and in the process realised that your change won't work for all objects that override stringification, but only those objects that are not built on blessed hashrefs or blessed arrayrefs. Apparently they always get serialised as if they were just plain hashrefs and arrayrefs (including any secret internal keys and values). I had forgotten that XMLout() worked that way and frankly I can't believe I ever thought it was a good idea. But the fact is it does work that way and has for over 15 years so it's safe to assume someone is relying on it..
I'm now trying to work out what a change to the docs would look like. If I mention that XMLout() will no longer die on an object that overrides stringification then some people are going to be surprised that the stringification overrides in their classes are not being called.