make location file name and column available in Colm code

7 files changed, 86 insertions, 24 deletions

diff --git a/src/bytecode.c b/src/bytecode.c
index a0cfdfac..74d1a25f 100644
--- a/src/bytecode.c
+++ b/src/bytecode.c
@@ -3093,25 +3093,49 @@ again:
 			colm_tree_downref( prg, sp, tree );
 			break;
 		}
-		case IN_GET_TOKEN_POS_R: {
-			debug( prg, REALM_BYTECODE, "IN_GET_TOKEN_POS_R\n" );
+		case IN_GET_TOKEN_FILE_R: {
+			debug( prg, REALM_BYTECODE, "IN_GET_TOKEN_FILE_R\n" );
+			tree_t *tree = vm_pop_tree();
+			tree_t *str = 0;
+			if ( tree->tokdata->location ) {
+				const char *fn = tree->tokdata->location->name;
+				head_t *data = string_alloc_full( prg, fn, strlen(fn) );
+				str = construct_string( prg, data );
+				colm_tree_upref( prg, str );
+			}
+			vm_push_tree( str );
+			colm_tree_downref( prg, sp, tree );
+			break;
+		}
+		case IN_GET_TOKEN_LINE_R: {
+			debug( prg, REALM_BYTECODE, "IN_GET_TOKEN_LINE_R\n" );
 
 			tree_t *tree = vm_pop_tree();
 			value_t integer = 0;
 			if ( tree->tokdata->location )
-				integer = tree->tokdata->location->byte;
+				integer = tree->tokdata->location->line;
 			vm_push_value( integer );
 			colm_tree_downref( prg, sp, tree );
 			break;
 		}
-		case IN_GET_TOKEN_LINE_R: {
-			debug( prg, REALM_BYTECODE, "IN_GET_TOKEN_LINE_R\n" );
+		case IN_GET_TOKEN_COL_R: {
+			debug( prg, REALM_BYTECODE, "IN_GET_TOKEN_COL_R\n" );
 
 			tree_t *tree = vm_pop_tree();
 			value_t integer = 0;
 			if ( tree->tokdata->location )
-				integer = tree->tokdata->location->line;
+				integer = tree->tokdata->location->column;
+			vm_push_value( integer );
+			colm_tree_downref( prg, sp, tree );
+			break;
+		}
+		case IN_GET_TOKEN_POS_R: {
+			debug( prg, REALM_BYTECODE, "IN_GET_TOKEN_POS_R\n" );
 
+			tree_t *tree = vm_pop_tree();
+			value_t integer = 0;
+			if ( tree->tokdata->location )
+				integer = tree->tokdata->location->byte;
 			vm_push_value( integer );
 			colm_tree_downref( prg, sp, tree );
 			break;
diff --git a/src/bytecode.h b/src/bytecode.h
index 5a37fdea..8793c909 100644
--- a/src/bytecode.h
+++ b/src/bytecode.h
@@ -142,8 +142,10 @@ typedef unsigned char uchar;
 #define IN_SET_TOKEN_DATA_WV     0x38
 #define IN_SET_TOKEN_DATA_BKT    0x39
 
-#define IN_GET_TOKEN_POS_R       0x3a
+#define IN_GET_TOKEN_FILE_R      0x80
 #define IN_GET_TOKEN_LINE_R      0x3b
+#define IN_GET_TOKEN_POS_R       0x3a
+#define IN_GET_TOKEN_COL_R       0x81
 
 #define IN_INIT_RHS_EL           0x3c
 #define IN_INIT_LHS_EL           0x3d
diff --git a/src/compiler.h b/src/compiler.h
index 46c33952..9ad38ee2 100644
--- a/src/compiler.h
+++ b/src/compiler.h
@@ -1012,8 +1012,10 @@ struct Compiler
 	CodeVect unwindCode;
 
 	ObjectField *makeDataEl();
-	ObjectField *makePosEl();
+	ObjectField *makeFileEl();
 	ObjectField *makeLineEl();
+	ObjectField *makeColEl();
+	ObjectField *makePosEl();
 
 	IterDef *findIterDef( IterDef::Type type, GenericType *generic );
 	IterDef *findIterDef( IterDef::Type type, Function *func );
diff --git a/src/declare.cc b/src/declare.cc
index 6e92bdf0..d521e7c6 100644
--- a/src/declare.cc
+++ b/src/declare.cc
@@ -926,22 +926,22 @@ ObjectField *Compiler::makeDataEl()
 	return el;
 }
 
-ObjectField *Compiler::makePosEl()
+ObjectField *Compiler::makeFileEl()
 {
-	/* Create the "data" field. */
-	TypeRef *typeRef = TypeRef::cons( internal, uniqueTypeInt );
+	/* Create the "file" field. */
+	TypeRef *typeRef = TypeRef::cons( internal, uniqueTypeStr );
 	ObjectField *el = ObjectField::cons( internal,
-			ObjectField::InbuiltFieldType, typeRef, "pos" );
+			ObjectField::InbuiltFieldType, typeRef, "file" );
 
 	el->isConst = true;
-	el->inGetR = IN_GET_TOKEN_POS_R;
-	el->inGetValR = IN_GET_TOKEN_POS_R;
+	el->inGetR = IN_GET_TOKEN_FILE_R;
+	el->inGetValR = IN_GET_TOKEN_FILE_R;
 	return el;
 }
 
 ObjectField *Compiler::makeLineEl()
 {
-	/* Create the "data" field. */
+	/* Create the "line" field. */
 	TypeRef *typeRef = TypeRef::cons( internal, uniqueTypeInt );
 	ObjectField *el = ObjectField::cons( internal,
 			ObjectField::InbuiltFieldType, typeRef, "line" );
@@ -952,6 +952,32 @@ ObjectField *Compiler::makeLineEl()
 	return el;
 }
 
+ObjectField *Compiler::makeColEl()
+{
+	/* Create the "col" field. */
+	TypeRef *typeRef = TypeRef::cons( internal, uniqueTypeInt );
+	ObjectField *el = ObjectField::cons( internal,
+			ObjectField::InbuiltFieldType, typeRef, "col" );
+
+	el->isConst = true;
+	el->inGetR = IN_GET_TOKEN_COL_R;
+	el->inGetValR = IN_GET_TOKEN_COL_R;
+	return el;
+}
+
+ObjectField *Compiler::makePosEl()
+{
+	/* Create the "data" field. */
+	TypeRef *typeRef = TypeRef::cons( internal, uniqueTypeInt );
+	ObjectField *el = ObjectField::cons( internal,
+			ObjectField::InbuiltFieldType, typeRef, "pos" );
+
+	el->isConst = true;
+	el->inGetR = IN_GET_TOKEN_POS_R;
+	el->inGetValR = IN_GET_TOKEN_POS_R;
+	return el;
+}
+
 /* Add a constant length field to the object. 
  * Opcode supplied by the caller. */
 void Compiler::addLengthField( ObjectDef *objDef, code_t getLength )
@@ -977,13 +1003,21 @@ void Compiler::declareTokenFields( )
 				ObjectField *dataEl = makeDataEl();
 				lel->objectDef->rootScope->insertField( dataEl->name, dataEl );
 
-				/* Create the "pos" field. */
-				ObjectField *posEl = makePosEl();
-				lel->objectDef->rootScope->insertField( posEl->name, posEl );
+				/* Create the "file" field. */
+				ObjectField *fileEl = makeFileEl();
+				lel->objectDef->rootScope->insertField( fileEl->name, fileEl );
 
 				/* Create the "line" field. */
 				ObjectField *lineEl = makeLineEl();
 				lel->objectDef->rootScope->insertField( lineEl->name, lineEl );
+
+				/* Create the "col" field. */
+				ObjectField *colEl = makeColEl();
+				lel->objectDef->rootScope->insertField( colEl->name, colEl );
+
+				/* Create the "pos" field. */
+				ObjectField *posEl = makePosEl();
+				lel->objectDef->rootScope->insertField( posEl->name, posEl );
 			}
 		}
 	}
diff --git a/src/input.c b/src/input.c
index 99e65ec5..a0b7a337 100644
--- a/src/input.c
+++ b/src/input.c
@@ -475,7 +475,7 @@ static void input_prepend_data( struct colm_program *prg, struct input_impl_seq
 
 	maybe_split( prg, si );
 
-	struct stream_impl *sub_si = colm_impl_new_text( "<text>", data, length );
+	struct stream_impl *sub_si = colm_impl_new_text( "<text1>", data, length );
 
 	struct seq_buf *new_buf = new_seq_buf();
 	new_buf->type = SB_ACCUM;
@@ -557,7 +557,7 @@ static void input_append_data( struct colm_program *prg, struct input_impl_seq *
 	if ( si->queue.tail == 0 || si->queue.tail->type != SB_ACCUM ) { 
 		debug( prg, REALM_INPUT, "input_append_data: creating accum\n" );
 
-		struct stream_impl *sub_si = colm_impl_new_accum( "<text>" );
+		struct stream_impl *sub_si = colm_impl_new_accum( "<text2>" );
 
 		struct seq_buf *new_buf = new_seq_buf();
 		new_buf->type = SB_ACCUM;
diff --git a/src/stream.c b/src/stream.c
index ee939fff..6fb3901b 100644
--- a/src/stream.c
+++ b/src/stream.c
@@ -234,7 +234,7 @@ static struct stream_impl *data_split_consumed( program_t *prg, struct stream_im
 	struct stream_impl *split_off = 0;
 	if ( sid->consumed > 0 ) {
 		debug( prg, REALM_INPUT, "maybe split: consumed is > 0, splitting\n" );
-		split_off = colm_impl_consumed( "<text>", sid->consumed );
+		split_off = colm_impl_consumed( "<text3>", sid->consumed );
 		sid->consumed = 0;
 	}
 	return split_off;
diff --git a/src/string.c b/src/string.c
index 6e0c7c38..8a852e8b 100644
--- a/src/string.c
+++ b/src/string.c
@@ -55,9 +55,9 @@ tree_t *construct_string( program_t *prg, head_t *s )
 
 /* 
  * In this system strings are not null terminated. Often strings come from a
- * parse, in which case the string is just a pointer into the the data string.
- * A block in a parsed stream can house many tokens and there is no room for
- * nulls.
+ * parse, in which case the string is just a pointer into the the data stream.
+ * A block in a parsed stream can hold many tokens and there is no room
+ * allocated for nulls.
  */
 
 head_t *string_copy( program_t *prg, head_t *head )