Minutes June 16 2020 - math-comp/math-comp GitHub Wiki
-
About the issues in order https://coq.zulipchat.com/#narrow/stream/237665-math-comp-devs/topic/hierarchy.2Eml
-
PY had problems extending order's hierarchy, KS tool has some issues and a workaround
- you can run it without compiling all.v but you have to manually Import (transitively)
- you can at least use the
all_something
to "factor" some imports
-
CC wonders if we could use
Export
in some .v files to transitively close automatically (if you Import A that Exports B, then you Import B automagically) -
the
Exports
modules are always Exported at the end of the files -
AM there was a bug w.r.t.
Export
and coercions (exporting twice the same module with a coercion was failing) -
conclusion: we should try this
-
PY: order is very complex, hard to do by hand, let's hope Hierarchy Builder comes soon
-
PY: CS indexed on the type, not the order relation, but you have many (on the same type) and today you have to alias the type.
- CC: we should have both. Roadmap:
- HB only indexes on types for now
- Next morphisms (index on functions)
- Next mixins on different indexes.
- ET: we should organize a meeting where we try to port order to HB
- CC: we should have both. Roadmap:
-
-
About OPAM naming policy
- we could not find a rule for cutting it
- maybe should be in sync with the GH organization
-
About packaging
- google points to Guillaume Claret blog which is outdated
- no single source of info/scripts
- ET: I would recommend coq-community
- but does not support choosing a prefix other than
coq-
for the name of the package
- but does not support choosing a prefix other than
- CC: improve nix template for coq-community, the "packager" tool should be used/advertised there
- The work should go into coq-community, not MC specific