Type derivation does not behave as specified

[Broes De Cat]

No description provided.

[Bart Bogaerts]

Bugfix in sort derivation.

See #683 Problem was that bottom-up propagation was forgotten.

Not yet finished: Should also happen for functerms, and I don't really know this code hence I am uncertain of correctness...

→ <<cset 3eb8009f7efa>>

[Bart Bogaerts]

Added test

see #683

→ <<cset bb78ee6b7186>>

[Bart Bogaerts]

Bugfix in deriveSorts:

Two problems are fixed here: first, during the loop, derivefuncs and derivepreds should be ran AFTER the accept As these two determine when to stop (i.e. _changed)

Now it was possible that the system stopped one iteration too early.

Second: _underivablevariables should be cleared every run. They are set during the top down phase and can never be removed from this set.

This fixes #683 for predicate symbols

For function symbols, there is a related issue, to be fixed soon

→ <<cset ab58ecbda15c>>

[Bart Bogaerts]

Added test

see #683

→ <<cset 406f229d8c80>>

[Broes De Cat]

Bugfix in deriveSorts:

Two problems are fixed here: first, during the loop, derivefuncs and derivepreds should be ran AFTER the accept As these two determine when to stop (i.e. _changed)

Now it was possible that the system stopped one iteration too early.

Second: _underivablevariables should be cleared every run. They are set during the top down phase and can never be removed from this set.

This fixes #683 for predicate symbols

For function symbols, there is a related issue, to be fixed soon

→ <<cset ea42f9e88b2a>>

[Broes De Cat]

Added test

see #683

→ <<cset 7c9c2e1641bf>>