Do a better job of emitting whitespace

1 files changed, 51 insertions, 5 deletions

diff --git a/main.c b/main.c
index 42a12bc..1eed7e2 100644
--- a/main.c
+++ b/main.c
@@ -48,6 +48,14 @@ typedef struct _zx_basic_line {
     uint16_t num, len;
 } zx_basic_line;
 
+enum zx_basic_token_type {
+    ZX_BASIC_TOKEN_UNKNOWN,
+    ZX_BASIC_TOKEN_ALNUM,
+    ZX_BASIC_TOKEN_QUOTE,
+    ZX_BASIC_TOKEN_SYMBOL,
+    ZX_BASIC_TOKEN_WORD,
+};
+
 #define ZX_BASIC_STATE_SIZE 116
 #define ZX_BASIC_LINE_LAST  0x7676
 
@@ -249,6 +257,33 @@ static inline int zx_print(uint8_t c, int tty) {
     return 0;
 }
 
+static inline enum zx_basic_token_type zx_basic_token_type_get(uint8_t b) {
+    if (ZX_CHAR_LOW(b)) {
+        uint32_t codepoint = zx_charset[b];
+
+        if ((codepoint >= 'A' && codepoint <= 'Z')
+         || (codepoint >= '0' && codepoint <= '9')) {
+            return ZX_BASIC_TOKEN_ALNUM;
+        } else {
+            return ZX_BASIC_TOKEN_SYMBOL;
+        }
+    } else if (ZX_CHAR_INVERSE(b)) {
+        return zx_basic_token_type_get(b - ZX_CHAR_INVERSE_START);
+    } else if (ZX_CHAR_TOKEN_LOW(b)) {
+        return ZX_BASIC_TOKEN_WORD;
+    } else if (ZX_CHAR_TOKEN_HIGH(b)) {
+        char *token = zx_tokens[b-ZX_CHAR_TOKEN_HIGH_START];
+
+        if (token[0] >= 'A' && token[0] <= 'Z') {
+            return ZX_BASIC_TOKEN_WORD;
+        } else {
+            return ZX_BASIC_TOKEN_SYMBOL;
+        }
+    }
+
+    return ZX_BASIC_TOKEN_UNKNOWN;
+}
+
 static ssize_t zx_dump_basic(int fd) {
     void *buf;
     ssize_t total = 0;
@@ -294,17 +329,28 @@ static ssize_t zx_dump_basic(int fd) {
 
         for (i=0; i<len; i++) {
             uint8_t c = ((uint8_t *)buf)[i];
-            int token      = ZX_CHAR_TOKEN(c),
-                token_last = ZX_CHAR_TOKEN(last);
 
-            if (token || token_last) {
-                if (putchar(' ') < 0) {
-                    goto error_io;
+            enum zx_basic_token_type type      = zx_basic_token_type_get(c),
+                                     type_last = zx_basic_token_type_get(last);
+
+            int space = 0;
+
+            if (type == ZX_BASIC_TOKEN_ALNUM) {
+                if (type_last == ZX_BASIC_TOKEN_WORD) {
+                    space = 1;
                 }
+            } else if (type == ZX_BASIC_TOKEN_SYMBOL) {
+                space = 0;
+            } else if (type == ZX_BASIC_TOKEN_WORD) {
+                space = 1;
             } else if (ZX_CHAR_TOKEN_INTEGRAL(c) || ZX_CHAR_TOKEN_FLOAT(c)) {
                 i += 5;
             }
 
+            if (space && putchar(' ') < 0) {
+                goto error_io;
+            }
+
             if (zx_print(c, tty) < 0) {
                 goto error_io;
             }