changed awk to convert a map to the size of the map in a scalar context.

allowed @include inside a normal block

qse/doc/page/awk-lang.md

@@ -14,24 +14,27 @@ QSEAWK reads an AWK program, recognizes various tokens contained while skipping
 comments and whitespaces that don't constinute a token, analyses syntax, and
 tranforms them to an internal form for execution.
 
-
-
  Program Structure
-----------------------
+--------------------------
 
-A QSEAWK program can be composed of the following elements at the top level.
+A QSEAWK program is composed of the following elements at the top level.
 
  - pattern-action block pair
- - action block without a pattern
+   - BEGIN action block pair
- - pattern without an action block
+   - END action block pair
+   - action block without a pattern
+   - pattern without an action block
  - user-defined function
- - \@global variable definition
+ - \@global variable declaration
- - \@include statement 
+ - \@include directive
 
 However, none of the above is mandatory. QSEAWK accepts an empty program.
 
-A typical pattern-action pair is composed of a pattern and an action block 
+ Patter-Action Block Pairs
-as shown below:
+--------------------------------
+
+A pattern-action pair is composed of a pattern and an action block as 
+shown below:
 
 ~~~~~{.awk}
  pattern {
@@ -162,8 +165,53 @@ third input line inclusive.
 > how to change the entry point programatically.
 
 
- Comments
+
---------------
+
+ \@include 
+----------------
+
+The \@include directive inserts the contents of the file specified in the
+following string as if they appeared in the source stream being processed.
+
+~~~~~{.awk}
+ @include "abc.awk"
+ BEGIN { func_in_abc (); }
+~~~~~
+
+A semicolon is optional after the included file name. The following is the 
+same as the sample above.
+
+~~~~~{.awk}
+ @include "abc.awk";
+ BEGIN { func_in_abc(); }
+~~~~~
+
+The directive can be used inside a block.
+
+~~~~~{.awk}
+ BEGIN {
+   @include "abc.awk";
+   print var_in_abc;
+ }
+~~~~~
+
+> ### Note ###
+> If #QSE_AWK_NEWLINE is off, a semicolon is required after the string.
+> 
+> See #qse_awk_sio_t for customizing file includsion handling.
+
+
+ Tokens
+------------
+
+When QSEAWK parses a program, it classifies a series of input characters 
+into meaningful tokens. It can extract the smallest meaningful unit through
+this tokenization process. 
+
+### Comments ###
+
+A comment is a part of the program text excluded during tokenization. You can
+put descriptive text about the program in a comment.
 
 A single-line comment is introduced by a hash character #, and is terminated at 
 the end of the same line. Additionally, it supports a C-style multi-line comment
@@ -178,16 +226,6 @@ string literals and regular expressions.
  */
 ~~~~~
 
-
-
-
- Tokens
-------------
-
-When QSEAWK parses a program, it classifies a series of input characters 
-into meaningful tokens. It can extract the smallest meaningful unit through
-this tokenization process. 
-
 ### Reserved Words ###
 
 The following words are reserved and cannot be used as a variable name,
@@ -374,8 +412,8 @@ converted to an integer 32 and the division is performed producing a
 floating-pointer number.
 
 You can create a hashed map by assigning a value to a variable indexed with
-a subscript placed within square brackets. You can't convert a hashed map to 
+a subscript placed within square brackets. A hashed map is converted to its size
-a scalar type implicitly and explcitly.
+in the numeric context and "#MAP" in the string context.
 
 A regular expression may or may not be viewed as a data type. You can't assign
 a compiled regular expression into a variable. A regular expression not placed
@@ -390,12 +428,18 @@ an empty string, and a nil value evaluate to false. All other values evaluate
 to true.
 
 > ### Note ###
+> The automatic conversion of a hashed map to a number or a string is 
+> disallowed if #QSE_AWK_FLEXMAP is off. This program ends up with an
+> error when it is off.
+> ~~~~~{.awk}
+>   BEGIN  { a[1]=1; a[30]=2; a[4]=4; print ("a=" a); }
+> ~~~~~
+> 
 > See #qse_awk_fnc_spec_t and #qse_awk_fnc_arg_t to define a function that 
 > can accept a regular expression.
 
 
 
-
  User-defined Variables
 ----------------------------
 
@@ -453,9 +497,8 @@ A local variable can shade a global variable. See the sample below:
 
 > ### Note ###
 > The QSEAWK engine allows the mixture of named variables and declared variables. 
-> However, the mixture may lead to confusion. Use #QSE_AWK_IMPLICIT and 
+> However, the mixture may lead to confusion. Use #QSE_AWK_IMPLICIT to allow
-> #QSE_AWK_EXPLICIT to control what to allow and disallow when configuring 
+> named variables when configuring the engine.
-> the engine.
 
 
  Built-in Variables
@@ -547,7 +590,7 @@ function name (arg1, arg2, ..., argn) {
 
 The caller must invoke a function with the same or less number of arguments 
 than the definition. When a function is called with the less number of 
-arguments, the redundant arguments are initialized to a nil value.
+arguments, the redundant parameters are initialized to a nil value.
 
 You can't define multiple functions with the same name. The function name
 can't confict with named variables and globals variables.
@@ -555,6 +598,11 @@ can't confict with named variables and globals variables.
  Built-in Functions
 -------------------------
 
+- index
+- length
+
+TBD.
+
 
  Function Calls
 -----------------------
@@ -658,7 +706,7 @@ The !== operator is a negated form of the === operator.
 -----------------------------
 
 When #QSE_AWK_TOLERANT is on, you can use a grouped expression without
-the 'in' operator. A grouped expression is a parentheses-enclosed list
+the **in** operator. A grouped expression is a parentheses-enclosed list
 of expressions separated with a comma. Each expression in the group is
 evaluated in the appearing order. The evaluation result of the last 
 expression in the group is returned as that of the group.
@@ -761,29 +809,7 @@ able to use this statement.
 ~~~~~
 
 
-### \@include ###
 
-The \@include directive inserts the contents of the object specified in the
-following string, typically a file name, as if they appeared in the source
-stream being processed. The directive can only be used at the outmost scope 
-where global variable declarations, *BEGIN*, *END*, and/or pattern-action 
-blocks appear. 
-
-~~~~~{.awk}
- @include "abc.awk"
- BEGIN { func_in_abc (); }
-~~~~~
-
-A semicolon is optional after the included file name. The following is the 
-same as the sample above.
-
-~~~~~{.awk}
- @include "abc.awk";
- BEGIN { func_in_abc(); }
-~~~~~
-
-> ### Note ###
-> If #QSE_AWK_NEWLINE is off, the semicolon is required.
 
 
 

qse/include/qse/awk/awk.h

@@ -512,6 +512,11 @@ struct qse_awk_sio_lxc_t
 typedef struct qse_awk_sio_lxc_t qse_awk_sio_lxc_t;
 
 typedef struct qse_awk_sio_arg_t qse_awk_sio_arg_t;
+
+/**
+ * The qse_awk_sio_arg_t type defines a structure to describe the source
+ * stream.
+ */
 struct qse_awk_sio_arg_t 
 {
 	/** 
@@ -699,18 +704,20 @@ typedef struct qse_awk_prm_t qse_awk_prm_t;
  * \code
  * qse_ssize_t in (
  *    qse_awk_t* awk, qse_awk_sio_cmd_t cmd,
+ *    qse_awk_sio_arg_t* arg,
  *    qse_char_t* buf, qse_size_t size)
  * {
- *    if (cmd == QSE_AWK_SIO_OPEN) open input stream;
+ *    if (cmd == QSE_AWK_SIO_OPEN) open input stream of arg->name;
  *    else if (cmd == QSE_AWK_SIO_CLOSE) close input stream;
  *    else read input stream and fill buf up to size characters;
  * }
  *
  * qse_ssize_t out (
  *    qse_awk_t* awk, qse_awk_sio_cmd_t cmd,
+ *    qse_awk_sio_arg_t* arg,
  *    qse_char_t* data, qse_size_t size)
  * {
- *    if (cmd == QSE_AWK_SIO_OPEN) open_output_stream;
+ *    if (cmd == QSE_AWK_SIO_OPEN) open_output_stream of arg->name;
  *    else if (cmd == QSE_AWK_SIO_CLOSE) close_output_stream;
  *    else write data of size characters to output stream;
  * }

qse/lib/awk/parse.c

@@ -1461,23 +1461,46 @@ static qse_awk_nde_t* parse_block (
 			if (get_token(awk) <= -1) return QSE_NULL;
 		}
 
-		if (!MATCH(awk,TOK_XLOCAL)) break;
+		if (MATCH(awk,TOK_XINCLUDE))
-
-		if (get_token(awk) <= -1) 
 		{
-			qse_lda_delete (
+			/* @include ... */
-				awk->parse.lcls, nlcls, 
+			if (awk->opt.depth.s.incl > 0 &&
-				QSE_LDA_SIZE(awk->parse.lcls)-nlcls);
+			    awk->parse.depth.incl >=  awk->opt.depth.s.incl)
-			return QSE_NULL;
+			{
-		}
+				SETERR_LOC (awk, QSE_AWK_EINCLTD, &awk->ptok.loc);
+				return QSE_NULL;
+			}
 		
-		if (collect_locals (awk, nlcls, istop) == QSE_NULL)
+			if (get_token(awk) <= -1) return QSE_NULL;
-		{
+	
-			qse_lda_delete (
+			if (!MATCH(awk,TOK_STR))
-				awk->parse.lcls, nlcls, 
+			{
-				QSE_LDA_SIZE(awk->parse.lcls)-nlcls);
+				SETERR_LOC (awk, QSE_AWK_EINCLSTR, &awk->ptok.loc);
-			return QSE_NULL;
+				return QSE_NULL;
+			}
+		
+			if (begin_include (awk) <= -1) return QSE_NULL;
 		}
+		else if (MATCH(awk,TOK_XLOCAL))
+		{
+			/* @local ... */
+			if (get_token(awk) <= -1) 
+			{
+				qse_lda_delete (
+					awk->parse.lcls, nlcls, 
+					QSE_LDA_SIZE(awk->parse.lcls)-nlcls);
+				return QSE_NULL;
+			}
+	
+			if (collect_locals (awk, nlcls, istop) == QSE_NULL)
+			{
+				qse_lda_delete (
+					awk->parse.lcls, nlcls, 
+					QSE_LDA_SIZE(awk->parse.lcls)-nlcls);
+				return QSE_NULL;
+			}
+		}
+		else break;
 	}
 
 	/* block body */
@@ -1517,38 +1540,60 @@ static qse_awk_nde_t* parse_block (
 			break;
 		}
 
-		/* parse an actual statement in a block */
+		if (MATCH(awk,TOK_XINCLUDE))
 		{
-			qse_awk_loc_t sloc = awk->tok.loc;
+			if (awk->opt.depth.s.incl > 0 &&
-			nde = parse_statement (awk, &sloc);
+			    awk->parse.depth.incl >=  awk->opt.depth.s.incl)
-		}
+			{
+				SETERR_LOC (awk, QSE_AWK_EINCLTD, &awk->ptok.loc);
+				return QSE_NULL;
+			}
 		
-		if (nde == QSE_NULL) 
+			if (get_token(awk) <= -1) return QSE_NULL;
-		{
-			qse_lda_delete (
-				awk->parse.lcls, nlcls, 
-				QSE_LDA_SIZE(awk->parse.lcls)-nlcls);
-			if (head != QSE_NULL) qse_awk_clrpt (awk, head);
-			return QSE_NULL;
-		}
 	
-		/* remove unnecessary statements such as adjacent 
+			if (!MATCH(awk,TOK_STR))
-		 * null statements */
+			{
-		if (nde->type == QSE_AWK_NDE_NULL) 
+				SETERR_LOC (awk, QSE_AWK_EINCLSTR, &awk->ptok.loc);
-		{
+				return QSE_NULL;
-			qse_awk_clrpt (awk, nde);
+			}
-			continue;
-		}
-		if (nde->type == QSE_AWK_NDE_BLK && 
-		    ((qse_awk_nde_blk_t*)nde)->body == QSE_NULL) 
-		{
-			qse_awk_clrpt (awk, nde);
-			continue;
-		}
 		
-		if (curr == QSE_NULL) head = nde;
+			if (begin_include (awk) <= -1) return QSE_NULL;
-		else curr->next = nde;	
+		}
-		curr = nde;
+		else
+		{
+			/* parse an actual statement in a block */
+			{
+				qse_awk_loc_t sloc = awk->tok.loc;
+				nde = parse_statement (awk, &sloc);
+			}
+
+			if (nde == QSE_NULL) 
+			{
+				qse_lda_delete (
+					awk->parse.lcls, nlcls, 
+					QSE_LDA_SIZE(awk->parse.lcls)-nlcls);
+				if (head != QSE_NULL) qse_awk_clrpt (awk, head);
+				return QSE_NULL;
+			}
+
+			/* remove unnecessary statements such as adjacent 
+			 * null statements */
+			if (nde->type == QSE_AWK_NDE_NULL) 
+			{
+				qse_awk_clrpt (awk, nde);
+				continue;
+			}
+			if (nde->type == QSE_AWK_NDE_BLK && 
+			    ((qse_awk_nde_blk_t*)nde)->body == QSE_NULL) 
+			{
+				qse_awk_clrpt (awk, nde);
+				continue;
+			}
+			
+			if (curr == QSE_NULL) head = nde;
+			else curr->next = nde;	
+			curr = nde;
+		}
 	}
 
 	block = (qse_awk_nde_blk_t*) qse_awk_callocmem (awk, QSE_SIZEOF(*block));

qse/lib/awk/val.c

@@ -1605,6 +1605,15 @@ int qse_awk_rtx_valtonum (
 			);
 		}
 
+		case QSE_AWK_VAL_MAP:
+		{
+			if (rtx->awk->opt.trait & QSE_AWK_FLEXMAP)
+			{
+				*l = QSE_HTB_SIZE(((qse_awk_val_map_t*)v)->map);
+				return 0; /* long */
+			}
+		}
+
 		case QSE_AWK_VAL_REF:
 		{
 			return val_ref_to_num (rtx, (qse_awk_val_ref_t*)v, l, r);