------------------------------------------------------------------------------ -- -- -- GNAT COMPILER COMPONENTS -- -- -- -- S E M _ E V A L -- -- -- -- S p e c -- -- -- -- Copyright (C) 1992-2018, Free Software Foundation, Inc. -- -- -- -- GNAT is free software; you can redistribute it and/or modify it under -- -- terms of the GNU General Public License as published by the Free Soft- -- -- ware Foundation; either version 3, or (at your option) any later ver- -- -- sion. GNAT is distributed in the hope that it will be useful, but WITH- -- -- OUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY -- -- or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License -- -- for more details. You should have received a copy of the GNU General -- -- Public License distributed with GNAT; see file COPYING3. If not, go to -- -- http://www.gnu.org/licenses for a complete copy of the license. -- -- -- -- GNAT was originally developed by the GNAT team at New York University. -- -- Extensive contributions were provided by Ada Core Technologies Inc. -- -- -- ------------------------------------------------------------------------------ -- This package contains various subprograms involved in compile time -- evaluation of expressions and checks for staticness of expressions and -- types. It also contains the circuitry for checking for violations of pure -- and preelaborated conditions (this naturally goes here, since these rules -- involve consideration of staticness). -- Note: the static evaluation for attributes is found in Sem_Attr even though -- logically it belongs here. We have done this so that it is easier to add -- new attributes to GNAT. with Types; use Types; with Uintp; use Uintp; with Urealp; use Urealp; package Sem_Eval is ------------------------------------ -- Handling of Static Expressions -- ------------------------------------ -- This package contains a set of routines that process individual -- subexpression nodes with the objective of folding (precomputing) the -- value of static expressions that are known at compile time and properly -- computing the setting of two flags that appear in every subexpression -- node: -- Is_Static_Expression -- This flag is set on any expression that is static according to the -- rules in (RM 4.9(3-32)). This flag should be tested during testing -- of legality of parts of a larger static expression. For all other -- contexts that require static expressions, use the separate predicate -- Is_OK_Static_Expression, since an expression that meets the RM 4.9 -- requirements, but raises a constraint error when evaluated in a non- -- static context does not meet the legality requirements. -- Raises_Constraint_Error -- This flag indicates that it is known at compile time that the -- evaluation of an expression raises constraint error. If the -- expression is static, and this flag is off, then it is also known at -- compile time that the expression does not raise constraint error -- (i.e. the flag is accurate for static expressions, and conservative -- for non-static expressions. -- If a static expression does not raise constraint error, then it will -- have the flag Raises_Constraint_Error flag False, and the expression -- must be computed at compile time, which means that it has the form of -- either a literal, or a constant that is itself (recursively) either a -- literal or a constant. -- The above rules must be followed exactly in order for legality checks to -- be accurate. For subexpressions that are not static according to the RM -- definition, they are sometimes folded anyway, but of course in this case -- Is_Static_Expression is not set. -- When we are analyzing and evaluating static expressions, we propagate -- both flags accurately. Usually if a subexpression raises a constraint -- error, then so will its parent expression, and Raise_Constraint_Error -- will be propagated to this parent. The exception is conditional cases -- like (True or else 1/0 = 0) which results in an expresion that has the -- Is_Static_Expression flag True, and Raises_Constraint_Error False. Even -- though 1/0 would raise an exception, the right operand is never actually -- executed, so the expression as a whole does not raise CE. -- For constructs in the language where static expressions are part of the -- required semantics, we need an expression that meets the 4.9 rules and -- does not raise CE. So nearly everywhere, callers should call function -- Is_OK_Static_Expression rather than Is_Static_Expression. -- Finally, the case of static predicates. These are applied only to entire -- expressions, not to subexpressions, so we do not have the case of having -- to propagate this information. We handle this case simply by resetting -- the Is_Static_Expression flag if a static predicate fails. Note that we -- can't use this simpler approach for the constraint error case because of -- the (True or else 1/0 = 0) example discussed above. ------------------------------- -- Compile-Time Known Values -- ------------------------------- -- For most legality checking purposes the flag Is_Static_Expression -- defined in Sinfo should be used. This package also provides a routine -- called Is_OK_Static_Expression which in addition of checking that an -- expression is static in the RM 4.9 sense, it checks that the expression -- does not raise constraint error. In fact for certain legality checks not -- only do we need to ascertain that the expression is static, but we must -- also ensure that it does not raise constraint error. -- Neither of Is_Static_Expression and Is_OK_Static_Expression should be -- used for compile time evaluation purposes. In fact certain expression -- whose value may be known at compile time are not static in the RM 4.9 -- sense. A typical example is: -- C : constant Integer := Record_Type'Size; -- The expression 'C' is not static in the technical RM sense, but for many -- simple record types, the size is in fact known at compile time. When we -- are trying to perform compile time constant folding (for instance for -- expressions like C + 1, Is_Static_Expression or Is_OK_Static_Expression -- are not the right functions to test if folding is possible. Instead, we -- use Compile_Time_Known_Value. All static expressions that do not raise -- constraint error (i.e. those for which Is_OK_Static_Expression is true) -- are known at compile time, but as shown by the above example, there may -- be cases of non-static expressions which are known at compile time. ----------------- -- Subprograms -- ----------------- procedure Check_Expression_Against_Static_Predicate (Expr : Node_Id; Typ : Entity_Id); -- Determine whether an arbitrary expression satisfies the static predicate -- of a type. The routine does nothing if Expr is not known at compile time -- or Typ lacks a static predicate, otherwise it may emit a warning if the -- expression is prohibited by the predicate. If the expression is a static -- expression and it fails a predicate that was not explicitly stated to be -- a dynamic predicate, then an additional warning is given, and the flag -- Is_Static_Expression is reset on Expr. procedure Check_Non_Static_Context (N : Node_Id); -- Deals with the special check required for a static expression that -- appears in a non-static context, i.e. is not part of a larger static -- expression (see RM 4.9(35)), i.e. the value of the expression must be -- within the base range of the base type of its expected type. A check is -- also made for expressions that are inside the base range, but outside -- the range of the expected subtype (this is a warning message rather than -- an illegality). -- -- Note: most cases of non-static context checks are handled within -- Sem_Eval itself, including all cases of expressions at the outer level -- (i.e. those that are not a subexpression). Currently the only outside -- customer for this procedure is Sem_Attr (because Eval_Attribute is -- there). There is also one special case arising from ranges (see body of -- Resolve_Range). -- -- Note: this procedure is also called by GNATprove on real literals -- that are not sub-expressions of static expressions, to convert them to -- machine numbers, as GNATprove cannot perform this conversion contrary -- to gigi. procedure Check_String_Literal_Length (N : Node_Id; Ttype : Entity_Id); -- N is either a string literal, or a constraint error node. In the latter -- case, the situation is already dealt with, and the call has no effect. -- In the former case, if the target type, Ttyp is constrained, then a -- check is made to see if the string literal is of appropriate length. type Compare_Result is (LT, LE, EQ, GT, GE, NE, Unknown); subtype Compare_GE is Compare_Result range EQ .. GE; subtype Compare_LE is Compare_Result range LT .. EQ; -- Result subtypes for Compile_Time_Compare subprograms function Compile_Time_Compare (L, R : Node_Id; Assume_Valid : Boolean) return Compare_Result; pragma Inline (Compile_Time_Compare); -- Given two expression nodes, finds out whether it can be determined at -- compile time how the runtime values will compare. An Unknown result -- means that the result of a comparison cannot be determined at compile -- time, otherwise the returned result indicates the known result of the -- comparison, given as tightly as possible (i.e. EQ or LT is preferred -- returned value to LE). If Assume_Valid is true, the result reflects -- the result of assuming that entities involved in the comparison have -- valid representations. If Assume_Valid is false, then the base type of -- any involved entity is used so that no assumption of validity is made. function Compile_Time_Compare (L, R : Node_Id; Diff : access Uint; Assume_Valid : Boolean; Rec : Boolean := False) return Compare_Result; -- This version of Compile_Time_Compare returns extra information if the -- result is GT or LT. In these cases, if the magnitude of the difference -- can be determined at compile time, this (positive) magnitude is returned -- in Diff.all. If the magnitude of the difference cannot be determined -- then Diff.all contains No_Uint on return. Rec is a parameter that is set -- True for a recursive call from within Compile_Time_Compare to avoid some -- infinite recursion cases. It should never be set by a client. function Compile_Time_Known_Bounds (T : Entity_Id) return Boolean; -- If T is an array whose index bounds are all known at compile time, then -- True is returned. If T is not an array type, or one or more of its index -- bounds is not known at compile time, then False is returned. function Compile_Time_Known_Value (Op : Node_Id) return Boolean; -- Returns true if Op is an expression not raising Constraint_Error whose -- value is known at compile time and for which a call to Expr_Value can -- be used to determine this value. This is always true if Op is a static -- expression, but can also be true for expressions which are technically -- non-static but which are in fact known at compile time. Some examples of -- such expressions are the static lower bound of a non-static range or the -- value of a constant object whose initial value is itself compile time -- known in the sense of this routine. Note that this routine is defended -- against unanalyzed expressions. Such expressions will not cause a -- blowup, they may cause pessimistic (i.e. False) results to be returned. -- In general we take a pessimistic view. False does not mean the value -- could not be known at compile time, but True means that absolutely -- definition it is known at compile time and it is safe to call -- Expr_Value[_XX] on the expression Op. -- -- Note that we don't define precisely the set of expressions that return -- True. Callers should not make any assumptions regarding the value that -- is returned for non-static expressions. Functional behavior should never -- be affected by whether a given non-static expression returns True or -- False when this function is called. In other words this is purely for -- efficiency optimization purposes. The code generated can often be more -- efficient with compile time known values, e.g. range analysis for the -- purpose of removing checks is more effective if we know precise bounds. function Compile_Time_Known_Value_Or_Aggr (Op : Node_Id) return Boolean; -- Similar to Compile_Time_Known_Value, but also returns True if the value -- is a compile-time-known aggregate, i.e. an aggregate all of whose -- constituent expressions are either compile-time-known values (based on -- calling Compile_Time_Known_Value) or compile-time-known aggregates. -- Note that the aggregate could still involve run-time checks that might -- fail (such as for subtype checks in component associations), but the -- evaluation of the expressions themselves will not raise an exception. function CRT_Safe_Compile_Time_Known_Value (Op : Node_Id) return Boolean; -- In the case of configurable run-times, there may be an issue calling -- Compile_Time_Known_Value with non-static expressions where the legality -- of the program is not well-defined. Consider this example: -- -- X := B ** C; -- -- Now if C is compile time known, and has the value 4, then inline code -- can be generated at compile time, instead of calling a run-time routine. -- That's fine in the normal case, but when we have a configurable run-time -- the run-time routine may not be available. This means that the program -- will be rejected if C is not known at compile time. We don't want the -- legality of a program to depend on how clever the implementation of this -- function is. If the run-time in use lacks the exponentiation routine, -- then what we say is that exponentiation is permitted if the exponent is -- officially static and has a value in the range 0 .. 4. -- -- In a case like this, we use CRT_Safe_Compile_Time_Known_Value to avoid -- this effect. This routine will return False for a non-static expression -- if we are in configurable run-time mode, even if the expression would -- normally be considered compile-time known. function Expr_Rep_Value (N : Node_Id) return Uint; -- This is identical to Expr_Value, except in the case of enumeration -- literals of types for which an enumeration representation clause has -- been given, in which case it returns the representation value rather -- than the pos value. This is the value that is needed for generating code -- sequences, while the Expr_Value value is appropriate for compile time -- constraint errors or getting the logical value. Note that this function -- does NOT concern itself with biased values, if the caller needs a -- properly biased value, the subtraction of the bias must be handled -- explicitly. function Expr_Value (N : Node_Id) return Uint; -- Returns the folded value of the expression N. This function is called in -- instances where it has already been determined that the expression is -- static or its value is compile time known (Compile_Time_Known_Value (N) -- returns True). This version is used for integer values, and enumeration -- or character literals. In the latter two cases, the value returned is -- the Pos value in the relevant enumeration type. It can also be used for -- fixed-point values, in which case it returns the corresponding integer -- value. It cannot be used for floating-point values. function Expr_Value_E (N : Node_Id) return Entity_Id; -- Returns the folded value of the expression. This function is called in -- instances where it has already been determined that the expression is -- static or its value known at compile time. This version is used for -- enumeration types and returns the corresponding enumeration literal. function Expr_Value_R (N : Node_Id) return Ureal; -- Returns the folded value of the expression. This function is called in -- instances where it has already been determined that the expression is -- static or its value known at compile time. This version is used for real -- values (including both the floating-point and fixed-point cases). In the -- case of a fixed-point type, the real value is returned (cf above version -- returning Uint). function Expr_Value_S (N : Node_Id) return Node_Id; -- Returns the folded value of the expression. This function is called -- in instances where it has already been determined that the expression -- is static or its value is known at compile time. This version is used -- for string types and returns the corresponding N_String_Literal node. procedure Eval_Actual (N : Node_Id); procedure Eval_Allocator (N : Node_Id); procedure Eval_Arithmetic_Op (N : Node_Id); procedure Eval_Call (N : Node_Id); procedure Eval_Case_Expression (N : Node_Id); procedure Eval_Character_Literal (N : Node_Id); procedure Eval_Concatenation (N : Node_Id); procedure Eval_Entity_Name (N : Node_Id); procedure Eval_If_Expression (N : Node_Id); procedure Eval_Indexed_Component (N : Node_Id); procedure Eval_Integer_Literal (N : Node_Id); procedure Eval_Logical_Op (N : Node_Id); procedure Eval_Membership_Op (N : Node_Id); procedure Eval_Named_Integer (N : Node_Id); procedure Eval_Named_Real (N : Node_Id); procedure Eval_Op_Expon (N : Node_Id); procedure Eval_Op_Not (N : Node_Id); procedure Eval_Real_Literal (N : Node_Id); procedure Eval_Relational_Op (N : Node_Id); procedure Eval_Shift (N : Node_Id); procedure Eval_Short_Circuit (N : Node_Id); procedure Eval_Slice (N : Node_Id); procedure Eval_String_Literal (N : Node_Id); procedure Eval_Qualified_Expression (N : Node_Id); procedure Eval_Type_Conversion (N : Node_Id); procedure Eval_Unary_Op (N : Node_Id); procedure Eval_Unchecked_Conversion (N : Node_Id); procedure Flag_Non_Static_Expr (Msg : String; Expr : Node_Id); -- This procedure is called after it has been determined that Expr is not -- static when it is required to be. Msg is the text of a message that -- explains the error. This procedure checks if an error is already posted -- on Expr, if so, it does nothing unless All_Errors_Mode is set in which -- case this flag is ignored. Otherwise the given message is posted using -- Error_Msg_F, and then Why_Not_Static is called on Expr to generate -- additional messages. The string given as Msg should end with ! to make -- it an unconditional message, to ensure that if it is posted, the entire -- set of messages is all posted. procedure Fold_Str (N : Node_Id; Val : String_Id; Static : Boolean); -- Rewrite N with a new N_String_Literal node as the result of the compile -- time evaluation of the node N. Val is the resulting string value from -- the folding operation. The Is_Static_Expression flag is set in the -- result node. The result is fully analyzed and resolved. Static indicates -- whether the result should be considered static or not (True = consider -- static). The point here is that normally all string literals are static, -- but if this was the result of some sequence of evaluation where values -- were known at compile time but not static, then the result is not -- static. The call has no effect if Raises_Constraint_Error (N) is True, -- since there is no point in folding if we have an error. procedure Fold_Uint (N : Node_Id; Val : Uint; Static : Boolean); -- Rewrite N with a (N_Integer_Literal, N_Identifier, N_Character_Literal) -- node as the result of the compile time evaluation of the node N. Val is -- the result in the integer case and is the position of the literal in the -- literals list for the enumeration case. Is_Static_Expression is set True -- in the result node. The result is fully analyzed/resolved. Static -- indicates whether the result should be considered static or not (True = -- consider static). The point here is that normally all integer literals -- are static, but if this was the result of some sequence of evaluation -- where values were known at compile time but not static, then the result -- is not static. The call has no effect if Raises_Constraint_Error (N) is -- True, since there is no point in folding if we have an error. procedure Fold_Ureal (N : Node_Id; Val : Ureal; Static : Boolean); -- Rewrite N with a new N_Real_Literal node as the result of the compile -- time evaluation of the node N. Val is the resulting real value from the -- folding operation. The Is_Static_Expression flag is set in the result -- node. The result is fully analyzed and result. Static indicates whether -- the result should be considered static or not (True = consider static). -- The point here is that normally all string literals are static, but if -- this was the result of some sequence of evaluation where values were -- known at compile time but not static, then the result is not static. -- The call has no effect if Raises_Constraint_Error (N) is True, since -- there is no point in folding if we have an error. function Is_In_Range (N : Node_Id; Typ : Entity_Id; Assume_Valid : Boolean := False; Fixed_Int : Boolean := False; Int_Real : Boolean := False) return Boolean; -- Returns True if it can be guaranteed at compile time that expression -- N is known to be in range of the subtype Typ. A result of False does -- not mean that the expression is out of range, merely that it cannot be -- determined at compile time that it is in range. If Typ is a floating -- point type or Int_Real is set, any integer value is treated as though it -- was a real value (i.e. the underlying real value is used). In this case -- we use the corresponding real value, both for the bounds of Typ, and for -- the value of the expression N. If Typ is a fixed type or a discrete type -- and Int_Real is False but flag Fixed_Int is True then any fixed-point -- value is treated as though it was discrete value (i.e. the underlying -- integer value is used). In this case we use the corresponding integer -- value, both for the bounds of Typ, and for the value of the expression -- N. If Typ is a discrete type and Fixed_Int as well as Int_Real are -- false, integer values are used throughout. -- -- If Assume_Valid is set True, then N is always assumed to contain a valid -- value. If Assume_Valid is set False, then N may be invalid (unless there -- is some independent way of knowing that it is valid, i.e. either it is -- an entity with Is_Known_Valid set, or Assume_No_Invalid_Values is True. function Is_Null_Range (Lo : Node_Id; Hi : Node_Id) return Boolean; -- Returns True if it can guarantee that Lo .. Hi is a null range. If it -- cannot (because the value of Lo or Hi is not known at compile time) then -- it returns False. function Is_OK_Static_Expression (N : Node_Id) return Boolean; -- An OK static expression is one that is static in the RM definition sense -- and which does not raise constraint error. For most legality checking -- purposes you should use Is_Static_Expression. For those legality checks -- where the expression N should not raise constraint error use this -- routine. This routine is *not* to be used in contexts where the test is -- for compile time evaluation purposes. Use Compile_Time_Known_Value -- instead (see section on "Compile-Time Known Values" above). function Is_OK_Static_Range (N : Node_Id) return Boolean; -- Determines if range is static, as defined in RM 4.9(26), and also checks -- that neither bound of the range raises constraint error, thus ensuring -- that both bounds of the range are compile-time evaluable (i.e. do not -- raise constraint error). A result of true means that the bounds are -- compile time evaluable. A result of false means they are not (either -- because the range is not static, or because one or the other bound -- raises CE). function Is_OK_Static_Subtype (Typ : Entity_Id) return Boolean; -- Determines whether a subtype fits the definition of an Ada static -- subtype as given in (RM 4.9(26)) with the additional check that neither -- bound raises constraint error (meaning that Expr_Value[_R|S] can be used -- on these bounds). -- -- This differs from Is_Static_Subtype in that it includes the constraint -- error checks, which are missing from Is_Static_Subtype. function Is_Out_Of_Range (N : Node_Id; Typ : Entity_Id; Assume_Valid : Boolean := False; Fixed_Int : Boolean := False; Int_Real : Boolean := False) return Boolean; -- Returns True if it can be guaranteed at compile time that expression is -- known to be out of range of the subtype Typ. True is returned if Typ is -- a scalar type, and the value of N can be determined to be outside the -- range of Typ. A result of False does not mean that the expression is in -- range, but rather merely that it cannot be determined at compile time -- that it is out of range. The parameters Assume_Valid, Fixed_Int, and -- Int_Real are as described for Is_In_Range above. function Is_Static_Subtype (Typ : Entity_Id) return Boolean; -- Determines whether a subtype fits the definition of an Ada static -- subtype as given in (RM 4.9(26)). -- -- This differs from Is_OK_Static_Subtype (which is what must be used by -- clients) in that it does not care whether the bounds raise a constraint -- error exception or not. Used for checking whether expressions are static -- in the 4.9 sense (without worrying about exceptions). function Is_Statically_Unevaluated (Expr : Node_Id) return Boolean; -- This function returns True if the given expression Expr is statically -- unevaluated, as defined in (RM 4.9 (32.1-32.6)). function In_Subrange_Of (T1 : Entity_Id; T2 : Entity_Id; Fixed_Int : Boolean := False) return Boolean; -- Returns True if it can be guaranteed at compile time that the range of -- values for scalar type T1 are always in the range of scalar type T2. A -- result of False does not mean that T1 is not in T2's subrange, only that -- it cannot be determined at compile time. Flag Fixed_Int is used as in -- routine Is_In_Range above. function Not_Null_Range (Lo : Node_Id; Hi : Node_Id) return Boolean; -- Returns True if it can guarantee that Lo .. Hi is not a null range. If -- it cannot (because the value of Lo or Hi is not known at compile time) -- then it returns False. function Predicates_Match (T1, T2 : Entity_Id) return Boolean; -- In Ada 2012, subtypes statically match if their static predicates -- match as well. This function performs the required check that -- predicates match. Separated out from Subtypes_Statically_Match so -- that it can be used in specializing error messages. function Subtypes_Statically_Compatible (T1 : Entity_Id; T2 : Entity_Id; Formal_Derived_Matching : Boolean := False) return Boolean; -- Returns true if the subtypes are unconstrained or the constraint on -- on T1 is statically compatible with T2 (as defined by 4.9.1(4)). -- Otherwise returns false. Formal_Derived_Matching indicates whether -- the type T1 is a generic actual being checked against ancestor T2 -- in a formal derived type association. function Subtypes_Statically_Match (T1 : Entity_Id; T2 : Entity_Id; Formal_Derived_Matching : Boolean := False) return Boolean; -- Determine whether two types T1, T2, which have the same base type, -- are statically matching subtypes (RM 4.9.1(1-2)). Also includes the -- extra GNAT rule that object sizes must match (this can be false for -- types that match in the RM sense because of use of 'Object_Size), -- except when testing a generic actual T1 against an ancestor T2 in a -- formal derived type association (indicated by Formal_Derived_Matching). procedure Test_Comparison (Op : Node_Id; Assume_Valid : Boolean; True_Result : out Boolean; False_Result : out Boolean); -- Determine the outcome of evaluating comparison operator Op using routine -- Compile_Time_Compare. Assume_Valid should be set when the operands are -- to be assumed valid. Flags True_Result and False_Result are set when the -- comparison evaluates to True or False respectively. procedure Why_Not_Static (Expr : Node_Id); -- This procedure may be called after generating an error message that -- complains that something is non-static. If it finds good reasons, it -- generates one or more error messages pointing the appropriate offending -- component of the expression. If no good reasons can be figured out, then -- no messages are generated. The expectation here is that the caller has -- already issued a message complaining that the expression is non-static. -- Note that this message should be placed using Error_Msg_F or -- Error_Msg_FE, so that it will sort before any messages placed by this -- call. Note that it is fine to call Why_Not_Static with something that -- is not an expression, and usually this has no effect, but in some cases -- (N_Parameter_Association or N_Range), it makes sense for the internal -- recursive calls. -- -- Note that these messages are not continuation messages, instead they are -- separate unconditional messages, marked with '!'. The reason for this is -- that they can be posted at a different location from the main message as -- documented above ("appropriate offending component"), and continuation -- messages must always point to the same location as the parent message. procedure Initialize; -- Initializes the internal data structures. Must be called before each -- separate main program unit (e.g. in a GNSA/ASIS context). private -- The Eval routines are all marked inline, since they are called once pragma Inline (Eval_Actual); pragma Inline (Eval_Allocator); pragma Inline (Eval_Character_Literal); pragma Inline (Eval_If_Expression); pragma Inline (Eval_Indexed_Component); pragma Inline (Eval_Named_Integer); pragma Inline (Eval_Named_Real); pragma Inline (Eval_Real_Literal); pragma Inline (Eval_Shift); pragma Inline (Eval_Slice); pragma Inline (Eval_String_Literal); pragma Inline (Eval_Unchecked_Conversion); pragma Inline (Is_OK_Static_Expression); end Sem_Eval;