aboutsummaryrefslogtreecommitdiff
path: root/interp/constrintern.ml
AgeCommit message (Collapse)Author
2012-12-04Backtrack on activating scopes with type casts (was r15978).herbelin
IT happens that there are some other uses of casts, e.g. in nat_int de MathClasses which uses a notation 0 for some class constructor zero, and a cast (0:nat) to force this notation 0 to represent the instance O in nat of the class of types having a zero (eventually, 0, i.e. "zero nat O" and "O" are convertible but the information of being a class instance is lost). git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16024 85f007b7-540e-0410-9357-904b9bb8a0f7
2012-11-25Monomorphization (interp)ppedrot
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15999 85f007b7-540e-0410-9357-904b9bb8a0f7
2012-11-25More equality functionsppedrot
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15998 85f007b7-540e-0410-9357-904b9bb8a0f7
2012-11-17Taking into account the type of a definition (if it exists), and theherbelin
type in "cast" to activate the temporary interpretation scope. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15978 85f007b7-540e-0410-9357-904b9bb8a0f7
2012-11-13Added a CString module.ppedrot
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15968 85f007b7-540e-0410-9357-904b9bb8a0f7
2012-11-08Monomorphized a lot of equalities over OCaml integers, thanks toppedrot
the new Int module. Only the most obvious were removed, so there are a lot more in the wild. This may sound heavyweight, but it has two advantages: 1. Monomorphization is explicit, hence we do not miss particular optimizations of equality when doing it carelessly with the generic equality. 2. When we have removed all the generic equalities on integers, we will be able to write something like "let (=) = ()" to retrieve all its other uses (mostly faulty) spread throughout the code, statically. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15957 85f007b7-540e-0410-9357-904b9bb8a0f7
2012-10-02Remove some more "open" and dead code thanks to OCaml4 warningsletouzey
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15844 85f007b7-540e-0410-9357-904b9bb8a0f7
2012-09-14As r15801: putting everything from Util.array_* to CArray.*.ppedrot
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15804 85f007b7-540e-0410-9357-904b9bb8a0f7
2012-09-14Moving Utils.list_* to a proper CList module, which includes stdlibppedrot
List module. That way, an "open Util" in the header permits using any function of CList in the List namespace (and in particular, this permits optimized reimplementations of the List functions, as, for example, tail-rec implementations. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15801 85f007b7-540e-0410-9357-904b9bb8a0f7
2012-09-14This patch removes unused "open" (automatically generated fromregisgia
compiler warnings). I was afraid that such a brutal refactoring breaks some obscure invariant about linking order and side-effects but the standard library still compiles. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15800 85f007b7-540e-0410-9357-904b9bb8a0f7
2012-08-08Updating headers.herbelin
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15715 85f007b7-540e-0410-9357-904b9bb8a0f7
2012-07-21Improving management of notations with binders (see #2708 where aherbelin
variable is used both as term and as binder, resulting in ununderstandable error message about scopes). git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15642 85f007b7-540e-0410-9357-904b9bb8a0f7
2012-06-22Added an indirection with respect to Loc in Compat. As many [open Compat]ppedrot
were closed (i.e. the only remaining ones are those of printing/parsing). Meanwhile, a simplified interface is provided in loc.mli. This also permits to put Pp in Clib, because it does not depend on CAMLP4/5 anymore. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15475 85f007b7-540e-0410-9357-904b9bb8a0f7
2012-06-14Internalization of pattern is done in two phases.pboutill
First Notations, syntactic definitions, primitive entries are tackled to build raw_cases_pattern_expr. Reference are revolved at this time too. Then raw_patterns are internalized as cases_pattern or applied inductive (dealing with implicit args, or_pattern refactoring, aliases). It is more uniform, it allows notations for non fully applied constructors but : - It does not raise a warning when an identifier is also a global_reference different than a constructor. - It looks for implicit arguments twice. (because finding scopes of arguments asks for implicit arguments). - It does not deal anymore with constants that evaluates to constructor. (This one is voluntary, dealing with implicit & notations is already a hell full of bugs so what will be next step ? Any terms that computes to a pattern ???) git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15439 85f007b7-540e-0410-9357-904b9bb8a0f7
2012-06-04Forward-port fixes from 8.4 (15358, 15353, 15333).msozeau
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15418 85f007b7-540e-0410-9357-904b9bb8a0f7
2012-05-29global_reference migrated from Libnames to new Globnames, less deps in ↵letouzey
grammar.cma git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15384 85f007b7-540e-0410-9357-904b9bb8a0f7
2012-05-29Basic stuff about constr_expr migrated from topconstr to constrexpr_opsletouzey
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15382 85f007b7-540e-0410-9357-904b9bb8a0f7
2012-05-29Stuff about notation_constr (ex-aconstr) now in notation_ops.mlletouzey
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15381 85f007b7-540e-0410-9357-904b9bb8a0f7
2012-05-29Pattern as a mli-only file, operations in Patternopsletouzey
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15376 85f007b7-540e-0410-9357-904b9bb8a0f7
2012-05-29New files intf/constrexpr.mli and intf/notation_term.mli out of Topconstrletouzey
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15375 85f007b7-540e-0410-9357-904b9bb8a0f7
2012-05-29Glob_term now mli-only, operations now in Glob_opsletouzey
Stuff about reductions now in genredexpr.mli, operations in redops.ml git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15374 85f007b7-540e-0410-9357-904b9bb8a0f7
2012-05-29locus.mli for occurrences+clauses, misctypes.mli for various little thingsletouzey
Corresponding operations in locusops.ml and miscops.ml The type of occurrences is now a clear algebraic one instead of a bool*list hard to understand. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15372 85f007b7-540e-0410-9357-904b9bb8a0f7
2012-05-29Evar_kinds.mli containing former Evd.hole_kind, avoid deps on Evdletouzey
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15371 85f007b7-540e-0410-9357-904b9bb8a0f7
2012-05-11Impossible branches inference fixup (bug 2761)pboutill
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15307 85f007b7-540e-0410-9357-904b9bb8a0f7
2012-04-12"A -> B" is a notation for "forall _ : A, B".pboutill
No good reason for that except uniformity so revert this commit if you find a reason against it. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15146 85f007b7-540e-0410-9357-904b9bb8a0f7
2012-04-07Fixing the "capture" check newly introduced in r15122 whenherbelin
internalizing declaration of contexts of the form "(x y : P x)": - don't forget to catch the error in intern_context; - check capture on glob_constr rather than constr_expr so that binders possibly introduced by notations are recognized as such; - give a more expressive error message. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15123 85f007b7-540e-0410-9357-904b9bb8a0f7
2012-04-06Unifying the different procedures for interning binders.herbelin
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15122 85f007b7-540e-0410-9357-904b9bb8a0f7
2012-04-06Fixing a few bugs (see #2571) related to interpretation of multiple bindersherbelin
- fixing missing spaces in the format of the exists' notations (Logic.v); - fixing wrong variable name in check_is_hole error message (topconstr.ml); - interpret expressions with open binders such as "forall x y, t" as "forall (x:_) (y:_),t" instead of "forall (x y:_),t" to avoid the "implicit type" of a variable being propagated to the type of another variable of different base name. An open question remains: when writing explicitly "forall (x y:_),t", should the types of x and y be the same or not. To avoid the "bug" that x and y have implicit types but the one of x takes precedences, I enforced the interpretation (in constrintern, not in parsing) that "forall (x y:_),t" means the same as "forall (x:_) (y:_),t". However, another choice could have been made. Then one would have to check that if x and y have implicit types, they are the same; also, glob_constr should ideally be changed to support a GProd and GLam with multiple names in the same type, especially if this type is an evar. On the contrary, one might also want e.g. "forall x y : list _, t" to mean "forall (x:list _) (y:list _), t" with distinct instanciations of "_" ...). git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15121 85f007b7-540e-0410-9357-904b9bb8a0f7
2012-03-20Fixing bug #2724 (using notations with binders in cases patternsherbelin
was provoking an anomaly instead of a regular error). git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15070 85f007b7-540e-0410-9357-904b9bb8a0f7
2012-03-14Second step of integration of Program:msozeau
- Remove useless functorization of Pretyping - Move Program coercion/cases code inside pretyping/, enabled according to a flag. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15033 85f007b7-540e-0410-9357-904b9bb8a0f7
2012-03-02Glob_term.predicate_pattern: No number of parameters with letins.pboutill
Detyping is wrong about it and as far as I understand no one but Constrextern uses it. Constrextern has now the same machinery for all patterns. Revert if I miss something. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15022 85f007b7-540e-0410-9357-904b9bb8a0f7
2012-03-02"Let in" in pattern hell, one more iterationpboutill
This reverts commit 28bcf05dd876beea8697f6ee47ebf916a8b94cdf. An other wrong externalize function git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15021 85f007b7-540e-0410-9357-904b9bb8a0f7
2012-03-02Noise for nothingpboutill
Util only depends on Ocaml stdlib and Utf8 tables. Generic pretty printing and loc functions are in Pp. Generic errors are in Errors. + Training white-spaces, useless open, prlist copies random erasure. Too many "open Errors" on the contrary. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15020 85f007b7-540e-0410-9357-904b9bb8a0f7
2012-02-29Fix compilation of Constrintern...pboutill
I love being push under presure to commit and do not try my fixup ! git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15003 85f007b7-540e-0410-9357-904b9bb8a0f7
2012-02-29RefMan update about match syntax.pboutill
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15002 85f007b7-540e-0410-9357-904b9bb8a0f7
2012-02-29In the syntax of pattern matching, the arguments of the inductive in the "in"pboutill
clause can be any pattern. It is expanded as a match in the return clause. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15001 85f007b7-540e-0410-9357-904b9bb8a0f7
2012-02-29In the syntax of pattern matching, "in" clauses are patterns.pboutill
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15000 85f007b7-540e-0410-9357-904b9bb8a0f7
2012-01-20Added documentation for "Set Parsing Explicit" + fixed mistakenlyherbelin
committed "assert" in commit r14928. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@14930 85f007b7-540e-0410-9357-904b9bb8a0f7
2012-01-20Added new command "Set Parsing Explicit" for deactivating parsing (andherbelin
printing) of implicit arguments (a priori useful for teaching). git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@14928 85f007b7-540e-0410-9357-904b9bb8a0f7
2012-01-19Boolean Option Patterns Compatibilitypboutill
switch between "all arguments no parameters" and "explicit params and args" for constructor arguments in patterns. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@14919 85f007b7-540e-0410-9357-904b9bb8a0f7
2012-01-16Parameters in pattern first step.pboutill
In interp/constrintern.ml, '_' for constructor parameter are required if you use AppExpl ('@Constr') and added in order to be erased at the last time if you do not use '@'. It makes the use of notation in pattern more permissive. For example, -8<------ Inductive I (A: Type) : Type := C : A -> I A. Notation "x <: T" := (C T x) (at level 38). Definition uncast A (x : I A) := match x with | x <: _ => x end. -8<------ was forbidden. Backward compatibility is strictly preserved expect for syntactic definitions: While defining "Notation SOME = @Some", SOME had a special treatment and used to be an alias of Some in pattern. It is now uniformly an alias of @Some ('_' must be provided for parameters). In interp/constrextern.ml, except if all the parameters are implicit and all the arguments explicit (This covers all the cases of the test-suite), pattern are generated with AppExpl (id est '@') (so with '_' for parameters) in order to become compatible in any case with future behavior. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@14909 85f007b7-540e-0410-9357-904b9bb8a0f7
2012-01-05Backtracking on r14876 (fix for bug #2267): extra scopes might beherbelin
useful in the presence of coercions to Funclass. Fixed the bug differently. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@14880 85f007b7-540e-0410-9357-904b9bb8a0f7
2012-01-04Fixing Arguments Scope bug when too many scopes are given (bug #2667).herbelin
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@14876 85f007b7-540e-0410-9357-904b9bb8a0f7
2011-12-17A pass on warning printings. Made systematic the use of msg_warning soherbelin
as to ensure the warning is flushed in real time. Made systematic the use of if_warn instead of if_verbose when the warning is intended to signal something anormal (if_warn is activated when compiling verbosely and when working interactively while if_verbose is activated only when working interactively and when loading verbosely). git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@14803 85f007b7-540e-0410-9357-904b9bb8a0f7
2011-11-16Fixing association of wrong "as"-pattern name when expandingherbelin
abbreviations of notations (most of the time inoffensive but has bad effect on compiling pattern-matching in the presence of dependencies). git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@14653 85f007b7-540e-0410-9357-904b9bb8a0f7
2011-10-28Remove dynamic stuff from constr_expr and glob_constrglondu
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@14621 85f007b7-540e-0410-9357-904b9bb8a0f7
2011-07-22Add a syntax entry for fully applied constructor patternpboutill
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@14292 85f007b7-540e-0410-9357-904b9bb8a0f7
2011-07-22Internalization of pattern carries a full intern_envpboutill
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@14291 85f007b7-540e-0410-9357-904b9bb8a0f7
2011-07-15A correct error message for an unknown field in a record definitionpboutill
fix bug 2571. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@14276 85f007b7-540e-0410-9357-904b9bb8a0f7
2011-05-26Partial fix for accepting bound variables with same name as implicitherbelin
parameters of inductive types when these variables cannot bind the conclusion of the inductive type (done for "return" predicates but still to be done for non strictly positive binding occurrence, as e.g. in "Set Implicit Arguments. Inductive I A:A->Prop:=C a:(forall A, A)->I a." which should morally be accepted but is not). git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@14159 85f007b7-540e-0410-9357-904b9bb8a0f7
generated by cgit v1.2.3 (git 2.25.1) at 2026-02-11 19:39:18 +0000