A better fix would be to get rid of eval entirely, but that would require an entire parser rewrite. This is a best-effort fix to get out the door for now.
Note that this change imposes (and enforces) a stricter format for the schema syntax. It should support the large majority of existing use cases, unless you're doing something very janky or unusual. Still, as a precaution we will bump the major version number to prevent breaking existing users.
Quick best-effort fix for #167.
A better fix would be to get rid of
eval
entirely, but that would require an entire parser rewrite. This is a best-effort fix to get out the door for now.Note that this change imposes (and enforces) a stricter format for the schema syntax. It should support the large majority of existing use cases, unless you're doing something very janky or unusual. Still, as a precaution we will bump the major version number to prevent breaking existing users.