testing left-recursive lists <* and <+ and sublists in productions

Using a transformation pass in prog.lm to create definitions for left-recursive
productions and sub-productions. So far just a sketch. Not all production
element forms are transformed and there can only be one sub-production in a
whole program. refs #74 refs #83 refs #84

5 files changed, 88 insertions, 4 deletions

diff --git a/colm/Makefile.am b/colm/Makefile.am
index ad2a78b2..fe594521 100644
--- a/colm/Makefile.am
+++ b/colm/Makefile.am
@@ -154,7 +154,7 @@ bootstrap2_LDADD = libprog.a libcolm.la
 
 endif
 
-tar/bootstrap3.tar: $(BUILD_PARSE_3_WITH) prog.lm
+tar/bootstrap3.tar: $(BUILD_PARSE_3_WITH) prog.lm colm.lm
 	mkdir -p gen
 	$(BUILD_PARSE_3_WITH) -c -o tar/parse3.c -e tar/if3.h -x tar/if3.cc prog.lm
 	tar -C tar -cf tar/bootstrap3.tar parse3.c if3.h if3.cc
diff --git a/colm/colm.lm b/colm/colm.lm
index 1c475adc..efbd7a65 100644
--- a/colm/colm.lm
+++ b/colm/colm.lm
@@ -472,9 +472,14 @@ def opt_id
 def ignore_def
 	[IGNORE opt_id LEX_FSLASH opt_lex_expr LEX_FSLASH]
 
+def prod_sublist
+	[prod_sublist BAR prod_el_list]
+|	[prod_el_list]
+
 def prod_el
 	[opt_prod_el_name region_qual id opt_repeat] :Id
 |	[opt_prod_el_name region_qual backtick_lit opt_repeat] :Lit
+|	[opt_prod_el_name POPEN prod_sublist PCLOSE opt_repeat ] :SubList
 
 def opt_prod_el_name
 	[id COLON] :Name
@@ -666,6 +671,8 @@ def opt_repeat
 	[STAR] :Star
 |	[PLUS] :Plus
 |	[QUESTION] :Question
+|	[LT STAR] :LeftStar
+|	[LT PLUS] :LeftPlus
 |	[]
 
 def opt_capture
diff --git a/colm/ctinput.cc b/colm/ctinput.cc
index 2adf2a94..bca2080a 100644
--- a/colm/ctinput.cc
+++ b/colm/ctinput.cc
@@ -459,7 +459,7 @@ int repl_undo_consume_data( struct colm_program *prg, struct input_impl_ct *ss,
 
 		/* Okay to go up to the front of the buffer. */
 		if ( length > avail ) {
-			ss->cons_item= ss->cons_item->prev;
+			ss->cons_item = ss->cons_item->prev;
 			ss->offset = ss->cons_item->data.length();
 			length -= avail;
 		}
diff --git a/colm/loadcolm.cc b/colm/loadcolm.cc
index 694a88fd..ac64f668 100644
--- a/colm/loadcolm.cc
+++ b/colm/loadcolm.cc
@@ -844,6 +844,11 @@ struct LoadColm
 		case opt_repeat::Question:
 			repeatType = RepeatOpt;
 			break;
+		case opt_repeat::LeftStar:
+		case opt_repeat::LeftPlus:
+			error( OptRepeat.loc() ) << "<* and <+ are implemented as a "
+					"colm transformation, they are not accepted at this stage" << endp;
+			break;
 		}
 		return repeatType;
 	}
@@ -972,10 +977,10 @@ struct LoadColm
 		}
 
 		RepeatType repeatType = walkOptRepeat( El.opt_repeat() );
-		NamespaceQual *nspaceQual = walkRegionQual( El.region_qual() );
-
 		switch ( El.prodName() ) {
 		case prod_el::Id: {
+			NamespaceQual *nspaceQual = walkRegionQual( El.region_qual() );
+
 			String typeName = El.id().data();
 			ProdEl *prodEl = prodElName( El.id().loc(), typeName,
 					nspaceQual, captureField, repeatType, false );
@@ -983,11 +988,17 @@ struct LoadColm
 			break;
 		}
 		case prod_el::Lit: {
+			NamespaceQual *nspaceQual = walkRegionQual( El.region_qual() );
+
 			String lit = El.backtick_lit().data();
 			ProdEl *prodEl = prodElLiteral( El.backtick_lit().loc(), lit,
 					nspaceQual, captureField, repeatType, false );
 			appendProdEl( list, prodEl );
 			break;
+		}
+		case prod_el::SubList: {
+			error( El.POPEN().loc() ) << "production sublist is implemented as a "
+					"colm transformation, it is not accepted at this stage" << endp;
 		}}
 	}
 
diff --git a/colm/prog.lm b/colm/prog.lm
index 43525993..e0ac8c22 100644
--- a/colm/prog.lm
+++ b/colm/prog.lm
@@ -7,5 +7,71 @@ A: str = argv->pop()
 F: stream = open( A, 'r' )
 parse P: start [ F ]
 
+def new_roots
+	[new_roots root_item]
+|	[]
+
+prod_list cons_prod( SLA: prod_sublist )
+{
+	if match SLA [Left: prod_sublist BAR prod_el_list]
+		return cons prod_list[ cons_prod(Left) ' | [ ' SLA.prod_el_list ' ] ' ]
+	else
+		return cons prod_list[ '[ ' SLA.prod_el_list ' ]' ]
+}
+
+void rewrite( P: ref<start> )
+{
+	Modified: bool = false
+	cons Additional: new_roots []
+
+	for PE: prod_el in P {
+		if match PE
+				[OptName: opt_prod_el_name POPEN PS: prod_sublist PCLOSE OptRep: opt_repeat]
+		{
+			PE = cons prod_el
+				[OptName "_sublist_" OptRep]
+
+			cons Def: cfl_def
+				"def _sublist_ [cons_prod(PS)]"
+
+			Additional = cons new_roots
+				[Additional Def]
+		}
+		else if match PE
+				[OptName: opt_prod_el_name Qual: region_qual Id: id "<*"]
+		{
+			parse LeftRepId: id "_lrep_[Id]"
+			PE = cons prod_el [OptName Qual LeftRepId]
+
+			cons Def: cfl_def "def [LeftRepId] \[[LeftRepId] [Id]\] | \[\]
+
+			Additional = cons new_roots [Additional Def]
+		}
+		else if match PE
+				[OptName: opt_prod_el_name Qual: region_qual Id: id "<+"]
+		{
+			parse LeftRepId: id "_lrep_[Id]"
+
+			PE = cons prod_el [OptName Qual LeftRepId]
+
+			cons Def: cfl_def "def [LeftRepId] \[[LeftRepId] [Id]\] | \[[Id]\]
+
+			Additional = cons new_roots [Additional Def]
+		}
+	}
+
+	while  ( Additional.root_item ) {
+		P = cons start [Additional.root_item P.RootItemList]
+		Additional = Additional._new_roots
+		Modified = true
+	}
+	return Modified
+}
+
+if P {
+	while ( rewrite( P ) ) {}
+	#rewrite( P )
+}
+
 ColmTree = P
 ColmError = error