Codebase list coq / debian/8.4pl2dfsg-2 proofs / goal.mli
debian/8.4pl2dfsg-2

Tree @debian/8.4pl2dfsg-2 (Download .tar.gz)

goal.mli @debian/8.4pl2dfsg-2raw · history · blame

(************************************************************************)
(*  v      *   The Coq Proof Assistant  /  The Coq Development Team     *)
(* <O___,, *   INRIA - CNRS - LIX - LRI - PPS - Copyright 1999-2012     *)
(*   \VV/  **************************************************************)
(*    //   *      This file is distributed under the terms of the       *)
(*         *       GNU Lesser General Public License Version 2.1        *)
(************************************************************************)

(* This module implements the abstract interface to goals *)

type goal

(* spiwack: this primitive is not extremely brilliant. It may be a good
    idea to define goals and proof views in the same file to avoid this
    sort of communication pipes. But I find it heavy. *)
val build : Evd.evar -> goal 

(* Gives a unique identifier to each goal. The identifier is
   guaranteed to contain no space. *)
val uid : goal -> string
(* Returns the goal (even if it has been partially solved)
   corresponding to a unique identifier obtained by {!uid}. *)
val get_by_uid : string -> goal

(* Debugging help *)
val pr_goal : goal -> Pp.std_ppcmds

(* [advance sigma g] returns [Some g'] if [g'] is undefined and 
    is the current avatar of [g] (for instance [g] was changed by [clear]
    into [g']). It returns [None] if [g] has been (partially) solved. *)
open Store.Field
val advance : Evd.evar_map -> goal -> goal option


(*** Goal tactics ***)


(* Goal tactics are [subgoal sensitive]-s *)
type subgoals = private { subgoals: goal list }

(* Goal sensitive values *)
type +'a sensitive

(* evaluates a goal sensitive value in a given goal (knowing the current evar_map). *)
val eval : 'a sensitive -> Environ.env -> Evd.evar_map -> goal -> 'a * Evd.evar_map

(* monadic bind on sensitive expressions *)
val bind : 'a sensitive -> ('a -> 'b sensitive) -> 'b sensitive

(* monadic return on sensitive expressions *)
val return : 'a -> 'a sensitive


(* interpretation of "open" constr *)
(* spiwack: it is a wrapper around [Constrintern.interp_open_constr]. 
    In an ideal world, this could/should be the other way round.
    As of now, though, it seems at least quite useful to build tactics. *)
val interp_constr : Topconstr.constr_expr -> Term.constr sensitive

(* Type of constr with holes used by refine. *)
type refinable

module Refinable : sig
  type t = refinable 
  type handle

  val make : (handle -> Term.constr sensitive) -> refinable sensitive
  val make_with : (handle -> (Term.constr*'a) sensitive) -> (refinable*'a) sensitive

  val mkEvar : handle -> Environ.env -> Term.types -> Term.constr sensitive

  (* [with_type c typ] constrains term [c] to have type [typ].  *)
  val with_type : Term.constr -> Term.types -> Term.constr sensitive

  val resolve_typeclasses : ?filter:(Evd.hole_kind -> bool) -> ?split:bool -> ?fail:bool -> unit -> unit sensitive


  (* [constr_of_raw h check_type resolve_classes] is a pretyping function.
      The [check_type] argument asks whether the term should have the same
      type as the conclusion. [resolve_classes] is a flag on pretyping functions
      which, if set to true, calls the typeclass resolver.
      The principal argument is a [glob_constr] which is then pretyped in the
      context of a term, the remaining evars are registered to the handle.
      It is the main component of the toplevel refine tactic.*)
  val constr_of_raw : 
    handle -> bool -> bool -> Glob_term.glob_constr -> Term.constr sensitive

  (* [constr_of_open_constr h check_type] transforms an open constr into a 
     goal-sensitive constr, adding the undefined variables to the set of subgoals.
     If [check_type] is true, the term is coerced to the conclusion of the goal.
     It allows to do refinement with already-built terms with holes.
  *)
  val constr_of_open_constr : handle -> bool -> Evd.open_constr -> Term.constr sensitive

end

(* [refine t] takes a refinable term and use it as a partial proof for current
    goal. *)
val refine : refinable -> subgoals sensitive


(*** Cleaning  goals ***)

(* Implements the [clear] tactic *)
val clear : Names.identifier list -> subgoals sensitive

(* Implements the [clearbody] tactic *)
val clear_body : Names.identifier list -> subgoals sensitive


(*** Conversion in goals ***)

(* Changes an hypothesis of the goal with a convertible type and body.
   Checks convertibility if the boolean argument is true. *)
val convert_hyp : bool -> Term.named_declaration -> subgoals sensitive

(* Changes the conclusion of the goal with a convertible type and body.
   Checks convertibility if the boolean argument is true. *)
val convert_concl : bool -> Term.constr -> subgoals sensitive

(*** Bureaucracy in hypotheses ***)

(* Renames a hypothesis. *)
val rename_hyp : Names.identifier -> Names.identifier -> subgoals sensitive

(*** Sensitive primitives ***)

(* [concl] is the conclusion of the current goal *)
val concl : Term.constr sensitive

(* [hyps] is the [named_context_val] representing the hypotheses
   of the current goal *)
val hyps : Environ.named_context_val sensitive

(* [env] is the current [Environ.env] containing both the 
   environment in which the proof is ran, and the goal hypotheses *)
val env : Environ.env sensitive

(* [defs] is the [Evd.evar_map] at the current evaluation point *)
val defs : Evd.evar_map sensitive

(* These four functions serve as foundation for the goal sensitive part
    of the tactic monad (see Proofview).
    [here] is a special sort of [return]: [here g a] is the value [a], but
    does not have any value (it raises an exception) if evaluated in
    any other goal than [g].
    [here_list] is the same, except with a list of goals rather than a single one.
    [plus a b] is the same as [a] if [a] is defined in the current goal, otherwise
    it is [b]. Effectively it's defined in the goals where [a] and [b] are defined.
    [null] is defined in no goal. (it is a neutral element for [plus]). *)
(* spiwack: these primitives are a bit hackish, but I couldn't find another way
    to pass information between goals, like for an intro tactic which gives to
    each goal the name of the variable it introduce.
    In pratice, in my experience, the primitives given in Proofview (in terms of
    [here] and [plus]) are sufficient to define any tactics, hence these might
    be another example of communication primitives between Goal and Proofview.
    Still, I can't see a way to prevent using the Proofview primitive to read
    a goal sensitive value out of its valid context. *)
val null : 'a sensitive

val plus : 'a sensitive -> 'a sensitive -> 'a sensitive 

val here : goal -> 'a -> 'a sensitive 

val here_list : goal list -> 'a -> 'a sensitive 

(*** Additional functions ***)

(* emulates List.map for functions of type 
   [Evd.evar_map -> 'a -> 'b * Evd.evar_map] on lists of type 'a, by propagating
   new evar_map to next definition *)
val list_map : (Evd.evar_map -> 'a -> 'b * Evd.evar_map) -> 
                        'a list -> 
                        Evd.evar_map -> 
                        'b list *Evd.evar_map

(* Layer to implement v8.2 tactic engine ontop of the new architecture. 
   Types are different from what they used to be due to a change of the
   internal types. *)
module V82 : sig

  (* Old style env primitive *)
  val env : Evd.evar_map -> goal -> Environ.env

  (* For printing *)
  val unfiltered_env : Evd.evar_map -> goal -> Environ.env

  (* Old style hyps primitive *)
  val hyps : Evd.evar_map -> goal -> Environ.named_context_val

  (* Access to ".evar_concl" *)
  val concl : Evd.evar_map -> goal -> Term.constr

  (* Access to ".evar_extra" *)
  val extra : Evd.evar_map -> goal -> Store.t

  (* Old style filtered_context primitive *)
  val filtered_context : Evd.evar_map -> goal -> Sign.named_context

  (* Old style mk_goal primitive, returns a new goal with corresponding 
       hypotheses and conclusion, together with a term which is precisely
       the evar corresponding to the goal, and an updated evar_map. *)
  val mk_goal : Evd.evar_map -> 
                         Environ.named_context_val ->
                         Term.constr ->
                         Store.t ->
                         goal * Term.constr * Evd.evar_map

  (* Equality function on goals *)
  val equal : Evd.evar_map -> goal -> goal -> bool

  (* Creates a dummy [goal sigma] for use in auto *)
  val dummy_goal : goal Evd.sigma

  (* Makes a goal out of an evar *)
  (* spiwack: used by [Proofview.init], not entirely clean probably, but it is
      the only way I could think of to preserve compatibility with previous Coq
      stuff. *)
  val build : Evd.evar -> goal 


  (* Instantiates a goal with an open term *)
  val partial_solution : Evd.evar_map -> goal -> Term.constr -> Evd.evar_map
   
  (* Principal part of the weak-progress tactical *)
  val weak_progress : goal list Evd.sigma -> goal Evd.sigma -> bool
   
  (* Principal part of the progress tactical *)
  val progress : goal list Evd.sigma -> goal Evd.sigma -> bool
    
 (* Principal part of tclNOTSAMEGOAL *)
  val same_goal : Evd.evar_map -> goal -> Evd.evar_map -> goal -> bool

 (* Used for congruence closure *)
  val new_goal_with : Evd.evar_map -> goal -> Sign.named_context -> goal Evd.sigma

  (* Used by the compatibility layer and typeclasses *)
  val nf_evar : Evd.evar_map -> goal -> goal * Evd.evar_map

  (* Goal represented as a type, doesn't take into account section variables *)
  val abstract_type : Evd.evar_map -> goal -> Term.types 

end