delehef
commented
9 months ago we even can't know if the spec has been fully implemented in the .lisp constraints files.
That's why constraints in the spec ought to be generated from Corset LaTeX export.
Closed letypequividelespoubelles closed 7 months ago
delehef
commented
9 months ago we even can't know if the spec has been fully implemented in the .lisp constraints files.
That's why constraints in the spec ought to be generated from Corset LaTeX export.
letypequividelespoubelles
commented
9 months ago That's why constraints in the spec ought to be generated from Corset LaTeX export.
Ought to, but are not imho. Because we want the spec to be beautiful /readable : we add comments, and the spec constraints are sometimes adapted when written in lisp to lower number of constraint.
We currently have no way to ensure traces are sufficiently constrained, but we even can't know if the spec has been fully implemented in the
.lispconstraints files. While debugging, it could be nice to have some checks to help finding missing constraint implementation:[ ] Some columns must be constrained on all cells apart obvious constraints (like byte decomposition, CT-constraints, etc). We could have a check
for cells in columns, it exists constraints apart obvious one.[ ] Let's assume that cells without a constraints are filled with
0. Corset could add a constraint, that if a cell is not constrained yet, then it should be zero.