You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
This is convenient when one is working with GATs rather than the associated theory modules but then need to do something with the theory module, e.g. using a theory morphism to do a model migration.
However, when testing this with Catlab, this no longer seems to work. Catlab adds a new theory, T, and during precompilation one can check that T is in GAT_MODULE_LOOKUP. But if one loads up GATlab and Catlab in the REPL, T is no longer in the cache (only the theories added in GATlab are there).
The
@theory
macro includes a line to register each theory module, indexed by the data of the theory.This is convenient when one is working with GATs rather than the associated theory modules but then need to do something with the theory module, e.g. using a theory morphism to do a model migration.
However, when testing this with Catlab, this no longer seems to work. Catlab adds a new theory,
T
, and during precompilation one can check thatT
is inGAT_MODULE_LOOKUP
. But if one loads up GATlab and Catlab in the REPL,T
is no longer in the cache (only the theories added in GATlab are there).This might be relevant https://discourse.julialang.org/t/mutating-global-variable-during-precompilation/51478/3 \
In the immediate future, one workaround could be to only use
@theorymap
modules (which have the dom and codom modules) for migration.The text was updated successfully, but these errors were encountered: