summaryrefslogtreecommitdiff
path: root/src/lmparse.kl
diff options
context:
space:
mode:
authorAdrian Thurston <thurston@complang.org>2012-07-01 12:48:22 -0400
committerAdrian Thurston <thurston@complang.org>2012-07-01 12:48:22 -0400
commit247904a84430b8c9151fa6afb68f01b60afb92c9 (patch)
tree58d498f783a935b02255120c814c387745dc6e41 /src/lmparse.kl
parentd8cdec468bb7efad768d25872147533312cffe91 (diff)
downloadcolm-247904a84430b8c9151fa6afb68f01b60afb92c9.tar.gz
moved 'colm' dir to 'src'
Diffstat (limited to 'src/lmparse.kl')
-rw-r--r--src/lmparse.kl2677
1 files changed, 2677 insertions, 0 deletions
diff --git a/src/lmparse.kl b/src/lmparse.kl
new file mode 100644
index 00000000..3ead7c98
--- /dev/null
+++ b/src/lmparse.kl
@@ -0,0 +1,2677 @@
+/*
+ * Copyright 2006-2012 Adrian Thurston <thurston@complang.org>
+ */
+
+/* This file is part of Colm.
+ *
+ * Colm 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 of the License, or
+ * (at your option) any later version.
+ *
+ * Colm 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 Colm; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
+ */
+
+#include <iostream>
+#include <errno.h>
+
+#include "config.h"
+#include "lmparse.h"
+#include "global.h"
+#include "input.h"
+#include "fsmrun.h"
+
+using std::cout;
+using std::cerr;
+using std::endl;
+
+ParserDict parserDict;
+
+%%{
+
+parser ColmParser;
+
+include "lmparse.kh";
+
+start: root_item_list
+ final {
+ if ( colm_log_compile ) {
+ cerr << "parsing complete" << endl;
+ }
+
+ pd->rootCodeBlock = new CodeBlock( $1->stmtList );
+ };
+
+nonterm root_item_list uses lang_stmt_list;
+
+root_item_list: root_item_list root_item
+ final {
+ $$->stmtList = $1->stmtList;
+
+ /* Maybe a statement. */
+ if ( $2->stmt != 0 )
+ $$->stmtList->append( $2->stmt );
+ };
+
+root_item_list:
+ final {
+ $$->stmtList = new StmtList;
+ };
+
+nonterm root_item uses statement;
+
+root_item: literal_def commit final { $$->stmt = 0; };
+root_item: rl_def commit final { $$->stmt = 0; };
+root_item: token_def commit final { $$->stmt = 0; };
+root_item: cfl_def commit final { $$->stmt = 0; };
+root_item: region_def commit final { $$->stmt = 0; };
+root_item: context_def commit final { $$->stmt = 0; };
+root_item: namespace_def commit final { $$->stmt = 0; };
+root_item: function_def commit final { $$->stmt = 0; };
+root_item: iter_def commit final { $$->stmt = 0; };
+root_item: global_def commit final { $$->stmt = $1->stmt; };
+root_item: statement commit final { $$->stmt = $1->stmt; };
+root_item: pre_eof commit final { $$->stmt = 0; };
+root_item: precedence commit final { $$->stmt = 0; };
+root_item: typedef commit final { $$->stmt = 0; };
+
+nonterm block_open
+{
+ ObjectDef *localFrame;
+};
+
+block_open: '{'
+ final {
+ /* Init the object representing the local frame. */
+ $$->localFrame = new ObjectDef( ObjectDef::FrameType,
+ "local", pd->nextObjectId++ );
+
+ pd->curLocalFrame = $$->localFrame;
+
+ /* Add captures to the local frame. We Depend on these becoming the
+ * first local variables so we can compute their location. */
+
+ /* Make local variables corresponding to the local capture vector. */
+ for ( ReCaptureVect::Iter c = reCaptureVect; c.lte(); c++ )
+ {
+ ObjField *objField = new ObjField( c->objField->loc,
+ c->objField->typeRef, c->objField->name );
+
+ /* Insert it into the field map. */
+ pd->curLocalFrame->insertField( objField->name, objField );
+ }
+ };
+
+block_close: '}'
+ final {
+ /* Pop the cur local frame, back to the root. */
+ pd->curLocalFrame = pd->rootLocalFrame;
+ };
+
+
+iter_def:
+ KW_Iter TK_Word '(' opt_param_list ')' block_open lang_stmt_list block_close
+ final {
+ CodeBlock *codeBlock = new CodeBlock( $7->stmtList );
+ codeBlock->localFrame = $6->localFrame;
+ Function *newFunction = new Function( 0, $2->data,
+ $4->paramList, codeBlock, pd->nextFuncId++, true );
+ pd->functionList.append( newFunction );
+ };
+
+function_def:
+ type_ref TK_Word '(' opt_param_list ')' block_open lang_stmt_list block_close
+ final {
+ CodeBlock *codeBlock = new CodeBlock( $7->stmtList );
+ codeBlock->localFrame = $6->localFrame;
+ Function *newFunction = new Function( $1->typeRef, $2->data,
+ $4->paramList, codeBlock, pd->nextFuncId++, false );
+ pd->functionList.append( newFunction );
+
+ if ( contextStack.length() > 0 )
+ newFunction->inContext = contextStack.top();
+ };
+
+nonterm opt_param_list uses param_list;
+
+opt_param_list: param_list
+ final {
+ $$->paramList = $1->paramList;
+ };
+
+opt_param_list:
+ final {
+ $$->paramList = new ParameterList;
+ };
+
+nonterm param_list
+{
+ ParameterList *paramList;
+};
+
+param_list: param_list param_var_def
+ final {
+ $$->paramList = $1->paramList;
+ $$->paramList->append( $2->objField );
+ };
+
+param_list: param_var_def
+ final {
+ /* Create the map and insert the first item. */
+ $$->paramList = new ParameterList;
+ $$->paramList->append( $1->objField );
+ };
+
+nonterm param_var_def uses var_def;
+
+param_var_def: TK_Word ':' type_ref
+ final {
+ $$->objField = new ObjField( $1->loc, $3->typeRef, $1->data );
+ $$->objField->isParam = true;
+ };
+param_var_def: TK_Word ':' reference_type_ref
+ final {
+ $$->objField = new ObjField( $1->loc, $3->typeRef, $1->data );
+ $$->objField->isParam = true;
+ };
+
+nonterm reference_type_ref uses type_ref;
+
+reference_type_ref: KW_Ref type_ref
+ final {
+ $$->typeRef = new TypeRef( TypeRef::Ref, $1->loc, $2->typeRef );
+ };
+
+nonterm global_def uses statement;
+
+global_def: KW_Export var_def opt_def_init
+ final {
+ $$->stmt = 0;
+
+ if ( contextStack.length() != 0 )
+ error($2->objField->loc) << "cannot export parser context variables" << endp;
+
+ ObjectDef *object = pd->globalObjectDef;
+
+ if ( object->checkRedecl( $2->objField->name ) != 0 )
+ error($2->objField->loc) << "object field renamed" << endp;
+
+ object->insertField( $2->objField->name, $2->objField );
+ $2->objField->isExport = true;
+
+ if ( $3->expr != 0 ) {
+ LangVarRef *varRef = new LangVarRef( $2->objField->loc,
+ new QualItemVect, $2->objField->name );
+
+ $$->stmt = new LangStmt( $2->objField->loc,
+ $3->assignType, varRef, $3->expr );
+ }
+ };
+
+global_def: KW_Global var_def opt_def_init
+ final {
+ $$->stmt = 0;
+
+ ObjectDef *object;
+ if ( contextStack.length() == 0 )
+ object = pd->globalObjectDef;
+ else {
+ Context *context = contextStack.top();
+ $2->objField->context = context;
+ object = context->contextObjDef;
+ }
+
+ if ( object->checkRedecl( $2->objField->name ) != 0 )
+ error($2->objField->loc) << "object field renamed" << endp;
+
+ object->insertField( $2->objField->name, $2->objField );
+
+ if ( $3->expr != 0 ) {
+ LangVarRef *varRef = new LangVarRef( $2->objField->loc,
+ new QualItemVect, $2->objField->name );
+
+ $$->stmt = new LangStmt( $2->objField->loc,
+ $3->assignType, varRef, $3->expr );
+ }
+ };
+
+precedence: pred_type pred_token_list final { pd->predValue++; };
+
+pred_type: KW_Left final { predType = PredLeft; };
+pred_type: KW_Right final { predType = PredRight; };
+pred_type: KW_Nonassoc final { predType = PredNonassoc; };
+
+pred_token_list: pred_token_list ',' pred_token
+ final {
+ };
+
+pred_token_list: pred_token;
+
+nonterm pred_token
+{
+ ProdEl *factor;
+ TypeRef *typeRef;
+};
+
+pred_token:
+ region_qual TK_Word
+ final {
+ TypeRef *typeRef = new TypeRef( $2->loc, $1->nspaceQual, $2->data );
+
+ PredDecl *predDecl = new PredDecl( typeRef, predType, pd->predValue );
+ pd->predDeclList.append( predDecl );
+ };
+
+pred_token:
+ region_qual TK_Literal
+ final {
+ PdaLiteral *literal = new PdaLiteral( $2->loc, *$2 );
+ TypeRef *typeRef = new TypeRef( $2->loc, $1->nspaceQual, literal );
+
+ PredDecl *predDecl = new PredDecl( typeRef, predType, pd->predValue );
+ pd->predDeclList.append( predDecl );
+ };
+
+typedef:
+ KW_Alias TK_Word type_ref
+ final {
+ Namespace *nspace = namespaceStack.top();
+ TypeAlias *typeAlias = new TypeAlias(
+ $1->loc, nspace, $2->data, $3->typeRef );
+ nspace->typeAliasList.append( typeAlias );
+ };
+
+cfl_def: cfl_def_head obj_var_list properties_list cfl_prod_list
+ final {
+ Namespace *nspace = namespaceStack.top();
+ NtDef *ntDef = new NtDef(
+ curDefineId,
+ nspace,
+ $4->defList,
+ pd->objectDef,
+ contextStack.length() > 0 ? contextStack.top() : 0,
+ $3->property & PROPERTY_REDUCE_FIRST );
+
+ nspace->ntDefList.append( ntDef );
+ };
+
+cfl_def_head: KW_Def TK_Word
+ final {
+ curDefineId = $2->data;
+ curDefList = new LelDefList;
+ };
+
+nonterm cfl_prod_list
+{
+ LelDefList *defList;
+};
+
+cfl_prod_list: cfl_prod_list '|' define_prod
+ final {
+ $$->defList = $1->defList;
+ $3->definition->prodNum = $$->defList->length();
+ $$->defList->append( $3->definition );
+ };
+cfl_prod_list: define_prod
+ final {
+ $$->defList = curDefList;
+ $1->definition->prodNum = $$->defList->length();
+ $$->defList->append( $1->definition );
+ };
+
+nonterm property
+{
+ long property;
+};
+
+nonterm properties_list uses property;
+
+properties_list: properties_list property
+ final {
+ $$->property = $1->property | $2->property;
+ };
+properties_list:
+ final {
+ $$->property = 0;
+ };
+
+property:
+ KW_ReduceFirst
+ final {
+ $$->property = PROPERTY_REDUCE_FIRST;
+ };
+
+nonterm opt_prec
+{
+ LangEl *predOf;
+};
+
+opt_prec:
+ final {
+ $$->predOf = 0;
+ };
+
+opt_prec:
+ KW_Prec pred_token
+ final {
+ $$->predOf = $2->factor->langEl;
+ };
+
+nonterm define_prod
+{
+ Definition *definition;
+};
+
+define_prod: '[' prod_el_list ']' opt_commit opt_reduce_code opt_prec
+ final {
+ const InputLoc &loc = $1->loc;
+ //const String &name = curDefineId;
+ ProdElList *prodElList = curProdElList;
+ bool commit = $4->commit;
+ CodeBlock *redBlock = $5->codeBlock;
+ LangEl *predOf = $6->predOf;
+
+ //Namespace *nspace = namespaceStack.top();
+
+ Definition *newDef = new Definition( loc, 0/*prodName*/,
+ prodElList, commit, redBlock,
+ pd->prodList.length(), 0, Definition::Production );
+ newDef->predOf = predOf;
+
+ pd->prodList.append( newDef );
+
+ $$->definition = newDef;
+ };
+
+obj_var_list: obj_var_list var_def
+ final {
+ if ( pd->objectDef->checkRedecl( $2->objField->name ) != 0 )
+ error() << "object field renamed" << endp;
+
+ pd->objectDef->insertField( $2->objField->name, $2->objField );
+ };
+
+obj_var_list:
+ final {
+ pd->objectDef = new ObjectDef( ObjectDef::UserType,
+ curDefineId, pd->nextObjectId++ );
+ };
+
+
+nonterm type_ref
+{
+ TypeRef *typeRef;
+};
+
+type_ref: basic_type_ref
+ final {
+ $$->typeRef = $1->typeRef;
+ };
+
+type_ref: KW_Map '<' type_ref type_ref '>'
+ final {
+ NamespaceQual *nspaceQual = new NamespaceQual(
+ namespaceStack.top(), regionStack.top() );
+ $$->typeRef = new TypeRef( TypeRef::Map, InputLoc(), nspaceQual,
+ $3->typeRef, $4->typeRef );
+ };
+
+type_ref: KW_List '<' type_ref '>'
+ final {
+ NamespaceQual *nspaceQual = new NamespaceQual(
+ namespaceStack.top(), regionStack.top() );
+ $$->typeRef = new TypeRef( TypeRef::List, InputLoc(), nspaceQual, $3->typeRef, 0 );
+ };
+type_ref: KW_Vector '<' type_ref '>'
+ final {
+ NamespaceQual *nspaceQual = new NamespaceQual(
+ namespaceStack.top(), regionStack.top() );
+ $$->typeRef = new TypeRef( TypeRef::Vector, InputLoc(), nspaceQual, $3->typeRef, 0 );
+ };
+type_ref: KW_Accum '<' type_ref '>'
+ final {
+ NamespaceQual *nspaceQual = new NamespaceQual(
+ namespaceStack.top(), regionStack.top() );
+ $$->typeRef = new TypeRef( TypeRef::Parser, InputLoc(), nspaceQual, $3->typeRef, 0 );
+ };
+
+nonterm basic_type_ref uses type_ref;
+
+basic_type_ref: region_qual TK_Word opt_repeat
+ final {
+ $$->typeRef = new TypeRef( $2->loc, $1->nspaceQual, $2->data );
+ $$->typeRef->repeatType = $3->repeatType;
+ };
+
+basic_type_ref: KW_Ptr region_qual TK_Word opt_repeat
+ final {
+ $$->typeRef = new TypeRef( $1->loc, $2->nspaceQual, $3->data );
+ $$->typeRef->repeatType = $4->repeatType;
+ $$->typeRef = new TypeRef( TypeRef::Ptr, $1->loc, $$->typeRef );
+ };
+
+
+nonterm var_def
+{
+ InputLoc loc;
+ ObjField *objField;
+};
+
+var_def: TK_Word ':' type_ref
+ final {
+ /* Return an object field object. The user of this nonterminal must
+ * load it into the approrpriate map and do error checking. */
+ $$->objField = new ObjField( $1->loc, $3->typeRef, $1->data );
+ };
+
+region_def:
+ region_head '{' root_item_list '}'
+ final {
+ /* Pop the top of the stack. */
+ regionStack.pop();
+ };
+
+region_head:
+ KW_Lex TK_Word
+ final {
+ /* Just for ignores. */
+ String scannerNameIgn( $2->data.length() + 2, "<%s>-ign", $2->data.data );
+ TokenRegion *tokenRegionIgn = createRegion( scannerNameIgn );
+
+ /* Just for collect ignores. Will use the ignore-only start state. */
+ String scannerNameCi( $2->data.length() + 2, "<%s>-ci", $2->data.data );
+ TokenRegion *tokenRegionCi = createRegion( scannerNameCi );
+
+ /* Just for tokens. */
+ String scannerNameTok( $2->data.length() + 2, "<%s>-tok", $2->data.data );
+ TokenRegion *tokenRegionTok = createRegion( scannerNameTok );
+
+ /* Make the new token region. */
+ String scannerName( $2->data.length() + 2, "<%s>", $2->data.data );
+ TokenRegion *tokenRegion = createRegion( scannerName );
+
+ regionStack.push( tokenRegion );
+
+ tokenRegion->ignoreOnlyRegion = tokenRegionIgn;
+ tokenRegion->tokenOnlyRegion = tokenRegionTok;
+ tokenRegion->ciRegion = tokenRegionCi;
+
+ tokenRegion->isFullRegion = true;
+ tokenRegionIgn->isIgnoreOnly = true;
+ tokenRegionCi->isCiOnly = true;
+ tokenRegionTok->isTokenOnly = true;
+
+ tokenRegionIgn->derivedFrom = tokenRegion;
+ tokenRegionCi->derivedFrom = tokenRegion;
+ tokenRegionTok->derivedFrom = tokenRegion;
+ };
+
+namespace_def:
+ namespace_head '{' root_item_list '}'
+ final {
+ namespaceStack.pop();
+ };
+
+namespace_head:
+ KW_Namespace TK_Word
+ final {
+ /* Make the new namespace. */
+ Namespace *nspace = new Namespace( InputLoc(), $2->data,
+ pd->namespaceList.length(), namespaceStack.top() );
+ namespaceStack.top()->childNamespaces.append( nspace );
+ pd->namespaceList.append( nspace );
+ namespaceStack.push( nspace );
+ };
+
+context_var_def:
+ var_def
+ final {
+ ObjectDef *object;
+ if ( contextStack.length() == 0 )
+ error($1->loc) << "internal error: no context stack items found" << endp;
+
+ Context *context = contextStack.top();
+ $1->objField->context = context;
+ object = context->contextObjDef;
+
+ if ( object->checkRedecl( $1->objField->name ) != 0 )
+ error($1->objField->loc) << "object field renamed" << endp;
+
+ object->insertField( $1->objField->name, $1->objField );
+ };
+
+
+context_item: context_var_def commit;
+context_item: literal_def commit;
+context_item: rl_def commit;
+context_item: token_def commit;
+context_item: cfl_def commit;
+context_item: region_def commit;
+context_item: context_def commit;
+context_item: function_def commit;
+context_item: iter_def commit;
+context_item: pre_eof commit;
+context_item: precedence commit;
+
+context_item_list:
+ context_item_list context_item;
+context_item_list:
+ ;
+
+context_def:
+ context_head '{' context_item_list '}'
+ final {
+ contextStack.pop();
+ namespaceStack.pop();
+ };
+
+context_head:
+ KW_Context TK_Word
+ final {
+ /* Make the new namespace. */
+ Namespace *nspace = new Namespace( InputLoc(), $2->data,
+ pd->namespaceList.length(), namespaceStack.top() );
+ namespaceStack.top()->childNamespaces.append( nspace );
+ pd->namespaceList.append( nspace );
+ namespaceStack.push( nspace );
+
+ Context *context = new Context( $1->loc, 0 );
+ contextStack.push( context );
+
+ ContextDef *contextDef = new ContextDef( $2->data, context, nspace );
+ nspace->contextDefList.append( contextDef );
+
+ context->contextObjDef = new ObjectDef( ObjectDef::UserType,
+ $2->data, pd->nextObjectId++ );
+ };
+
+pattern_list: pattern_list pattern;
+pattern_list: init_pattern_list pattern;
+
+init_pattern_list:
+ final {
+ patternItemList = new PatternItemList;
+ };
+
+pattern: '"' litpat_el_list '"';
+pattern: '[' pattern_el_list ']';
+
+litpat_el_list: litpat_el_list litpat_el;
+litpat_el_list: ;
+
+litpat_el: TK_LitPat
+ final {
+ PatternItem *patternItem = new PatternItem( $1->loc, $1->data,
+ PatternItem::InputText );
+ patternItemList->append( patternItem );
+ };
+
+litpat_el: '[' pattern_el_list ']';
+
+pattern_el_list: pattern_el_list pattern_el;
+pattern_el_list: ;
+
+pattern_el: opt_label pattern_el_type_or_lit
+ final {
+ /* Store the variable reference in the pattern itemm. */
+ $2->patternItem->varRef = $1->varRef;
+
+ if ( $1->varRef != 0 ) {
+ if ( pd->curLocalFrame->checkRedecl( $1->varRef->name ) != 0 ) {
+ error( $1->varRef->loc ) << "variable " << $1->varRef->name <<
+ " redeclared" << endp;
+ }
+
+ TypeRef *typeRef = $2->patternItem->factor->typeRef;
+ ObjField *objField = new ObjField( InputLoc(), typeRef, $1->varRef->name );
+
+ /* Insert it into the field map. */
+ pd->curLocalFrame->insertField( $1->varRef->name, objField );
+ }
+ };
+
+pattern_el: '"' litpat_el_list '"';
+pattern_el: '?' TK_Word
+ final {
+ /* FIXME: Implement */
+ assert(false);
+ };
+
+nonterm pattern_el_type_or_lit
+{
+ PatternItem *patternItem;
+};
+
+pattern_el_type_or_lit: region_qual TK_Word opt_repeat
+ final {
+ TypeRef *typeRef = new TypeRef( $2->loc, $1->nspaceQual, $2->data );
+ typeRef->repeatType = $3->repeatType;
+ ProdEl *factor = new ProdEl( ProdEl::ReferenceType, $2->loc, 0, false, typeRef, 0 );
+ $$->patternItem = new PatternItem( $2->loc, factor, PatternItem::FactorType );
+ patternItemList->append( $$->patternItem );
+ };
+
+pattern_el_type_or_lit: region_qual TK_Literal opt_repeat
+ final {
+ PdaLiteral *literal = new PdaLiteral( $2->loc, *$2 );
+ TypeRef *typeRef = new TypeRef( $2->loc, $1->nspaceQual, literal );
+ typeRef->repeatType = $3->repeatType;
+
+ ProdEl *factor = new ProdEl( ProdEl::ReferenceType, $2->loc, 0, false, typeRef, 0 );
+ $$->patternItem = new PatternItem( $2->loc, factor, PatternItem::FactorType );
+ patternItemList->append( $$->patternItem );
+ };
+
+nonterm opt_label
+{
+ /* Variable reference. */
+ LangVarRef *varRef;
+};
+
+opt_label: TK_Word ':'
+ final {
+ $$->varRef = new LangVarRef( $1->loc, new QualItemVect, $1->data );
+ };
+opt_label:
+ final {
+ $$->varRef = 0;
+ };
+
+#
+# Replacement
+#
+
+repl_list: repl_list replacement;
+repl_list: init_repl_list replacement;
+
+init_repl_list:
+ final {
+ replItemList = new ReplItemList;
+ };
+
+replacement: '"' lit_repl_el_list '"';
+replacement: '[' repl_el_list ']';
+
+lit_repl_el_list: lit_repl_el_list lit_repl_el;
+lit_repl_el_list: ;
+
+lit_repl_el: TK_LitPat
+ final {
+ ReplItem *replItem = new ReplItem( $1->loc, ReplItem::InputText, $1->data );
+ replItemList->append( replItem );
+ };
+
+lit_repl_el: '[' repl_el_list ']';
+
+repl_el_list: repl_el_list repl_el;
+repl_el_list: ;
+
+repl_el: region_qual TK_Literal
+ final {
+ PdaLiteral *literal = new PdaLiteral( $2->loc, *$2 );
+ TypeRef *typeRef = new TypeRef( $2->loc, $1->nspaceQual, literal );
+ typeRef->repeatType = RepeatNone;
+ ProdEl *factor = new ProdEl( ProdEl::LiteralType, $2->loc, 0, false, typeRef, 0 );
+ ReplItem *replItem = new ReplItem( $2->loc, ReplItem::FactorType, factor );
+ replItemList->append( replItem );
+ };
+repl_el: '"' lit_repl_el_list '"';
+
+repl_el: code_expr
+ final {
+ ReplItem *replItem = new ReplItem( $1->expr->loc, ReplItem::ExprType, $1->expr );
+ replItemList->append( replItem );
+ };
+
+#
+# Accum
+#
+accumulate: init_repl_list accum_list;
+accumulate: init_repl_list code_expr
+ final {
+ ReplItem *replItem = new ReplItem( $2->expr->loc, ReplItem::ExprType, $2->expr );
+ replItemList->append( replItem );
+ };
+
+accum_list: accum_list accum;
+accum_list: accum;
+
+init_accum_list:
+ final {
+ replItemList = new ReplItemList;
+ };
+
+accum: '"' lit_accum_el_list '"';
+accum: '[' accum_el_list ']';
+
+lit_accum_el_list: lit_accum_el_list lit_accum_el;
+lit_accum_el_list: ;
+
+lit_accum_el: TK_LitPat
+ final {
+ ReplItem *replItem = new ReplItem( $1->loc, ReplItem::InputText, $1->data );
+ replItemList->append( replItem );
+ };
+
+lit_accum_el: '[' accum_el_list ']';
+
+accum_el_list: accum_el_list accum_el;
+accum_el_list: ;
+
+#accum_el: region_qual TK_Literal
+# final {
+# PdaLiteral *literal = new PdaLiteral( $2->loc, *$2 );
+# ProdEl *factor = new ProdEl( $2->loc, false, $1->nspaceQual,
+# literal, 0 );
+# ReplItem *replItem = new ReplItem( $2->loc, ReplItem::FactorType, factor );
+# replItemList->append( replItem );
+# };
+accum_el: code_expr
+ final {
+ ReplItem *replItem = new ReplItem( $1->expr->loc, ReplItem::ExprType, $1->expr );
+ replItemList->append( replItem );
+ };
+
+accum_el: '"' lit_accum_el_list '"';
+
+
+#
+# String
+#
+
+string_list: string_list string;
+string_list: init_string_list string;
+
+init_string_list:
+ final {
+ replItemList = new ReplItemList;
+ };
+
+string: '"' lit_string_el_list '"';
+string: '[' string_el_list ']';
+
+lit_string_el_list: lit_string_el_list lit_string_el;
+lit_string_el_list: ;
+
+lit_string_el: TK_LitPat
+ final {
+ ReplItem *replItem = new ReplItem( $1->loc, ReplItem::InputText, $1->data );
+ replItemList->append( replItem );
+ };
+
+lit_string_el: '[' string_el_list ']';
+
+string_el_list: string_el_list string_el;
+string_el_list: ;
+
+#accum_el: region_qual TK_Literal
+# final {
+# PdaLiteral *literal = new PdaLiteral( $2->loc, *$2 );
+# ProdEl *factor = new ProdEl( $2->loc, false, $1->nspaceQual,
+# literal, 0 );
+# ReplItem *replItem = new ReplItem( $2->loc, ReplItem::FactorType, factor );
+# replItemList->append( replItem );
+# };
+string_el: code_expr
+ final {
+ ReplItem *replItem = new ReplItem( $1->expr->loc, ReplItem::ExprType, $1->expr );
+ replItemList->append( replItem );
+ };
+
+string_el: '"' lit_string_el_list '"';
+
+prod_el_list:
+ prod_el_list prod_el
+ final {
+ curProdElList->append( $2->factor );
+ };
+
+prod_el_list:
+ final { curProdElList = new ProdElList; };
+
+nonterm opt_no_ignore { bool value; };
+
+opt_no_ignore: KW_Ni final { $$->value = true; };
+opt_no_ignore: final { $$->value = false; };
+
+nonterm prod_el
+{
+ ProdEl *factor;
+};
+
+prod_el:
+ opt_capture opt_commit region_qual TK_Word opt_repeat
+ final {
+ TypeRef *typeRef = new TypeRef( $4->loc, $3->nspaceQual, $4->data );
+ typeRef->repeatType = $5->repeatType;
+ $$->factor = new ProdEl( ProdEl::ReferenceType, $4->loc, $1->objField, $2->commit, typeRef, 0 );
+
+ /* If there is a capture, create the field. */
+ if ( $1->objField != 0 ) {
+ /* Might already exist. */
+ ObjField *objField = pd->objectDef->checkRedecl( $1->objField->name );
+ if ( objField == 0 ) {
+ objField = $1->objField;
+ objField->typeRef = typeRef;
+ pd->objectDef->insertField( objField->name, objField );
+ }
+ else {
+ /* FIXME: check the types are the same. */
+ //error() << "object field renamed" << endp;
+ }
+
+ objField->isRhsGet = true;
+ RhsVal rhsVal( curDefList->length(), curProdElList->length() );
+ objField->rhsVal.append( RhsVal( curDefList->length(), curProdElList->length() ) );
+ }
+ };
+
+prod_el:
+ opt_capture opt_commit region_qual TK_Literal opt_repeat
+ final {
+ /* Create a new factor node going to a concat literal. */
+ PdaLiteral *literal = new PdaLiteral( $4->loc, *$4 );
+ TypeRef *typeRef = new TypeRef( $4->loc, $3->nspaceQual, literal );
+ typeRef->repeatType = $5->repeatType;
+ $$->factor = new ProdEl( ProdEl::LiteralType, $4->loc, $1->objField, $2->commit, typeRef, 0 );
+
+ /* If there is a capture, create the field. */
+ if ( $1->objField != 0 ) {
+ $1->objField->typeRef = typeRef;
+ if ( pd->objectDef->checkRedecl( $1->objField->name ) != 0 )
+ error() << "object field renamed" << endp;
+
+ pd->objectDef->insertField( $1->objField->name, $1->objField );
+ }
+ };
+
+nonterm opt_repeat
+{
+ bool opt;
+ bool repeat;
+ RepeatType repeatType;
+};
+
+opt_repeat: '*' final { $$->opt = false; $$->repeat = true; $$->repeatType = RepeatRepeat; };
+opt_repeat: '+' final { $$->opt = false; $$->repeat = false; $$->repeatType = RepeatList; };
+opt_repeat: '?' final { $$->opt = true; $$->repeat = false; $$->repeatType = RepeatOpt; };
+opt_repeat: final { $$->opt = false; $$->repeat = false; $$->repeatType = RepeatNone; };
+
+nonterm region_qual
+{
+ NamespaceQual *nspaceQual;
+};
+
+region_qual: region_qual TK_Word TK_DoubleColon
+ final {
+ $$->nspaceQual = $1->nspaceQual;
+ $$->nspaceQual->qualNames.append( $2->data );
+ };
+
+region_qual:
+ final {
+ $$->nspaceQual = new NamespaceQual( namespaceStack.top(), regionStack.top() );
+ };
+
+literal_def: KW_Literal literal_list;
+
+literal_list: literal_list ',' literal_item;
+literal_list: literal_item;
+
+literal_item: opt_no_ignore TK_Literal opt_no_ignore
+ final {
+ /* Create a name for the literal. */
+ String name( 32, "_literal_%.4x", pd->nextTokenId );
+
+ bool insideRegion = regionStack.top() != pd->rootRegion;
+ if ( !insideRegion ) {
+ /* Just for ignores. */
+ String scannerNameIgn( name.length() + 2, "<%s>-ign", name.data );
+ TokenRegion *tokenRegionIgn = createRegion( scannerNameIgn );
+
+ /* Just for collect ignores. Will use the ignore-only start state. */
+ String scannerNameCi( name.length() + 2, "<%s>-ci", name.data );
+ TokenRegion *tokenRegionCi = createRegion( scannerNameCi );
+
+ /* Just for tokens. */
+ String scannerNameTok( name.length() + 2, "<%s>-tok", name.data );
+ TokenRegion *tokenRegionTok = createRegion( scannerNameTok );
+
+ /* Make a new token region just for the token. */
+ String scannerName( name.length() + 2, "<%s>", name.data );
+ TokenRegion *tokenRegion = createRegion( scannerName );
+
+ regionStack.push( tokenRegion );
+
+ tokenRegion->ignoreOnlyRegion = tokenRegionIgn;
+ tokenRegion->tokenOnlyRegion = tokenRegionTok;
+ tokenRegion->ciRegion = tokenRegionCi;
+
+ tokenRegion->isFullRegion = true;
+ tokenRegionIgn->isIgnoreOnly = true;
+ tokenRegionCi->isCiOnly = true;
+ tokenRegionTok->isTokenOnly = true;
+
+ tokenRegionIgn->derivedFrom = tokenRegion;
+ tokenRegionCi->derivedFrom = tokenRegion;
+ tokenRegionTok->derivedFrom = tokenRegion;
+ }
+
+ bool unusedCI;
+ String interp;
+ prepareLitString( interp, unusedCI, $2->data, $2->loc );
+
+ /* Look for the production's associated region. */
+ Namespace *nspace = namespaceStack.top();
+ TokenRegion *region = regionStack.top();
+
+
+ LiteralDictEl *ldel = nspace->literalDict.find( interp );
+ if ( ldel != 0 )
+ error( $2->loc ) << "literal already defined in this namespace" << endp;
+ else {
+ Join *join = new Join( new Expression( new Term( new FactorWithAug(
+ new FactorWithRep( $2->loc, new FactorWithNeg( $2->loc, new Factor(
+ new Literal( $2->loc, $2->data,
+ Literal::LitString ) ) ) ) ) ) ) );
+
+ if ( strcmp( interp.data, "" ) == 0 ) {
+ TokenDef *tokenDef = new TokenDef( name, $2->data, true, false, join,
+ 0, $2->loc, pd->nextTokenId++, nspace, region, 0, 0, 0 );
+
+ //region->tokenDefList.append( tokenDef );
+
+ ldel = nspace->literalDict.insert( interp, tokenDef );
+ nspace->tokenDefList.append( tokenDef );
+
+ tokenDef->isZero = true;
+ }
+ else {
+ TokenDef *tokenDef = new TokenDef( name, $2->data, true, false, join,
+ 0, $2->loc, pd->nextTokenId++, nspace, region, 0, 0, 0 );
+ region->tokenDefList.append( tokenDef );
+ ldel = nspace->literalDict.insert( interp, tokenDef );
+ nspace->tokenDefList.append( tokenDef );
+
+ if ( $1->value )
+ tokenDef->noPreIgnore = true;
+ if ( $3->value )
+ tokenDef->noPostIgnore = true;
+
+ TokenDef *tokenDefTok = new TokenDef( name + "_tok", $2->data, true, false, join,
+ 0, $2->loc, pd->nextTokenId++, nspace, region->tokenOnlyRegion, 0, 0, 0 );
+ tokenDefTok->dupOf = tokenDef;
+ region->tokenOnlyRegion->tokenDefList.append( tokenDefTok );
+ ldel = nspace->literalDict.insert( "|" + interp + "_tok", tokenDefTok );
+ nspace->tokenDefList.append( tokenDefTok );
+ }
+ }
+
+ if ( !insideRegion ) {
+ /* Leave the region just for this token. */
+ regionStack.pop();
+ }
+ };
+
+
+# These two productions are responsible for setting and unsetting the Regular
+# language scanning context.
+enter_rl:
+ try {
+ enterRl = true;
+ }
+ undo {
+ enterRl = false;
+ };
+leave_rl:
+ try {
+ enterRl = false;
+ }
+ undo {
+ enterRl = true;
+ };
+
+token_def:
+ token_or_ignore token_def_name obj_var_list
+ enter_rl opt_no_ignore '/' opt_rl_join leave_rl '/' opt_no_ignore
+ opt_translate
+ final {
+ bool ignore = $1->ignore;
+ String name = $2->name;
+ Join *join = $7->join;
+ CodeBlock *transBlock = $11->transBlock;
+
+ /* Check the region if this is for an ignore. */
+ if ( ignore && !pd->insideRegion )
+ error($1->loc) << "ignore tokens can only appear inside scanners" << endp;
+
+ /* Check the name if this is a token. */
+ if ( !ignore && name == 0 )
+ error($1->loc) << "tokens must have a name" << endp;
+
+ /* Give a default name to ignores. */
+ if ( name == 0 )
+ name.setAs( 32, "_ignore_%.4x", pd->nextTokenId );
+
+ Namespace *nspace = namespaceStack.top();
+ TokenRegion *region = regionStack.top();
+
+ TokenDef *tokenDef = new TokenDef( name, String(), false, ignore, join,
+ transBlock, $1->loc, pd->nextTokenId++, nspace, region,
+ &reCaptureVect, pd->objectDef,
+ contextStack.length() > 0 ? contextStack.top() : 0 );
+
+ region->tokenDefList.append( tokenDef );
+ nspace->tokenDefList.append( tokenDef );
+
+ if ( $5->value )
+ tokenDef->noPreIgnore = true;
+ if ( $10->value )
+ tokenDef->noPostIgnore = true;
+
+ /* All again for the ignore. */
+ if ( ignore ) {
+ TokenDef *tokenDefIgn = new TokenDef( name + "_ign", String(), false, ignore, join,
+ 0, $1->loc, pd->nextTokenId++, nspace, region->ignoreOnlyRegion,
+ &reCaptureVect, pd->objectDef,
+ contextStack.length() > 0 ? contextStack.top() : 0 );
+
+ tokenDefIgn->dupOf = tokenDef;
+
+ region->ignoreOnlyRegion->tokenDefList.append( tokenDefIgn );
+ nspace->tokenDefList.append( tokenDefIgn );
+ }
+ else {
+ TokenDef *tokenDefTok = new TokenDef( name + "_tok", String(), false, ignore, join,
+ 0, $1->loc, pd->nextTokenId++, nspace, region->tokenOnlyRegion,
+ &reCaptureVect, pd->objectDef,
+ contextStack.length() > 0 ? contextStack.top() : 0 );
+
+ tokenDefTok->dupOf = tokenDef;
+
+ region->tokenOnlyRegion->tokenDefList.append( tokenDefTok );
+ nspace->tokenDefList.append( tokenDefTok );
+ }
+
+ /* This is created and pushed in the name. */
+ if ( !pd->insideRegion ) {
+ /* Leave the region that we made just for this token. */
+ regionStack.pop();
+ }
+
+ if ( join != 0 ) {
+ /* Create a regular language definition so the token can be used to
+ * make other tokens */
+ addRegularDef( $1->loc, namespaceStack.top(), name, join );
+ }
+
+
+ reCaptureVect.empty();
+ };
+
+nonterm token_or_ignore
+{
+ InputLoc loc;
+ bool ignore;
+};
+
+token_or_ignore: KW_Token
+ final { $$->loc = $1->loc; $$->ignore = false; };
+
+token_or_ignore: KW_Ignore
+ final { $$->loc = $1->loc; $$->ignore = true; };
+
+nonterm class token_def_name
+{
+ String name;
+};
+
+token_def_name:
+ opt_name
+ final {
+ String name = $1->name;
+
+ $$->name = name;
+ pd->insideRegion = regionStack.top() != pd->rootRegion;
+ curDefineId = name;
+
+ if ( !pd->insideRegion ) {
+ /* For just ignores. */
+ String scannerNameIgn( name.length() + 2, "<%s>-ign", name.data );
+ TokenRegion *tokenRegionIgn = createRegion( scannerNameIgn );
+
+ /* Just for explicitly collecting ignores. */
+ String scannerNameCi( name.length() + 2, "<%s>-ci", name.data );
+ TokenRegion *tokenRegionCi = createRegion( scannerNameCi );
+
+ /* Just for tokens. */
+ String scannerNameTok( name.length() + 2, "<%s>-tok", name.data );
+ TokenRegion *tokenRegionTok = createRegion( scannerNameTok );
+
+ /* If not inside a region, make one for the token. */
+ String scannerName( name.length() + 2, "<%s>", name.data );
+ TokenRegion *tokenRegion = createRegion( scannerName );
+
+ regionStack.push( tokenRegion );
+
+ tokenRegion->ignoreOnlyRegion = tokenRegionIgn;
+ tokenRegion->tokenOnlyRegion = tokenRegionTok;
+ tokenRegion->ciRegion = tokenRegionCi;
+
+ tokenRegion->isFullRegion = true;
+ tokenRegionIgn->isIgnoreOnly = true;
+ tokenRegionCi->isCiOnly = true;
+ tokenRegionTok->isTokenOnly = true;
+
+ tokenRegionIgn->derivedFrom = tokenRegion;
+ tokenRegionCi->derivedFrom = tokenRegion;
+ tokenRegionTok->derivedFrom = tokenRegion;
+ }
+
+ /* Reset the lable id counter. */
+ pd->nextLabelId = 0;
+ };
+
+nonterm class opt_name
+{
+ String name;
+};
+
+opt_name: TK_Word final { $$->name = $1->data; };
+opt_name: ;
+
+nonterm opt_translate
+{
+ CodeBlock *transBlock;
+};
+
+opt_translate:
+ block_open lang_stmt_list block_close
+ final {
+ $$->transBlock = new CodeBlock( $2->stmtList );
+ $$->transBlock->localFrame = $1->localFrame;
+ $$->transBlock->context = contextStack.length() == 0 ? 0 : contextStack.top();
+ };
+
+opt_translate:
+ final {
+ $$->transBlock = 0;
+ };
+
+pre_eof:
+ KW_Preeof block_open lang_stmt_list block_close
+ final {
+ bool insideRegion = regionStack.top() != pd->rootRegion;
+ if ( !insideRegion )
+ error($1->loc) << "preeof must be used inside an existing region" << endl;
+
+ CodeBlock *codeBlock = new CodeBlock( $3->stmtList );
+ codeBlock->localFrame = $2->localFrame;
+ codeBlock->context = contextStack.length() == 0 ? 0 : contextStack.top();
+
+ TokenRegion *region = regionStack.top();
+ region->preEofBlock = codeBlock;
+ };
+
+rl_def:
+ KW_Rl machine_name enter_rl '/' rl_join leave_rl '/'
+ final {
+ /* Generic creation of machine for instantiation and assignment. */
+ addRegularDef( $2->loc, namespaceStack.top(), $2->data, $5->join );
+
+ if ( reCaptureVect.length() > 0 )
+ error($1->loc) << "rl definitions cannot capture vars" << endl;
+ };
+
+type class token_data
+{
+ InputLoc loc;
+ String data;
+};
+
+nonterm machine_name uses token_data;
+
+machine_name:
+ TK_Word
+ final {
+ /* Make/get the priority key. The name may have already been referenced
+ * and therefore exist. */
+ PriorDictEl *priorDictEl;
+ if ( pd->priorDict.insert( $1->data, pd->nextPriorKey, &priorDictEl ) )
+ pd->nextPriorKey += 1;
+ pd->curDefPriorKey = priorDictEl->value;
+
+ /* Make/get the local error key. */
+ LocalErrDictEl *localErrDictEl;
+ if ( pd->localErrDict.insert( $1->data, pd->nextLocalErrKey, &localErrDictEl ) )
+ pd->nextLocalErrKey += 1;
+ pd->curDefLocalErrKey = localErrDictEl->value;
+
+ $$->loc = $1->loc;
+ $$->data = $1->data;
+ };
+
+#
+# Reduce statements
+#
+
+nonterm opt_reduce_code
+{
+ CodeBlock *codeBlock;
+};
+
+opt_reduce_code:
+ final { $$->codeBlock = 0; };
+
+opt_reduce_code:
+ start_reduce lang_stmt_list block_close
+ final {
+ $$->codeBlock = new CodeBlock( $2->stmtList );
+ $$->codeBlock->localFrame = $1->localFrame;
+ $$->codeBlock->context = contextStack.length() == 0 ? 0 : contextStack.top();
+ };
+
+nonterm start_reduce uses block_open;
+
+start_reduce:
+ block_open
+ final {
+ $$->localFrame = $1->localFrame;
+ };
+
+nonterm lang_stmt_list
+{
+ StmtList *stmtList;
+};
+
+lang_stmt_list: rec_stmt_list opt_require_stmt
+ final {
+ $$->stmtList = $1->stmtList;
+ if ( $2->stmt != 0 )
+ $$->stmtList->append( $2->stmt );
+ };
+
+nonterm rec_stmt_list uses lang_stmt_list;
+
+rec_stmt_list: rec_stmt_list statement
+ final {
+ $$->stmtList = $1->stmtList;
+
+ /* Maybe a statement was generated. */
+ if ( $2->stmt != 0 )
+ $$->stmtList->append( $2->stmt );
+ };
+
+rec_stmt_list:
+ final {
+ $$->stmtList = new StmtList;
+ };
+
+nonterm opt_def_init
+{
+ LangExpr *expr;
+ LangStmt::Type assignType;
+};
+
+opt_def_init: '=' code_expr
+ final {
+ $$->expr = $2->expr;
+ $$->assignType = LangStmt::AssignType;
+ };
+opt_def_init:
+ final {
+ $$->expr = 0;
+ };
+
+scope_push:
+ final {
+ pd->curLocalFrame->pushScope();
+ //cout << "push scope" << endl;
+ };
+
+scope_pop:
+ final {
+ pd->curLocalFrame->popScope();
+ //cout << "pop scope" << endl;
+ };
+
+nonterm statement
+{
+ LangStmt *stmt;
+};
+nonterm for_scope uses statement;
+
+statement: var_def opt_def_init
+ final {
+ /* By default no statement here. Maybe will add an initialization. */
+ $$->stmt = 0;
+
+ /* Check for redeclaration. */
+ if ( pd->curLocalFrame->checkRedecl( $1->objField->name ) != 0 ) {
+ error( $1->objField->loc ) << "variable " << $1->objField->name <<
+ " redeclared" << endp;
+ }
+
+ /* Insert it into the field map. */
+ pd->curLocalFrame->insertField( $1->objField->name, $1->objField );
+
+ //cout << "var def " << $1->objField->name << endl;
+
+ if ( $2->expr != 0 ) {
+ LangVarRef *varRef = new LangVarRef( $1->objField->loc,
+ new QualItemVect, $1->objField->name );
+
+ $$->stmt = new LangStmt( $1->objField->loc,
+ $2->assignType, varRef, $2->expr );
+ }
+ };
+statement: var_ref '=' code_expr
+ final {
+ $$->stmt = new LangStmt( $2->loc, LangStmt::AssignType, $1->varRef, $3->expr );
+ };
+statement: KW_Print '(' code_expr_list ')'
+ final {
+ $$->stmt = new LangStmt( $1->loc, LangStmt::PrintType, $3->exprVect );
+ };
+statement: KW_PrintXMLAC '(' code_expr_list ')'
+ final {
+ $$->stmt = new LangStmt( $1->loc, LangStmt::PrintXMLACType, $3->exprVect );
+ };
+statement: KW_PrintXML '(' code_expr_list ')'
+ final {
+ $$->stmt = new LangStmt( $1->loc, LangStmt::PrintXMLType, $3->exprVect );
+ };
+statement: KW_PrintStream '(' code_expr_list ')'
+ final {
+ $$->stmt = new LangStmt( $1->loc, LangStmt::PrintStreamType, $3->exprVect );
+ };
+statement: code_expr
+ final {
+ $$->stmt = new LangStmt( InputLoc(), LangStmt::ExprType, $1->expr );
+ };
+statement: if_stmt
+ final {
+ $$->stmt = $1->stmt;
+ };
+statement: KW_Reject
+ final {
+ $$->stmt = new LangStmt( $1->loc, LangStmt::RejectType );
+ };
+statement: KW_While scope_push code_expr block_or_single scope_pop
+ final {
+ $$->stmt = new LangStmt( LangStmt::WhileType, $3->expr, $4->stmtList );
+ };
+
+for_scope: TK_Word ':' type_ref KW_In iter_call block_or_single
+ final {
+ /* Check for redeclaration. */
+ if ( pd->curLocalFrame->checkRedecl( $1->data ) != 0 )
+ error( $1->loc ) << "variable " << $1->data << " redeclared" << endp;
+
+ /* Note that we pass in a null type reference. This type is dependent
+ * on the result of the iter_call lookup since it must contain a reference
+ * to the iterator that is called. This lookup is done at compile time. */
+ ObjField *iterField = new ObjField( $1->loc, (TypeRef*)0, $1->data );
+ pd->curLocalFrame->insertField( $1->data, iterField );
+
+ $$->stmt = new LangStmt( $1->loc, LangStmt::ForIterType,
+ iterField, $3->typeRef, $5->langTerm, $6->stmtList );
+ };
+
+statement: KW_For scope_push for_scope scope_pop
+ final {
+ $$->stmt = $3->stmt;
+ };
+
+statement: KW_Return code_expr
+ final {
+ $$->stmt = new LangStmt( $1->loc, LangStmt::ReturnType, $2->expr );
+ };
+statement: KW_Break
+ final {
+ $$->stmt = new LangStmt( LangStmt::BreakType );
+ };
+statement: KW_Yield var_ref
+ final {
+ $$->stmt = new LangStmt( LangStmt::YieldType, $2->varRef );
+ };
+statement: var_ref TK_LtLt accumulate
+ final {
+ Namespace *nspace = namespaceStack.top();
+ TokenRegion *region = regionStack.top();
+ ParserText *parserText = new ParserText( $2->loc, nspace, region, replItemList );
+ pd->parserTextList.append( parserText );
+
+ $$->stmt = new LangStmt( LangStmt::ParserType, $1->varRef, parserText );
+ };
+statement: KW_Send var_ref accumulate
+ final {
+ Namespace *nspace = namespaceStack.top();
+ TokenRegion *region = regionStack.top();
+ ParserText *parserText = new ParserText( $1->loc, nspace, region, replItemList );
+ pd->parserTextList.append( parserText );
+
+ $$->stmt = new LangStmt( LangStmt::ParserType, $2->varRef, parserText );
+ };
+
+nonterm opt_require_stmt uses statement;
+
+opt_require_stmt:
+ scope_push require_pattern lang_stmt_list scope_pop
+ final {
+ $$->stmt = new LangStmt( LangStmt::IfType, $2->expr, $3->stmtList, 0 );
+ };
+opt_require_stmt:
+ final {
+ $$->stmt = 0;
+ };
+
+nonterm require_pattern uses code_expr;
+
+require_pattern:
+ KW_Require var_ref pattern_list
+ final {
+ Namespace *nspace = namespaceStack.top();
+ TokenRegion *region = regionStack.top();
+ Pattern *pattern = new Pattern( $1->loc, nspace, region,
+ patternItemList, pd->nextPatReplId++ );
+ pd->patternList.append( pattern );
+
+ $$->expr = new LangExpr(
+ new LangTerm( LangTerm::MatchType, $2->varRef, pattern ) );
+ };
+
+nonterm block_or_single uses lang_stmt_list;
+
+block_or_single: '{' lang_stmt_list '}'
+ final {
+ $$->stmtList = $2->stmtList;
+ };
+block_or_single: statement
+ final {
+ $$->stmtList = new StmtList;
+ $$->stmtList->append( $1->stmt );
+ };
+
+nonterm iter_call
+{
+ LangTerm *langTerm;
+};
+
+iter_call: var_ref '(' opt_code_expr_list ')'
+ final {
+ $$->langTerm = new LangTerm( $1->varRef, $3->exprVect );
+ };
+iter_call: TK_Word
+ final {
+ $$->langTerm = new LangTerm( LangTerm::VarRefType,
+ new LangVarRef( $1->loc, new QualItemVect, $1->data ) );
+ };
+
+#
+# If Statements
+#
+
+nonterm if_stmt uses statement;
+
+if_stmt: KW_If scope_push code_expr block_or_single scope_pop elsif_list
+ final {
+ $$->stmt = new LangStmt( LangStmt::IfType, $3->expr, $4->stmtList, $6->stmt );
+ };
+
+nonterm elsif_list
+{
+ LangStmt *stmt;
+};
+
+elsif_list:
+ elsif_clause elsif_list
+ final {
+ /* Put any of the followng elseif part, an else, or null into the elsePart. */
+ $$->stmt = $1->stmt;
+ $$->stmt->elsePart = $2->stmt;
+ };
+elsif_list:
+ optional_else
+ final {
+ $$->stmt = $1->stmt;
+ };
+
+nonterm elsif_clause
+{
+ LangStmt *stmt;
+};
+
+elsif_clause:
+ KW_Elsif scope_push code_expr block_or_single scope_pop
+ final {
+ $$->stmt = new LangStmt( LangStmt::IfType, $3->expr, $4->stmtList, 0 );
+ };
+
+nonterm optional_else
+{
+ LangStmt *stmt;
+};
+
+optional_else:
+ KW_Else scope_push block_or_single scope_pop
+ final {
+ $$->stmt = new LangStmt( LangStmt::ElseType, $3->stmtList );
+ };
+
+optional_else:
+ final {
+ $$->stmt = 0;
+ };
+
+#
+# Code Expression Lists.
+#
+nonterm code_expr_list
+{
+ ExprVect *exprVect;
+};
+
+code_expr_list: code_expr_list code_expr
+ final {
+ $$->exprVect = $1->exprVect;
+ $$->exprVect->append( $2->expr );
+ };
+code_expr_list: code_expr
+ final {
+ $$->exprVect = new ExprVect;
+ $$->exprVect->append( $1->expr );
+ };
+
+nonterm opt_code_expr_list uses code_expr_list;
+
+opt_code_expr_list: code_expr_list
+ final {
+ $$->exprVect = $1->exprVect;
+ };
+
+opt_code_expr_list:
+ final {
+ $$->exprVect = 0;
+ };
+
+#
+# Type list
+#
+
+nonterm type_list
+{
+ TypeRefVect *typeRefVect;
+};
+
+type_list: type_list ',' type_ref
+ final {
+ $$->typeRefVect = $1->typeRefVect;
+ $$->typeRefVect->append( $3->typeRef );
+ };
+type_list: type_ref
+ final {
+ $$->typeRefVect = new TypeRefVect;
+ $$->typeRefVect->append( $1->typeRef );
+ };
+
+nonterm opt_type_list uses type_list;
+
+opt_type_list: type_list
+ final {
+ $$->typeRefVect = $1->typeRefVect;
+ };
+
+opt_type_list:
+ final {
+ $$->typeRefVect = 0;
+ };
+
+
+#
+# Variable reference
+#
+
+nonterm var_ref
+{
+ LangVarRef *varRef;
+};
+
+var_ref: qual TK_Word
+ final {
+ $$->varRef = new LangVarRef( $2->loc, $1->qual, $2->data );
+ };
+
+nonterm qual
+{
+ QualItemVect *qual;
+};
+
+qual: qual TK_Word '.'
+ final {
+ $$->qual = $1->qual;
+ $$->qual->append( QualItem( $2->loc, $2->data, QualItem::Dot ) );
+ };
+qual: qual TK_Word TK_RightArrow
+ final {
+ $$->qual = $1->qual;
+ $$->qual->append( QualItem( $2->loc, $2->data, QualItem::Arrow ) );
+ };
+qual:
+ final {
+ $$->qual = new QualItemVect;
+ };
+
+#
+# Code expression
+#
+
+nonterm code_expr
+{
+ LangExpr *expr;
+};
+
+code_expr: code_expr TK_AmpAmp code_relational
+ final {
+ $$->expr = new LangExpr( $2->loc, $1->expr, OP_LogicalAnd, $3->expr );
+ };
+
+code_expr: code_expr TK_BarBar code_relational
+ final {
+ $$->expr = new LangExpr( $2->loc, $1->expr, OP_LogicalOr, $3->expr );
+ };
+
+code_expr: code_relational
+ final {
+ $$->expr = $1->expr;
+ };
+
+nonterm code_relational uses code_expr;
+
+code_relational: code_relational TK_DoubleEql code_additive
+ final {
+ $$->expr = new LangExpr( $2->loc, $1->expr, OP_DoubleEql, $3->expr );
+ };
+
+code_relational: code_relational TK_NotEql code_additive
+ final {
+ $$->expr = new LangExpr( $2->loc, $1->expr, OP_NotEql, $3->expr );
+ };
+
+code_relational: code_relational '<' code_additive
+ final {
+ $$->expr = new LangExpr( $2->loc, $1->expr, '<', $3->expr );
+ };
+
+code_relational: code_relational '>' code_additive
+ final {
+ $$->expr = new LangExpr( $2->loc, $1->expr, '>', $3->expr );
+ };
+
+code_relational: code_relational TK_LessEql code_additive
+ final {
+ $$->expr = new LangExpr( $2->loc, $1->expr, OP_LessEql, $3->expr );
+ };
+
+code_relational: code_relational TK_GrtrEql code_additive
+ final {
+ $$->expr = new LangExpr( $2->loc, $1->expr, OP_GrtrEql, $3->expr );
+ };
+
+
+code_relational: code_additive
+ final {
+ $$->expr = $1->expr;
+ };
+
+nonterm code_additive uses code_expr;
+
+code_additive: code_additive '+' code_multiplicitive
+ final {
+ $$->expr = new LangExpr( $2->loc, $1->expr, '+', $3->expr );
+ };
+
+code_additive: code_additive '-' code_multiplicitive
+ final {
+ $$->expr = new LangExpr( $2->loc, $1->expr, '-', $3->expr );
+ };
+
+code_additive: code_multiplicitive
+ final {
+ $$->expr = $1->expr;
+ };
+
+nonterm code_multiplicitive uses code_expr;
+
+code_multiplicitive: code_multiplicitive '*' code_unary
+ final {
+ $$->expr = new LangExpr( $2->loc, $1->expr, '*', $3->expr );
+ };
+
+code_multiplicitive: code_multiplicitive '/' code_unary
+ final {
+ $$->expr = new LangExpr( $2->loc, $1->expr, '/', $3->expr );
+ };
+
+code_multiplicitive: code_unary
+ final {
+ $$->expr = $1->expr;
+ };
+
+nonterm code_unary uses code_expr;
+code_unary: '!' code_factor
+ final {
+ $$->expr = new LangExpr( $1->loc, '!', $2->expr );
+ };
+code_unary: '$' code_factor
+ final {
+ $$->expr = new LangExpr( $1->loc, '$', $2->expr );
+ };
+code_unary: '^' code_factor
+ final {
+ $$->expr = new LangExpr( $1->loc, '^', $2->expr );
+ };
+code_unary: '%' code_factor
+ final {
+ $$->expr = new LangExpr( $1->loc, '%', $2->expr );
+ };
+code_unary: code_factor
+ final {
+ $$->expr = $1->expr;
+ };
+
+nonterm opt_capture uses var_def;
+
+opt_capture: TK_Word ':'
+ final {
+ $$->objField = new ObjField( $1->loc, 0, $1->data );
+ };
+opt_capture:
+ final {
+ $$->objField = 0;
+ };
+
+nonterm code_factor uses code_expr;
+
+code_factor: TK_Number
+ final {
+ $$->expr = new LangExpr( new LangTerm( LangTerm::NumberType, $1->data ) );
+ };
+code_factor: TK_Literal
+ final {
+ $$->expr = new LangExpr( new LangTerm( LangTerm::StringType, $1->data ) );
+ };
+code_factor: var_ref '(' opt_code_expr_list ')'
+ final {
+ $$->expr = new LangExpr( new LangTerm( $1->varRef, $3->exprVect ) );
+ };
+code_factor: var_ref
+ final {
+ $$->expr = new LangExpr( new LangTerm( LangTerm::VarRefType, $1->varRef ) );
+ };
+code_factor: KW_Match var_ref pattern_list
+ final {
+ Namespace *nspace = namespaceStack.top();
+ TokenRegion *region = regionStack.top();
+ Pattern *pattern = new Pattern( $1->loc, nspace, region,
+ patternItemList, pd->nextPatReplId++ );
+ pd->patternList.append( pattern );
+
+ $$->expr = new LangExpr( new LangTerm( LangTerm::MatchType, $2->varRef, pattern ) );
+ };
+code_factor: KW_New code_factor
+ final {
+ $$->expr = new LangExpr( new LangTerm( LangTerm::NewType, $2->expr ) );
+ };
+code_factor:
+ KW_Construct opt_capture type_ref opt_field_init repl_list
+ final {
+ Namespace *nspace = namespaceStack.top();
+ TokenRegion *region = regionStack.top();
+ Replacement *replacement = new Replacement( $1->loc, nspace, region,
+ replItemList, pd->nextPatReplId++ );
+ pd->replList.append( replacement );
+
+ LangVarRef *varRef = 0;
+ if ( $2->objField != 0 )
+ varRef = new LangVarRef( $2->objField->loc, new QualItemVect, $2->objField->name );
+
+ $$->expr = new LangExpr( new LangTerm( $1->loc, LangTerm::ConstructType,
+ varRef, $2->objField, $3->typeRef, $4->fieldInitVect, replacement ) );
+
+ /* Check for redeclaration. */
+ if ( $2->objField != 0 ) {
+ if ( pd->curLocalFrame->checkRedecl( $2->objField->name ) != 0 ) {
+ error( $2->objField->loc ) << "variable " << $2->objField->name <<
+ " redeclared" << endp;
+ }
+
+ /* Insert it into the field map. */
+ $2->objField->typeRef = $3->typeRef;
+ pd->curLocalFrame->insertField( $2->objField->name, $2->objField );
+ }
+ };
+code_factor: KW_Parse opt_capture type_ref '(' opt_code_expr_list ')'
+ final {
+ String parserName = $3->typeRef->typeName + "_parser";
+
+ /* Get the language element. */
+ Namespace *nspace = namespaceStack.top();
+
+ GenericType *generic = 0;
+
+ NamespaceQual *nspaceQual = new NamespaceQual(
+ namespaceStack.top(), regionStack.top() );
+ TypeRef *parserTypeRef = new TypeRef( TypeRef::Parser,
+ InputLoc(), nspaceQual, $3->typeRef, 0 );
+
+ Replacement *replacement = new Replacement( $1->loc, nspace, pd->rootRegion,
+ new ReplItemList, pd->nextPatReplId++ );
+ pd->replList.append( replacement );
+
+ LangVarRef *varRef = 0;
+ if ( $2->objField != 0 )
+ varRef = new LangVarRef( $2->objField->loc, new QualItemVect, $2->objField->name );
+
+ $$->expr = new LangExpr( new LangTerm( $1->loc, LangTerm::ParseType,
+ varRef, $2->objField, $3->typeRef, generic, parserTypeRef, replacement ) );
+ $$->expr->term->args = $5->exprVect;
+
+ /* Check for redeclaration. */
+ if ( $2->objField != 0 ) {
+ if ( pd->curLocalFrame->checkRedecl( $2->objField->name ) != 0 ) {
+ error( $2->objField->loc ) << "variable " << $2->objField->name <<
+ " redeclared" << endp;
+ }
+
+ /* Insert it into the field map. */
+ $2->objField->typeRef = $3->typeRef;
+ pd->curLocalFrame->insertField( $2->objField->name, $2->objField );
+ }
+ };
+code_factor: KW_ParseStop opt_capture type_ref '(' opt_code_expr_list ')'
+ final {
+ /* This is a silly clone. To be fixed later. */
+ String parserName = $3->typeRef->typeName + "_parser";
+
+ /* Get the language element. */
+ Namespace *nspace = namespaceStack.top();
+
+ GenericType *generic = 0;
+
+ NamespaceQual *nspaceQual = new NamespaceQual(
+ namespaceStack.top(), regionStack.top() );
+ TypeRef *parserTypeRef = new TypeRef( TypeRef::Parser,
+ InputLoc(), nspaceQual, $3->typeRef, 0 );
+
+ Replacement *replacement = new Replacement( $1->loc, nspace, pd->rootRegion,
+ new ReplItemList, pd->nextPatReplId++ );
+ pd->replList.append( replacement );
+
+ LangVarRef *varRef = 0;
+ if ( $2->objField != 0 )
+ varRef = new LangVarRef( $2->objField->loc, new QualItemVect, $2->objField->name );
+
+ $$->expr = new LangExpr( new LangTerm( $1->loc, LangTerm::ParseStopType,
+ varRef, $2->objField, $3->typeRef, generic, parserTypeRef, replacement ) );
+ $$->expr->term->args = $5->exprVect;
+
+ /* Check for redeclaration. */
+ if ( $2->objField != 0 ) {
+ if ( pd->curLocalFrame->checkRedecl( $2->objField->name ) != 0 ) {
+ error( $2->objField->loc ) << "variable " << $2->objField->name <<
+ " redeclared" << endp;
+ }
+
+ /* Insert it into the field map. */
+ $2->objField->typeRef = $3->typeRef;
+ pd->curLocalFrame->insertField( $2->objField->name, $2->objField );
+ }
+
+ };
+code_factor: KW_TypeId '<' type_ref '>'
+ final {
+ $$->expr = new LangExpr( new LangTerm( $1->loc,
+ LangTerm::TypeIdType, $3->typeRef ) );
+ };
+code_factor: type_ref KW_In var_ref
+ final {
+ $$->expr = new LangExpr( new LangTerm( $2->loc,
+ LangTerm::SearchType, $1->typeRef, $3->varRef ) );
+ };
+code_factor: KW_Nil
+ final {
+ $$->expr = new LangExpr( new LangTerm( $1->loc,
+ LangTerm::NilType ) );
+ };
+code_factor: KW_True
+ final {
+ $$->expr = new LangExpr( new LangTerm( $1->loc,
+ LangTerm::TrueType ) );
+ };
+code_factor: KW_False
+ final {
+ $$->expr = new LangExpr( new LangTerm( $1->loc,
+ LangTerm::FalseType ) );
+ };
+code_factor: '(' code_expr ')'
+ final {
+ $$->expr = $2->expr;
+ };
+code_factor: KW_MakeTree '(' opt_code_expr_list ')'
+ final {
+ $$->expr = new LangExpr( new LangTerm( $1->loc,
+ LangTerm::MakeTreeType, $3->exprVect ) );
+ };
+code_factor: KW_MakeToken '(' opt_code_expr_list ')'
+ final {
+ $$->expr = new LangExpr( new LangTerm( $1->loc,
+ LangTerm::MakeTokenType, $3->exprVect ) );
+ };
+code_factor: KW_Deref code_expr
+ final {
+ $$->expr = new LangExpr( $1->loc, OP_Deref, $2->expr );
+ };
+code_factor: string_list
+ final {
+ $$->expr = new LangExpr( new LangTerm( replItemList ) );
+ };
+
+nonterm opt_field_init uses field_init_list;
+
+opt_field_init: '(' opt_field_init_list ')'
+ final {
+ $$->fieldInitVect = $2->fieldInitVect;
+ };
+opt_field_init:
+ final {
+ $$->fieldInitVect = 0;
+ };
+
+nonterm opt_field_init_list uses field_init_list;
+
+opt_field_init_list: field_init_list
+ final {
+ $$->fieldInitVect = $1->fieldInitVect;
+ };
+opt_field_init_list:
+ final {
+ $$->fieldInitVect = 0;
+ };
+
+nonterm field_init_list
+{
+ FieldInitVect *fieldInitVect;
+};
+
+field_init_list: field_init_list field_init
+ final {
+ $$->fieldInitVect = $1->fieldInitVect;
+ $$->fieldInitVect->append( $2->fieldInit );
+ };
+field_init_list: field_init
+ final {
+ $$->fieldInitVect = new FieldInitVect;
+ $$->fieldInitVect->append( $1->fieldInit );
+ };
+
+nonterm field_init
+{
+ FieldInit *fieldInit;
+};
+
+field_init: code_expr
+ final {
+ $$->fieldInit = new FieldInit( InputLoc(), "_name", $1->expr );
+ };
+
+#
+# Regular Expressions
+#
+
+nonterm opt_rl_join uses rl_join;
+
+opt_rl_join: rl_join opt_context
+ final {
+ $$->join = $1->join;
+ $$->context = $2->context;
+
+ if ( $2->context != 0 ) {
+ /* Create the enter and leaving actions that will mark the substring. */
+ Action *mark = new Action( MarkMark, pd->nextMatchEndNum++ );
+ pd->actionList.append( mark );
+
+ $$->join->context = $2->context;
+ $$->join->mark = mark;
+ }
+ };
+
+opt_rl_join:
+ final {
+ $$->join = 0;
+ $$->context = 0;
+ };
+
+nonterm rl_join
+{
+ Join *join;
+ Join *context;
+};
+
+rl_join:
+ rl_join ',' rl_expr
+ final {
+ /* Append the expression to the list and return it. */
+ $1->join->exprList.append( $3->expression );
+ $$->join = $1->join;
+ };
+rl_join:
+ rl_expr
+ final {
+ $$->join = new Join( $1->expression );
+ };
+
+# Context at the end of a pattern that is not included in the match
+nonterm opt_context uses rl_join;
+
+opt_context: '@' rl_join final { $$->context = $2->join; };
+opt_context: final { $$->context = 0; };
+
+nonterm rl_expr
+{
+ Expression *expression;
+};
+
+rl_expr:
+ rl_expr '|' rl_term_short final {
+ $$->expression = new Expression( $1->expression,
+ $3->term, Expression::OrType );
+ };
+rl_expr:
+ rl_expr '&' rl_term_short final {
+ $$->expression = new Expression( $1->expression,
+ $3->term, Expression::IntersectType );
+ };
+# This priority specification overrides the innermost parsing strategy which
+# results ordered choice interpretation of the grammar.
+rl_expr:
+ rl_expr '-' rl_term_short final {
+ $$->expression = new Expression( $1->expression,
+ $3->term, Expression::SubtractType );
+ };
+rl_expr:
+ rl_expr TK_DashDash rl_term_short final {
+ $$->expression = new Expression( $1->expression,
+ $3->term, Expression::StrongSubtractType );
+ };
+rl_expr:
+ rl_term_short final {
+ $$->expression = new Expression( $1->term );
+ };
+
+nonterm rl_term_short
+{
+ Term *term;
+};
+
+shortest rl_term_short;
+
+rl_term_short: rl_term
+ final { $$->term = $1->term; };
+
+nonterm rl_term
+{
+ Term *term;
+};
+
+rl_term:
+ rl_term factor_with_label final {
+ $$->term = new Term( $1->term, $2->factorWithAug );
+ };
+rl_term:
+ rl_term '.' factor_with_label final {
+ $$->term = new Term( $1->term, $3->factorWithAug );
+ };
+rl_term:
+ rl_term TK_ColonGt factor_with_label final {
+ $$->term = new Term( $1->term, $3->factorWithAug, Term::RightStartType );
+ };
+rl_term:
+ rl_term TK_ColonGtGt factor_with_label final {
+ $$->term = new Term( $1->term, $3->factorWithAug, Term::RightFinishType );
+ };
+rl_term:
+ rl_term TK_LtColon factor_with_label final {
+ $$->term = new Term( $1->term,
+ $3->factorWithAug, Term::LeftType );
+ };
+rl_term:
+ factor_with_label final {
+ $$->term = new Term( $1->factorWithAug );
+ };
+
+nonterm factor_with_label
+{
+ FactorWithAug *factorWithAug;
+};
+
+factor_with_label:
+ factor_with_ep final {
+ $$->factorWithAug = $1->factorWithAug;
+ };
+
+factor_with_label:
+ TK_Word ':' factor_with_label final {
+ $$->factorWithAug = $3->factorWithAug;
+
+ if ( pd->objectDef->checkRedecl( $1->data ) != 0 )
+ error($1->loc) << "label name \"" << $1->data << "\" already in use" << endp;
+
+ /* Create the object field. */
+ NamespaceQual *qual = new NamespaceQual( namespaceStack.top(), regionStack.top() );
+ TypeRef *typeRef = new TypeRef( $1->loc, qual, "str" );
+ ObjField *objField = new ObjField( $1->loc, typeRef, $1->data );
+
+ /* Insert it into the map. */
+ pd->objectDef->insertField( $1->data, objField );
+
+ /* Create the enter and leaving actions that will mark the substring. */
+ Action *enter = new Action( MarkMark, pd->nextMatchEndNum++ );
+ Action *leave = new Action( MarkMark, pd->nextMatchEndNum++ );
+ pd->actionList.append( enter );
+ pd->actionList.append( leave );
+
+ /* Add entering and leaving actions. */
+ $$->factorWithAug->actions.append( ParserAction( $1->loc, at_start, 0, enter ) );
+ $$->factorWithAug->actions.append( ParserAction( $1->loc, at_leave, 0, leave ) );
+
+ reCaptureVect.append( ReCapture( enter, leave, objField ) );
+ };
+
+nonterm factor_with_ep
+{
+ FactorWithAug *factorWithAug;
+};
+
+factor_with_ep:
+ factor_with_aug final {
+ $$->factorWithAug = $1->factorWithAug;
+ };
+
+nonterm factor_with_aug
+{
+ FactorWithAug *factorWithAug;
+};
+
+factor_with_aug:
+ factor_with_rep final {
+ $$->factorWithAug = new FactorWithAug( $1->factorWithRep );
+ };
+
+
+# The fourth level of precedence. These are the trailing unary operators that
+# allow for repetition.
+
+nonterm factor_with_rep
+{
+ FactorWithRep *factorWithRep;
+};
+
+factor_with_rep:
+ factor_with_rep '*' final {
+ $$->factorWithRep = new FactorWithRep( $2->loc, $1->factorWithRep,
+ 0, 0, FactorWithRep::StarType );
+ };
+factor_with_rep:
+ factor_with_rep TK_StarStar final {
+ $$->factorWithRep = new FactorWithRep( $2->loc, $1->factorWithRep,
+ 0, 0, FactorWithRep::StarStarType );
+ };
+factor_with_rep:
+ factor_with_rep '?' final {
+ $$->factorWithRep = new FactorWithRep( $2->loc, $1->factorWithRep,
+ 0, 0, FactorWithRep::OptionalType );
+ };
+factor_with_rep:
+ factor_with_rep '+' final {
+ $$->factorWithRep = new FactorWithRep( $2->loc, $1->factorWithRep,
+ 0, 0, FactorWithRep::PlusType );
+ };
+factor_with_rep:
+ factor_with_rep '{' factor_rep_num '}' final {
+ $$->factorWithRep = new FactorWithRep( $2->loc, $1->factorWithRep,
+ $3->rep, 0, FactorWithRep::ExactType );
+ };
+factor_with_rep:
+ factor_with_rep '{' ',' factor_rep_num '}' final {
+ $$->factorWithRep = new FactorWithRep( $2->loc, $1->factorWithRep,
+ 0, $4->rep, FactorWithRep::MaxType );
+ };
+factor_with_rep:
+ factor_with_rep '{' factor_rep_num ',' '}' final {
+ $$->factorWithRep = new FactorWithRep( $2->loc, $1->factorWithRep,
+ $3->rep, 0, FactorWithRep::MinType );
+ };
+factor_with_rep:
+ factor_with_rep '{' factor_rep_num ',' factor_rep_num '}' final {
+ $$->factorWithRep = new FactorWithRep( $2->loc, $1->factorWithRep,
+ $3->rep, $5->rep, FactorWithRep::RangeType );
+ };
+factor_with_rep:
+ factor_with_neg final {
+ $$->factorWithRep = new FactorWithRep(
+ $1->factorWithNeg->loc, $1->factorWithNeg );
+ };
+
+nonterm factor_rep_num
+{
+ int rep;
+};
+
+factor_rep_num:
+ TK_UInt final {
+ // Convert the priority number to a long. Check for overflow.
+ errno = 0;
+ int rep = strtol( $1->data, 0, 10 );
+ if ( errno == ERANGE && rep == LONG_MAX ) {
+ // Repetition too large. Recover by returing repetition 1. */
+ error($1->loc) << "repetition number " << $1->data << " overflows" << endl;
+ $$->rep = 1;
+ }
+ else {
+ // Cannot be negative, so no overflow.
+ $$->rep = rep;
+ }
+ };
+
+
+#
+# The fifth level up in precedence. Negation.
+#
+
+nonterm factor_with_neg
+{
+ FactorWithNeg *factorWithNeg;
+};
+
+factor_with_neg:
+ '!' factor_with_neg final {
+ $$->factorWithNeg = new FactorWithNeg( $1->loc,
+ $2->factorWithNeg, FactorWithNeg::NegateType );
+ };
+factor_with_neg:
+ '^' factor_with_neg final {
+ $$->factorWithNeg = new FactorWithNeg( $1->loc,
+ $2->factorWithNeg, FactorWithNeg::CharNegateType );
+ };
+factor_with_neg:
+ rl_factor final {
+ $$->factorWithNeg = new FactorWithNeg( $1->factor->loc, $1->factor );
+ };
+
+nonterm rl_factor
+{
+ Factor *factor;
+};
+
+rl_factor:
+ TK_Literal final {
+ /* Create a new factor node going to a concat literal. */
+ $$->factor = new Factor( new Literal( $1->loc, $1->data, Literal::LitString ) );
+ };
+rl_factor:
+ alphabet_num final {
+ /* Create a new factor node going to a literal number. */
+ $$->factor = new Factor( new Literal( $1->loc,
+ $1->data, Literal::Number ) );
+ };
+rl_factor:
+ TK_Word final {
+ /* Find the named graph. */
+ Namespace *nspace = namespaceStack.top();
+
+ while ( nspace != 0 ) {
+ GraphDictEl *gdNode = nspace->rlMap.find( $1->data );
+ if ( gdNode != 0 ) {
+ if ( gdNode->isInstance ) {
+ /* Recover by retuning null as the factor node. */
+ error($1->loc) << "references to graph instantiations not allowed "
+ "in expressions" << endl;
+ $$->factor = 0;
+ }
+ else {
+ /* Create a factor node that is a lookup of an expression. */
+ $$->factor = new Factor( $1->loc, gdNode->value );
+ }
+ break;
+ }
+
+ nspace = nspace->parentNamespace;
+ }
+
+ if ( nspace == 0 ) {
+ /* Recover by returning null as the factor node. */
+ error($1->loc) << "graph lookup of \"" << $1->data << "\" failed" << endl;
+ $$->factor = 0;
+ }
+ };
+rl_factor:
+ TK_SqOpen regular_expr_or_data TK_SqClose final {
+ /* Create a new factor node going to an OR expression. */
+ $$->factor = new Factor( new ReItem( $1->loc, $2->reOrBlock, ReItem::OrBlock ) );
+ };
+rl_factor:
+ TK_SqOpenNeg regular_expr_or_data TK_SqClose final {
+ /* Create a new factor node going to a negated OR expression. */
+ $$->factor = new Factor( new ReItem( $1->loc, $2->reOrBlock, ReItem::NegOrBlock ) );
+ };
+rl_factor:
+ range_lit TK_DotDot range_lit final {
+ /* Create a new factor node going to a range. */
+ $$->factor = new Factor( new Range( $1->literal, $3->literal ) );
+ };
+rl_factor:
+ '(' rl_join ')' final {
+ /* Create a new factor going to a parenthesized join. */
+ $$->factor = new Factor( $2->join );
+ };
+
+nonterm range_lit
+{
+ Literal *literal;
+};
+
+# Literals which can be the end points of ranges.
+range_lit:
+ TK_Literal final {
+ /* Range literas must have only one char. We restrict this in the parse tree. */
+ $$->literal = new Literal( $1->loc, $1->data, Literal::LitString );
+ };
+range_lit:
+ alphabet_num final {
+ /* Create a new literal number. */
+ $$->literal = new Literal( $1->loc, $1->data, Literal::Number );
+ };
+
+nonterm alphabet_num uses token_data;
+
+# Any form of a number that can be used as a basic machine. */
+alphabet_num:
+ TK_UInt final {
+ $$->loc = $1->loc;
+ $$->data = $1->data;
+ };
+alphabet_num:
+ '-' TK_UInt final {
+ $$->loc = $1->loc;
+ $$->data = '+';
+ $$->data += $2->data;
+ };
+alphabet_num:
+ TK_Hex final {
+ $$->loc = $1->loc;
+ $$->data = $1->data;
+ };
+
+#
+# Regular Expressions.
+#
+
+
+# The data inside of a [] expression in a regular expression. Accepts any
+# number of characters or ranges. */
+nonterm regular_expr_or_data
+{
+ ReOrBlock *reOrBlock;
+};
+
+regular_expr_or_data:
+ regular_expr_or_data regular_expr_or_char final {
+ /* An optimization to lessen the tree size. If an or char is directly
+ * under the left side on the right and the right side is another or
+ * char then paste them together and return the left side. Otherwise
+ * just put the two under a new or data node. */
+ if ( $2->reOrItem->type == ReOrItem::Data &&
+ $1->reOrBlock->type == ReOrBlock::RecurseItem &&
+ $1->reOrBlock->item->type == ReOrItem::Data )
+ {
+ /* Append the right side to right side of the left and toss the
+ * right side. */
+ $1->reOrBlock->item->data += $2->reOrItem->data;
+ delete $2->reOrItem;
+ $$->reOrBlock = $1->reOrBlock;
+ }
+ else {
+ /* Can't optimize, put the left and right under a new node. */
+ $$->reOrBlock = new ReOrBlock( $1->reOrBlock, $2->reOrItem );
+ }
+ };
+regular_expr_or_data:
+ final {
+ $$->reOrBlock = new ReOrBlock();
+ };
+
+# A single character inside of an or expression. Can either be a character or a
+# set of characters.
+nonterm regular_expr_or_char
+{
+ ReOrItem *reOrItem;
+};
+
+regular_expr_or_char:
+ TK_ReChar final {
+ $$->reOrItem = new ReOrItem( $1->loc, $1->data );
+ };
+regular_expr_or_char:
+ TK_ReChar TK_Dash TK_ReChar final {
+ $$->reOrItem = new ReOrItem( $2->loc, $1->data[0], $3->data[0] );
+ };
+
+# A local state reference. Cannot have :: prefix.
+local_state_ref:
+ no_name_sep state_ref_names;
+
+# Clear the name ref structure.
+no_name_sep:
+ final {
+ nameRef.empty();
+ };
+
+# A qualified state reference.
+state_ref: opt_name_sep state_ref_names;
+
+# Optional leading name separator.
+opt_name_sep:
+ TK_NameSep
+ final {
+ /* Insert an initial null pointer val to indicate the existence of the
+ * initial name seperator. */
+ nameRef.setAs( 0 );
+ };
+opt_name_sep:
+ final {
+ nameRef.empty();
+ };
+
+# List of names separated by ::
+state_ref_names:
+ state_ref_names TK_NameSep TK_Word
+ final {
+ nameRef.append( $3->data );
+ };
+state_ref_names:
+ TK_Word
+ final {
+ nameRef.append( $1->data );
+ };
+
+nonterm opt_commit
+{
+ bool commit;
+};
+
+opt_commit: final { $$->commit = false; };
+opt_commit: KW_Commit final { $$->commit = true; };
+
+#
+# Grammar Finished
+#
+
+ write types;
+ write data;
+}%%
+
+void ColmParser::init()
+{
+ /* Set up the root namespace. */
+ const char *rootNamespaceName = "___ROOT_NAMESPACE";
+ Namespace *rootNamespace = new Namespace( InputLoc(),
+ rootNamespaceName, pd->namespaceList.length(), 0 );
+ pd->namespaceList.append( rootNamespace );
+ namespaceStack.push( rootNamespace );
+ pd->rootNamespace = rootNamespace;
+
+ /* Set up the root token region. */
+ const char *rootRegionName = "___ROOT_REGION";
+
+ TokenRegion *rootRegion = new TokenRegion( InputLoc(), rootRegionName,
+ pd->regionList.length(), 0 );
+ pd->regionList.append( rootRegion );
+ addRegionDef( InputLoc(), namespaceStack.top(), rootRegionName, rootRegion );
+
+ regionStack.push( rootRegion );
+
+ pd->rootRegion = rootRegion;
+
+ /* Set up the global object. */
+ String global = "global";
+ pd->globalObjectDef = new ObjectDef( ObjectDef::UserType,
+ global, pd->nextObjectId++ );
+
+ /* The eofTokenRegion defaults to the root region. */
+ pd->eofTokenRegion = rootRegion;
+
+ /* Initialize the dictionary of graphs. This is our symbol table. The
+ * initialization needs to be done on construction which happens at the
+ * beginning of a machine spec so any assignment operators can reference
+ * the builtins. */
+ pd->initGraphDict();
+
+ pd->rootLocalFrame = new ObjectDef( ObjectDef::FrameType,
+ "local", pd->nextObjectId++ );
+ pd->curLocalFrame = pd->rootLocalFrame;
+
+ %% write init;
+
+ addArgvList();
+}
+
+void ColmParser::addArgvList()
+{
+ NamespaceQual *nspaceQual1 = new NamespaceQual(
+ namespaceStack.top(), regionStack.top() );
+ TypeRef *typeRef = new TypeRef( InputLoc(), nspaceQual1, "str" );
+
+ NamespaceQual *nspaceQual2 = new NamespaceQual(
+ namespaceStack.top(), regionStack.top() );
+
+ pd->argvTypeRef = new TypeRef( TypeRef::List, InputLoc(),
+ nspaceQual2, typeRef, 0 );
+}
+
+int ColmParser::parseLangEl( int type, const Token *token )
+{
+ %% write exec;
+ return errCount == 0 ? 0 : -1;
+}
+
+void ColmParser::addRegularDef( const InputLoc &loc, Namespace *nspace,
+ const String &name, Join *join )
+{
+ GraphDictEl *newEl = nspace->rlMap.insert( name );
+ if ( newEl != 0 ) {
+ /* New element in the dict, all good. */
+ newEl->value = new VarDef( name, join );
+ newEl->isInstance = false;
+ newEl->loc = loc;
+ }
+ else {
+ // Recover by ignoring the duplicate.
+ error(loc) << "regular definition \"" << name << "\" already exists" << endl;
+ }
+}
+
+TokenRegion *ColmParser::createRegion( String &scannerName )
+{
+ TokenRegion *tokenRegion = new TokenRegion( InputLoc(), scannerName,
+ pd->regionList.length(), regionStack.top() );
+
+ regionStack.top()->childRegions.append( tokenRegion );
+
+ pd->regionList.append( tokenRegion );
+
+ addRegionDef( InputLoc(), namespaceStack.top(), scannerName, tokenRegion );
+
+ return tokenRegion;
+}
+
+
+void ColmParser::addRegionDef( const InputLoc &loc, Namespace *nspace,
+ const String &name, TokenRegion *tokenRegion )
+{
+ RegionGraphDictEl *newEl = nspace->graphDict.insert( name );
+ if ( newEl != 0 ) {
+ /* New element in the dict, all good. */
+ newEl->value = new RegionDef( name, tokenRegion );
+ newEl->isInstance = true;
+ newEl->loc = loc;
+
+ /* It it is an instance, put on the instance list. */
+ pd->instanceList.append( newEl );
+ }
+ else {
+ // Recover by ignoring the duplicate.
+ error(loc) << "regular definition \"" << name << "\" already exists" << endl;
+ }
+}
+
+ostream &ColmParser::parse_error( int tokId, Token &token )
+{
+ /* Maintain the error count. */
+ gblErrorCount += 1;
+
+ cerr << token.loc.fileName << ":" << token.loc.line << ":" << token.loc.col << ": ";
+ cerr << "at token ";
+ if ( tokId < 128 )
+ cerr << "\"" << ColmParser_lelNames[tokId] << "\"";
+ else
+ cerr << ColmParser_lelNames[tokId];
+ if ( token.data != 0 )
+ cerr << " with data \"" << token.data << "\"";
+ cerr << ": ";
+
+ return cerr;
+}
+
+int ColmParser::token( InputLoc &loc, int tokId, char *tokstart, int toklen )
+{
+ Token token;
+
+ if ( toklen > 0 )
+ token.data.setAs( tokstart, toklen );
+
+ token.loc = loc;
+ int res = parseLangEl( tokId, &token );
+ if ( res < 0 ) {
+ parse_error(tokId, token) << "parse error" << endl;
+ exit(1);
+ }
+ return res;
+}