| Age | Commit message (Collapse) | Author |
|---|
|
![proof_stack] is equivalent to the old meaning of ![proof]: the body
has type `pstate:Proof_global.t option -> Proof_global.t option`
The other specifiers are for the following body types:
~~~
![open_proof] `is_ontop:bool -> pstate`
![maybe_open_proof] `is_ontop:bool -> pstate option`
![proof] `pstate:pstate -> pstate`
![proof_opt_query] `pstate:pstate option -> unit`
![proof_query] `pstate:pstate -> unit`
~~~
The `is_ontop` is only used for the warning message when declaring a
section variable inside a proof, we could also just stop warning.
The specifiers look closely related to stm classifiers, but currently
they're unconnected. Notably this means that a ![proof_query] doesn't
have to be classified QUERY.
![proof_stack] is only used by g_rewrite/rewrite whose behaviour I
don't fully understand, maybe we can drop it in the future.
For compat we may want to consider keeping ![proof] with its old
meaning and using some new name for the new meaning. OTOH fixing
plugins to be stricter is easier if we change it as the errors tell us
where it's used.
|
|
Typically instead of [start_proof : ontop:Proof_global.t option -> bla ->
Proof_global.t] we have [start_proof : bla -> Proof_global.pstate] and
the pstate is pushed on the stack by a caller around the
vernacentries/mlg level.
Naming can be a bit awkward, hopefully it can be improved (maybe in a
followup PR).
We can see some patterns appear waiting for nicer combinators, eg in
mlg we often only want to work with the current proof, not the stack.
Behaviour should be similar modulo bugs, let's see what CI says.
|
|
|
|
We never use this id in rewrite.ml so don't bother threading it around.
|
|
|
|
The only lawbreaker was the Add Ring command. We generate a type for
the declaration to fix the code.
|
|
Reviewed-by: gares
Reviewed-by: ppedrot
|
|
To prevent confusion, forbidding a mix of the "injection term as pat1
... patn" and of the "injection term as [= pat1 ... patn]" syntax: If
a "[= ...]" occurs, this should be a singleton list of patterns.
|
|
Reviewed-by: Zimmi48
|
|
Since their introduction, these notations were incorrectly using the
proof-local environment.
|
|
Using pstate makes no sense for printing global stuff
|
|
We simply pass them as arguments, now that they are not called by the
kernel anymore.
The checker definitely needs to access the opaque proofs. In order not to
touch the API at all, I added a hook there, but it could also be provided
as an additional argument, at the cost of changing all the upwards callers.
|
|
|
|
|
|
Ack-by: SkySkimmer
Ack-by: Zimmi48
Reviewed-by: vbgl
|
|
|
|
Typeclasses resolution is not used anymore for lia.
Typeclasses resolution is still used by lra but only to access a
database of declared constants.
|
|
This lets us avoid passing ~ontop to do_definition and co, and after #10050
to even more functions.
|
|
It's used a few times in the stdlib (a couple of which need no other
change when removing the !) and not at all throughout our CI.
Considering that I think it's fair enough to remove it.
|
|
Reviewed-by: SkySkimmer
Reviewed-by: gares
Reviewed-by: maximedenes
|
|
This clean-up removes the dependency of the current proof mode (and hence
the parsing state) on unification.
The current proof mode can now be known simply by parsing and elaborating
attributes. We give access to attributes from the classifier for this purpose.
We remove the infamous `VtUnknown` code path in the STM which is known to
be buggy.
Fixes #3632 #3890 #4638.
|
|
|
|
|
|
Ack-by: SkySkimmer
Reviewed-by: gares
Ack-by: herbelin
Reviewed-by: maximedenes
Reviewed-by: ppedrot
|
|
|
|
|
|
This impacts a lot of code, apparently in the good, removing several
conversions back and forth constr.
|
|
Reviewed-by: SkySkimmer
|
|
Reviewed-by: herbelin
|
|
|
|
The current situation is a mess, some functions set it by default, but other
no. Making it mandatory ensures that the expected value is the correct one.
|
|
Some of them are significant so presumably it will take a bit of
effort to fix overlays.
I left out the removal of `nf_enter` for now as MTac2 needs some
serious porting in order to avoid it.
|
|
Before this patch, `x` was "simplified" to `x / 1`.
|
|
Ack-by: SkySkimmer
Ack-by: ejgallego
Reviewed-by: ppedrot
|
|
It was fairly easy, the plugin defined an argument that was only used in
a vernacular extension. Thus marking it as VERNAC was enough not to link
to Ltac.
|
|
Ack-by: Zimmi48
Ack-by: herbelin
Reviewed-by: ppedrot
|
|
|
|
** Changed definition of `simpl_rel` to `T -> `simpl_pred T`, so that
`inE` will now expand `a \in r b`, when `r := [rel x y | R]` to `R{b/x,
a/y}`, as the expanding coercion is now only inserted in the _last_
application.
The old definition made it possible to have a `simpl_rel >-> rel`
coercion that does not block expansion, but this can now be achieved
more economically with the `Arguments … /.` annotation.
** Deleted the `[rel of P]` notation which is no longer needed with
the new `simpl_rel` definition, and was broken anyway.
** Added `relpre f R` definition of functional preimage of a notation.
** `comp` and `idfun` are now proper definitions, using the `Arguments
… /.` annotation to specify simplification on application.
** Added `{pred T}` syntax for the alias of `pred T` in the `pred_sort`
coercion class; deleted the `pred_class` alias: one should either
use `pred_sort` in `Coercion` declarations, or `{pred T}` in type casts.
Used `{pred T}` as appropriate in localised predicate (`{in …, …}`) theory.
Extended and corrected `pred` coercion internal documentation.
** Simplified the `predType` structure by removing the redundant
explicit `mem_pred` subfield, and replacing it with an interlocked
projection; deleted `mkPredType`, now replaced by `PredType`.
** Added (and extensively documented) a `nonPropType` interface
matching types that do _not_ have sort `Prop`, and used it to remove
the non-standard maximal implicits annotation on `Some_inj` introduced
in #6911 by @anton-trumov; included `test-suite` entry for `nonPropType`.
** Documented the design of the four structures used to control the
matching of `inE` and related predicate rewriting lemmas; added `test-suite`
entry covering the `pred` rewriting control idioms.
** Used `only printing` annotations to get rid of token concatenation
hacks.
** Fixed boolean and general `if b return t then …` notation so that
`b` is bound in `t`. This is a minor source of incompatibility for
misuses of this syntax when `b` is _not_ bound in `t`, and `(if b then
…) : t` should have been used instead.
** Reserved all `ssreflect`, `ssrfun` and `ssrbool` notation at the top
of the file, adding some printing boxes, and removing some spurious
`[pred .. => ..]` reserved notation.
** Fixed parsing precedence and format of `<hidden n>` notation, and
declared and put it in an explicit `ssr_scope`.
** Used module-and-functor idiom to ensure that the `simpl_pred T >-
pred T` _and_ `simpl_pred T >-> {pred T}` coercions are realised by the
_same_ Gallina constant.
** Updated `CREDITS`.
The policy implied by this PR: that `{pred T}` should systematically
be used as the generic collective predicate type, was implemented in MathComp
math-comp/math-comp#237. As a result `simpl_pred >-> pred_sort` coercions
became more frequent, as it turned out they were not, as incorrectly stated
in `ssrbool` internal comments, impossible: while the `simplPredType`
canonical instance does solve all `simpl_pred T =~= pred_sort ?pT`
instances, it does _not_ solve `simpl_pred T =~= {pred T}`, and so the
coercion will be used in that case. However it appeared that having two
different coercion constants confused the SSReflect keyed matching
heuristic, hence the fix introduced here. This has entailed some
rearrangement of `ssrbool`: the large `Predicates` section had to be
broken up as the module-functor idiom for aliasing coercions cannot be
used inside a section.
|
|
|
|
Reviewed-by: gares
Ack-by: herbelin
|
|
Reviewed-by: CohenCyril
Ack-by: Zimmi48
Ack-by: erikmd
Ack-by: gares
Ack-by: jfehrle
|
|
We have some discrepancy with the package names for plugins in the
META / Dune case. This PR fixes it.
At some point there was a need for plugin package names having to be
named as their directories.
I think this is not true anymore, but taking the "dir_name ==
package_name" convention just in case.
|
|
This has been a mess for quite a while, we try to improve it.
|
|
See also PR math-comp/math-comp#73
|
|
|
|
So if the underlying tactic "contains a ;" one should actually write:
under eq_bigl => i do [rewrite andb_idl; first by move/eqP->].
|
|
so it acts "more naturally" like (under eq_bigl; [hnf|]); move=> [*|].
Also: replace "by over." in the doc example with "over."
|
|
as suggested by @gares, and:
* Rename some Under_* terms for better uniformity;
* Update & Improve minor details in the documentation.
|
|
* Add tests accordingly.
|
|
|
