First draft of the B module code; a module name can now be specified as a

compiler flag, which is used to set the name of the patch table. The compiler
now understands C preprocessor line directives. Extend the standard library
somewhat.

lang/b/compiler/b.h

@@ -4,6 +4,7 @@
 #include <string.h>
 #include <unistd.h>
 #include <em.h>
+#include "astring.h"
 
 #define	NCPS	8	/* chars per symbol */
 #define	HSHSIZ	400	/* hash table size */
@@ -40,6 +41,7 @@ struct	swtab {
 };
 
 extern int wordsize;
+extern const char* modulename;
 int paramsize;
 struct	hshtab hshtab[HSHSIZ];
 int	hshused;
@@ -89,6 +91,7 @@ char* manglename(char* name, char prefix);
 #define	RPARN	7
 #define	COLON	8
 #define	COMMA	9
+#define HASH	10
 
 #define	MCALL	15
 #define	CALL	16

lang/b/compiler/b0.c

@@ -14,6 +14,7 @@ int	contlab = -1;
 int	brklab = -1;
 
 int wordsize = 4;
+const char* modulename = "b_module_main";
 int bsymb_part;
 int code_part;
 int string_part;
@@ -38,7 +39,7 @@ main(int argc, char *argv[])
 {
 
 	for (;;) {
-		int opt = getopt(argc, argv, "-w:i:o:");
+		int opt = getopt(argc, argv, "-w:B:i:o:");
 		if (opt == -1)
 			break;
 
@@ -47,6 +48,10 @@ main(int argc, char *argv[])
 				wordsize = atoi(optarg);
 				break;
 
+			case 'B':
+				modulename = aprintf("b_module_%s", optarg);
+				break;
+
 			case 'i':
 				if (freopen(optarg, "r", stdin) == NULL) {
 					error("Can't find %s", optarg);
@@ -62,7 +67,7 @@ main(int argc, char *argv[])
 				break;
 
 			derfault:
-				error("Usage: em_b [-w wordsize] [-i inputfile] [-o outputfile]");
+				error("Usage: em_b [-w wordsize] [-B modulename] [-i inputfile] [-o outputfile]");
 				exit(1);
 		}
 	}
@@ -97,8 +102,8 @@ main(int argc, char *argv[])
 	if (string_part)
 		C_insertpart(string_part);
 
-	C_exa_dnam("bsymb_patch_table");
+	C_exa_dnam((char*) modulename);
-	C_df_dnam("bsymb_patch_table");
+	C_df_dnam((char*) modulename);
 	if (bsymb_part)
 		C_insertpart(bsymb_part);
 	C_rom_cst(0);
@@ -340,11 +345,30 @@ loop:
 
 	case NEWLN:
 		line++;
-
+		/* fall through */
 	case SPACE:
 		c = getchar();
 		goto loop;
 
+	case HASH:
+		/* # is invalid in B; but we handle it out of convenience so that we can read
+		 * in input files that have been run through the C preprocessor. Ideally we
+		 * should only recognise it when it's the first character in a line, but as
+		 * it's not used anywhere else we can get away with recognising it anywhere.
+		 */
+
+		while ((c = getchar()) == ' ')
+			;
+
+		peekc = c;
+		getnum();
+		line = cval;
+
+		while ((c = getchar()) != '\n')
+			;
+
+		goto loop;
+
 	case PLUS:
 		return subseq(c,PLUS,INCBEF);
 
@@ -1242,7 +1266,7 @@ char ctab[128] = {
 	LETTER,	SPACE,	NEWLN,	SPACE,	SPACE,	UNKN,	UNKN,	UNKN,
 	UNKN,	UNKN,	UNKN,	UNKN,	UNKN,	UNKN,	UNKN,	UNKN,
 	UNKN,	UNKN,	UNKN,	UNKN,	UNKN,	UNKN,	UNKN,	UNKN,
-	SPACE,	EXCLA,	DQUOTE,	UNKN,	UNKN,	MOD,	AND,	SQUOTE,
+	SPACE,	EXCLA,	DQUOTE,	HASH,	UNKN,	MOD,	AND,	SQUOTE,
 	LPARN,	RPARN,	TIMES,	PLUS,	COMMA,	MINUS,	UNKN,	DIVIDE,
 	DIGIT,	DIGIT,	DIGIT,	DIGIT,	DIGIT,	DIGIT,	DIGIT,	DIGIT,
 	DIGIT,	DIGIT,	COLON,	SEMI,	LESS,	ASSIGN,	GREAT,	QUEST,

lang/b/compiler/build.lua

@@ -8,6 +8,7 @@ cprogram {
 		"./*.h",
 		"modules+headers",
 		"modules/src/alloc+lib",
+		"modules/src/data+lib",
 		"modules/src/em_code+lib_k",
 		"modules/src/em_data+lib",
 		"modules/src/em_mes+lib",

lang/b/lib/b.h

@@ -14,6 +14,9 @@
 #error Unsupported EM_PSIZE
 #endif
 
+extern FILE* input_unit;
 extern FILE* output_unit;
 
+#define END 4
+
 #endif

lang/b/lib/main.c

@@ -1,8 +1,12 @@
 #include "b.h"
+#include <stdarg.h>
+#include <string.h>
+#include <limits.h>
 
-extern uintptr_t* bsymb_patch_table[];
+extern uintptr_t* b_module_main[];
 extern intptr_t i_main(intptr_t argc, const char* argv[]);
 
+FILE* input_unit;
 FILE* output_unit;
 
 static intptr_t i_char(intptr_t s, intptr_t n)
@@ -19,7 +23,7 @@ static void i_lchar(intptr_t s, intptr_t n, intptr_t c)
 
 static intptr_t i_getchar(void)
 {
-    return fgetc(output_unit);
+    return fgetc(input_unit);
 }
 
 static intptr_t i_putchar(intptr_t c)
@@ -28,17 +32,139 @@ static intptr_t i_putchar(intptr_t c)
     return c;
 }
 
+static intptr_t i_putstr(intptr_t s)
+{
+	char* p = (char*)(s<<SHIFT);
+	char c;
+
+	while ((c = *p++) != END)
+		fputc(c, output_unit);
+
+	return s;
+}
+
+static intptr_t i_getstr(intptr_t s)
+{
+	char* p = (char*)(s<<SHIFT);
+
+	for (;;)
+	{
+		int c = fgetc(input_unit);
+		if ((c == -1) || (c == '\n'))
+			break;
+		*p++ = c;
+	}
+
+	*p++ = END;
+	return s;
+}
+	
+static void i_flush(void)
+{
+	fflush(output_unit);
+}
+
+static int tochar(int n)
+{
+	if (n <= 9)
+		return n + '0';
+	return n - 10 + 'a';
+}
+
+static void putnum(intptr_t value, int base)
+{
+	int i;
+	char s[32];
+
+	if (value < 0)
+	{
+		fputc('-', output_unit);
+		value = -value;
+	}
+
+	i = 0;
+	do
+		s[i++] = tochar(value % base);
+	while ((value /= base) > 0);
+
+	do
+		fputc(s[--i], output_unit);
+	while (i > 0);
+}
+
+static void i_printf(intptr_t s, ...)
+{
+	char* p = (char*)(s<<SHIFT);
+	char c;
+
+	va_list ap;
+	va_start(ap, s);
+
+	while ((c = *p++) != END)
+	{
+		switch (c)
+		{
+			case '%':
+			{
+				intptr_t ss = va_arg(ap, intptr_t);
+
+				switch ((c = *p++))
+				{
+					case 's':
+						i_putstr(ss);
+						break;
+
+					case 'd':
+						putnum(ss, 10);
+						break;
+
+					case 'o':
+						putnum(ss, 8);
+						break;
+
+					case 'x':
+						putnum(ss, 16);
+						break;
+
+					case 'c':
+						fputc(ss, output_unit);
+						break;
+
+					default:
+						fputc('?', output_unit);
+						break;
+				}
+				break;
+			}
+				
+			default:
+				fputc(c, output_unit);
+				break;
+		}
+	}
+
+	va_end(ap);
+}
+
 uintptr_t b_char = (uintptr_t)i_char;
 uintptr_t b_lchar = (uintptr_t)i_lchar;
 uintptr_t b_getchar = (uintptr_t)i_getchar;
 uintptr_t b_putchar = (uintptr_t)i_putchar;
+uintptr_t b_putstr = (uintptr_t)i_putstr;
+uintptr_t b_getstr = (uintptr_t)i_getstr;
+uintptr_t b_flush = (uintptr_t)i_flush;
+uintptr_t b_printf = (uintptr_t)i_printf;
 
-static uintptr_t* lib_patch_table[] =
+static uintptr_t* b_module_stdlib[] =
 {
     &b_char,
     &b_lchar,
     &b_getchar,
     &b_putchar,
+	&b_putstr,
+	&b_getstr,
+	&b_flush,
+	&b_printf,
     0
 };
 
@@ -53,8 +179,9 @@ static void patch_addresses(uintptr_t** p)
 
 int main(int argc, const char* argv[])
 {
-    patch_addresses(bsymb_patch_table);
+    patch_addresses(b_module_main);
-    patch_addresses(lib_patch_table);
+    patch_addresses(b_module_stdlib);
+    input_unit = stdin;
     output_unit = stdout;
     return i_main(argc, NULL);
-}
+}