Jeff Garland wrote:
Robert Ramey wrote:
I want everyone to be successful using the library - even if they complain that it requires them to write good code.
^^^^^^^^^ IMO the current policy of the library actually forces me to write bad/ugly code because if I ever want to serialize something in a non-const context I have to force it to const artificially.
This will be short because I know we've been over this point endlessly in the past, but I'll try one more time. Make a macro to control this behavior:
BOOST_SERIALIZATION_CONST_STRICT
or whatever. Make this optional. When someone posts an issue with an archive, tell them to recompile with this option turned on so that they can help find the problem. Then the rest of us can write normal code and the library will behave as expected. Less people will need to read the fine print of the library.
This is a grand idea.
BTW, I suspect the decrease in bug reports might be a result of less folks being able to get "hello serialization" to compile and subsequently giving up on the library instead of the 'better' code that this 'enforces'.
Right. We want people to use the library, not scare them off. -Thorsten