https://github.com/Xazax-hun commented:

I wonder if collecting in buckets indexed by `StmtClass<QualType>` is the right 
thing to do. 

If the type and the ast node kind are expected to be correlated, i.e., we 
expect to have holes where a combination of the type and the kind is the root 
cause something is not covered than this is the right way to slice the data.

On the other hand, if we expect these to be more independent, i.e., some types 
are not covered and some kinds of ast nodes are not covered and there is no 
correlation between them, I think it might be better to collect across these 
two independent axises.

So print one statistics that maps from statement class (ids, no need for 
strings there) to counts and print another one that maps from types (again no 
need for string maps) to counts.

https://github.com/llvm/llvm-project/pull/166568
_______________________________________________
cfe-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

Reply via email to