diff options
Diffstat (limited to 'src/vhdl/sem_names.ads')
| -rw-r--r-- | src/vhdl/sem_names.ads | 159 |
1 files changed, 159 insertions, 0 deletions
diff --git a/src/vhdl/sem_names.ads b/src/vhdl/sem_names.ads new file mode 100644 index 0000000..3bc8530 --- /dev/null +++ b/src/vhdl/sem_names.ads @@ -0,0 +1,159 @@ +-- Semantic analysis. +-- Copyright (C) 2002, 2003, 2004, 2005 Tristan Gingold +-- +-- GHDL is free software; you can redistribute it and/or modify it under +-- the terms of the GNU General Public License as published by the Free +-- Software Foundation; either version 2, or (at your option) any later +-- version. +-- +-- GHDL is distributed in the hope that it will be useful, but WITHOUT 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 +-- along with GHDL; see the file COPYING. If not, write to the Free +-- Software Foundation, 59 Temple Place - Suite 330, Boston, MA +-- 02111-1307, USA. +with Iirs; use Iirs; + +package Sem_Names is + -- In VHDL, most of name notations are ambiguous: + -- P.N is either + -- an expanded name or + -- a selected name for an element (with a possible implicit dereference) + -- P (A1, A2, ...) can be + -- an indexed name (with a possible implicit dereference) + -- a slice name (with a possible implicit dereference) + -- a subprogram call + -- a type conversion + + -- The name analysis resolves two ambiguities: notation and overload. + -- In a first pass, all possible meaning are collected as an overload + -- list in the Named_Entity field of the name. Prefixes in that list + -- are always declarations and not simple or expanded names. This is done + -- to avoid creating nodes for simple or expanded names, as they cannot be + -- shared in the prefixes because they can have several meanings. + -- + -- In a second pass, when the caller has resolved the overloading (using + -- the context), the name is rewritten: parenthesis and selected names are + -- replaced (by slice, index, call, element selection...). Prefixes are + -- simple or expanded names (and never declarations). Checks are also + -- performed on the result (pure, all sensitized). + -- + -- The result of the name analysis may not be a name: a function_call or + -- a type conversion are not names. + + -- Analyze NAME: perform the first pass only. In case of error, a message + -- is displayed and the named entity is error_mark. + procedure Sem_Name (Name : Iir; Keep_Alias : Boolean := False); + + -- Finish semantisation of NAME, if necessary. The named entity must not + -- be an overload list (ie the overload resolution must have been done). + -- This make remaining checks, transforms function names into calls... + function Finish_Sem_Name (Name : Iir) return Iir; + + -- Analyze NAME as a type mark. NAME must be either a simple name or an + -- expanded name, and the denoted entity must be either a type or a subtype + -- declaration. Return the name (possibly modified) and set named_entity + -- and type. In case of error, the type is error_mark. NAME may have + -- already been analyzed by Sem_Name. + -- Incomplete types are allowed only if INCOMPLETE is True. + function Sem_Type_Mark (Name : Iir; Incomplete : Boolean := False) + return Iir; + + -- Same as Sem_Name but without any side-effect: + -- * do not report error + -- * do not set xrefs + -- Currently, only simple names (and expanded names) are handled. + -- This is to be used during sem of associations. Because there is no side + -- effect, NAME is not modified. + procedure Sem_Name_Soft (Name : Iir); + + -- Remove every named_entity of NAME. + -- If NAME is Null_Iir then this is no op. + -- To be used only for names (weakly) semantized by sem_name_soft. + procedure Sem_Name_Clean (Name : Iir); + + -- Return TRUE if NAME is a name that designate an object (ie a constant, + -- a variable, a signal or a file). + function Is_Object_Name (Name : Iir) return Boolean; + + -- Return an object node if NAME designates an object (ie either is an + -- object or a name for an object). + -- Otherwise, returns NULL_IIR. + function Name_To_Object (Name : Iir) return Iir; + + -- If NAME is a selected name whose prefix is a protected variable, set + -- method_object of CALL. + procedure Name_To_Method_Object (Call : Iir; Name : Iir); + + -- Convert name NAME to an expression (ie, can create function call). + -- A_TYPE is the expected type of the expression. + -- FIXME: it is unclear wether the result must be an expression or not + -- (ie, it *must* have a type, but may be a range). + function Name_To_Expression (Name : Iir; A_Type : Iir) return Iir; + + -- Finish analyze of NAME and expect a range (either a type or subtype + -- declaration or a range attribute). Return Error_Mark in case of error. + function Name_To_Range (Name : Iir) return Iir; + + -- Return true if AN_IIR is an overload list. + function Is_Overload_List (An_Iir: Iir) return Boolean; + pragma Inline (Is_Overload_List); + + -- Create an overload list, that must be destroyed by Destroy_Overload_List. + function Get_Overload_List return Iir_Overload_List; + pragma Inline (Get_Overload_List); + + function Create_Overload_List (List : Iir_List) return Iir_Overload_List; + pragma Inline (Create_Overload_List); + + -- Free the list node (and the list itself). + procedure Free_Overload_List (N : in out Iir_Overload_List); + + -- Display an error message if the overload resolution for EXPR find more + -- than one interpretation. + procedure Error_Overload (Expr: Iir); + + -- Disp the overload list LIST. + procedure Disp_Overload_List (List : Iir_List; Loc : Iir); + + -- Convert a list to either Null_Iir, an element or an overload list. + function Simplify_Overload_List (List : Iir_List) return Iir; + + -- Add new interpretation DECL to RES. + -- Create an overload_list if necessary. + -- Before the first call, RES should be set to NULL_IIR. + procedure Add_Result (Res : in out Iir; Decl : Iir); + + -- Free a Parenthesis_Name. This is a special case as in general the + -- Association_Chain field must be freed too. + procedure Free_Parenthesis_Name (Name : Iir; Res : Iir); + + -- Return TRUE iff TYPE1 and TYPE2 are closely related. + function Are_Types_Closely_Related (Type1, Type2 : Iir) return Boolean; + + -- From the list LIST of function or enumeration literal, extract the + -- list of (return) types. + -- If there is only one type, return it. + -- If there is no types, return NULL. + -- Otherwise, return the list as an overload list. + function Create_List_Of_Types (List : Iir_List) return Iir; + + function Sem_Index_Specification (Name : Iir_Parenthesis_Name; Itype : Iir) + return Iir; + + -- Analyze denoting name NAME. NAME must be either a simple name or an + -- expanded name and so is the result. + function Sem_Denoting_Name (Name: Iir) return Iir; + + -- Like Sem_Denoting_Name but expect a terminal name. + function Sem_Terminal_Name (Name : Iir) return Iir; + + -- Emit an error for NAME that doesn't match its class CLASS_NAME. + procedure Error_Class_Match (Name : Iir; Class_Name : String); + + -- Create an error node for name ORIG; set its expr staticness to none. + function Create_Error_Name (Orig : Iir) return Iir; +end Sem_Names; |