The debug output from tree_buf::internal::read_root(&bytes); is useful to get a sense for what a file contains, but for large schemas, it is not organized in a way to answer questions at a glance. See, for example, in #6, that the output is pages long and full of noise.
An experimental API should be added that makes it easy to answer questions like:
What fields take up the most space?
What percentage of the file uses a particular encoding?
How many bytes are schema overhead and top-level fields?
These together help to answer what the improvement ceiling would be for any given feature.
The debug output from
tree_buf::internal::read_root(&bytes);
is useful to get a sense for what a file contains, but for large schemas, it is not organized in a way to answer questions at a glance. See, for example, in #6, that the output is pages long and full of noise.An experimental API should be added that makes it easy to answer questions like:
These together help to answer what the improvement ceiling would be for any given feature.