| Age | Commit message (Collapse) | Author |
|---|
|
PR #13183 introduced quite a bit of duplicate code, we refactor it and
expose less internals on the way. That should make the API more
robust.
|
|
We store bound variable names instead of functions for both branches and
predicate, and we furthermore add the parameters in the node. Let bindings
are not taken into account and require an environment lookup for retrieval.
|
|
|
|
The syntax is the one of boolean attributes, that is to say
`#[typing($flag={yes,no}]` where `$flag` is one of `guarded`,
`universes`, `positive`.
We had to instrument the pretyper in a few places, it is interesting
that it is doing so many checks.
|
|
|
|
|
|
|
|
This shall be for Record fields consumption.
|
|
In our quest to unify all the declaration paths, an important step
is to account for the state pertaining to `Program` declarations.
Whereas regular proofs keep are kept in a stack-like structure;
obligations for constants defined by `Program` are stored in a global
map which is manipulated by almost regular open/close proof primitives.
We make this manipulation explicit by handling the program state
functionally, in a similar way than we already do for lemmas.
This requires to extend the proof DSL a bit; but IMO changes are
acceptable given the gain.
Most of the PR is routine; only remarkable change is that the hook is
called explicitly in `finish_admitted` as it had to learn about the
different types of proof_endings.
Note that we could have gone deeper and use the type system to refine
the core proof type; IMO it is still too preliminary so it is better
to do this step as an intermediate one towards a deeper unification.
|
|
Fix #12651
|
|
|
|
|
|
This allows us to remove a large chunk of the internal API, and is the
pre-requisite to get rid of [Proof_ending], and even more refactoring
on the declare path.
|
|
|
|
We unify information about constants so it is shared among all the
paths [interactive, NI, obligations].
IMHO the current setup looks pretty good, with information split into
a per-constant record `CInfo.t` and variables affecting mutual
definitions at once, which live in `Info.t`.
Main information outside our `Info` record is `opaque`, which is
provided at different moments in several cases.
There are a few nits regarding interactive proofs, which will go away
in the next commits.
|
|
This improves the interface, and allows even more sealing of the API.
This is yet work in progress.
|
|
At this point the record in lemmas was just a stub; next commit will
stop exposing the internals of mutual information, and pave the way
for the refactoring of `Info.t` handling in the Declare interface.
|
|
|
|
This is needed as a first step to refactor and unify the obligation
save path and state; in particular `Equations` is a heavy user of
Hooks to modify obligations state, thus in order to make the hook
aware of this we need to place the obligation state before the hook.
As a good side-effect, `inline_private_constants` and `Hook.call` are
not exported from `Declare` anymore.
|
|
The API in `DeclareDef` should become the recommended API in `Declare`.
This greatly reduces the exposure of internals; we still have a large
offender in `Lemmas` but that will be taken care of in the next
commit; effectively removing quite some chunks from `declare.mli`.
This PR originally introduced a dependency cycle due to:
- `Declare`: uses `Vernacexpr.decl_notation list`
- `Vernacexpr`: uses `ComHint.hint_expr`
- `ComHint`: uses `Declare.declare_constant`
This is a real cycle in the sense that `ComHint` would have also move
to `DeclareDef` in the medium term.
There were quite a few ways to solve it, we have chosen to
move the hints ast to `Vernacexpr` as it is not very invasive
and seems consistent with the current style.
Alternatives, which could be considered at a later stage are for
example moving the notations AST to `Metasyntax`, having `Declare` not
to depend on `Vernacexpr` [which seems actually a good thing to do in
the medium term], reworking notation support more deeply...
|
|
Co-authored-by: Jasper Hugunin <jasper@hugunin.net>
Co-authored-by: Gaëtan Gilbert <gaetan.gilbert@skyskimmer.net>
|
|
Ack-by: Matafou
Reviewed-by: SkySkimmer
|
|
This will allow to share the definition metadata for example with
obligations; a bit more work is needed to finally move the preparation
of interactive proofs from Proof_global to `prepare_entry`.
|
|
The only reasons that `prepare_definition` returned a sigma were:
- to obtain the universe binders to be passed to declare
- to obtain the UState.t to be passed to the equations hook
We handle this automatically now; it seems like a reasonably good API
improvement.
However, it is not clear what we do now is right for all cases; must
check.
|
|
We don't the parameter anymore as the paths are too different now.
Note that this removes a duplicate `check_evars` that has been in
`master` quite some time [double check in `ComDefinition` and in
`DeclareDef.prepare_definition`]
|
|
Preparation of obligation/program entries requires low-level
manipulation that does break the abstraction over `proof_entry`; we
thus introduce `prepare_obligation`, and move the code that prepares
the obligation entry to its own module.
This seems to improve separation of concerns, and helps clarify the
two of three current models in which Coq operates w.r.t. definitions:
- single, ground entries with possibly mutual definitions [regular lemmas]
- single, non-ground entries with possibly mutual definitions [obligations]
- multiple entries [equations]
|
|
Proof "preparation" [as in `DeclareDef.prepare_definition`] is fairly
more complicated in the Program case; in particular, it includes
checking the existential variables, and elaborating a list of entries
from the holes.
Indeed, in the `Program` case we cannot use
`DeclareDef.prepare_definition` while keeping a good level of
abstraction, so we should introduce a `prepare_obligation` function.
This PR is in preparation for that.
|
|
We clean the hack bypassing the interpretation of "'pat" in binders
and move it to comDefinition.ml. In particular, we fix the exact
subterm to which Eval has to apply in the hack, and remove the
artificial cast we had introduced.
|
|
In anticipation for more consolidation of duplicated functionality.
|
|
Most of the parameters were named, we fix the remaining cases.
|
|
Add headers to a few files which were missing them.
|
|
This is useful to remove some duplicate bits in other declare files.
|
|
top of declare.
This PR is a follow up to #10406 , moving the then introduced
`proof_entry` type to `Declare`.
This makes sense as `Declare` is the main consumer of the entry type,
and already provides the constructors for it.
This is a step towards making the entry type private, which will allow
us to enforce / handle invariants on entry data better.
A side-effect of this PR is that now `Proof_global` does depend on
`Declare`, not the other way around, but that makes sense given that
closing an interactive proof will be a client of declare.
Indeed, all `Declare` / `Pfedit` / and `Proof_global` are tied into
tactics due to `abstract`, at some point we may be able to unify all
them into a single file in `vernac`.
|
|
This is more in-line with attributes and the rest of the API, and
makes some code significantly clearer (as in `foo true false false`,
etc...)
|
|
We split `{goal,declaration,assumption}_kind` into their
components. This makes sense as each part of this triple is handled by
a different layer, namely:
- `polymorphic` status: necessary for the lower engine layers;
- `locality`: only used in `vernac` top-level constants
- `kind`: merely used for cosmetic purposes [could indeed be removed /
pushed upwards]
We also profit from this refactoring to add some named parameters to
the top-level definition API which is quite parameter-hungry.
More refactoring is possible and will come in further commits, in
particular this is a step towards unifying the definition / lemma save path.
|
|
This allows to desynchronize the kernel-facing API from the proof-facing one.
|
|
|
|
|
|
|
|
We move the role data into the evarmap instead.
|
|
This lets us avoid passing ~ontop to do_definition and co, and after #10050
to even more functions.
|
|
This impacts a lot of code, apparently in the good, removing several
conversions back and forth constr.
|
|
We also update the plugin tutorial.
This was already tried [in the same exact way] in #8811, however the
bench time was not convincing then, but now things seem a bit better,
likely due to the removal of some extra normalization somewhere.
Some more changes from #8811 are still pending.
|
|
|
|
Supersedes #8718.
|
|
|
|
We remove all calls to `Flags.is_program_mode` except one (to compute
the default value of the attribute). Everything else is passed
explicitely, and we remove the special logic in the interpretation loop
to set/unset the flag.
This is especially important since the value of the flag has an impact on
proof modes, so on the separation of parsing and execution phases.
|
|
We make `declaration_hook`s optional arguments everywhere, and thus we
avoid some "fake" functions having to be passed.
This identifies positively the code really using hooks [funind,
rewrite, coercions, program, and canonicals] and helps moving toward
some hope of reification.
|
|
(same for solve_remaining_evars)
This is the standard way to use these functions, with 1 exception in
Unification.
|
|
terms
This is necessary for programs like Equations that call add_definition
and want to later update in their hook some separate datastructures
which refer to the obligations that are defined by Program. We give back
a map from obligation name to a constr defined in the program's universe
state which the hook returns as well. (Obligation names also correspond
to undefined evars in the original terms through
Obligations.eterm_obligations).
Using this, I can avoid ucontext_of_aucontext in Equations, allowing PR
#8601 to go through.
|
