I'm not convinced (5) and (8) are adequately resolved.
(5) If we make a plan to fix lack of const in 'MirBuffer *' everywhere soon, then OK. If you can change it to "const" in the below diff right now without breaking existing code then please do it before this lands.
(8) Similarly if we can cull that function and existing code still works, please do so. I'm not convinced we have the right design for handling 'age'. It's also not an important feature -- both uses of buffer age in toolkits I've ever seen were buggy and had to avoid it in the end.
I'm not convinced (5) and (8) are adequately resolved.
(5) If we make a plan to fix lack of const in 'MirBuffer *' everywhere soon, then OK. If you can change it to "const" in the below diff right now without breaking existing code then please do it before this lands.
(8) Similarly if we can cull that function and existing code still works, please do so. I'm not convinced we have the right design for handling 'age'. It's also not an important feature -- both uses of buffer age in toolkits I've ever seen were buggy and had to avoid it in the end.