Issue #407 - Modified the reader to handle escaped intraline whitespace properly, per R7RS.

CHANGELOG.md

@@ -11,9 +11,10 @@ Bug Fixes
 
 - Updated the `Dockerfile` to use a non-interactive install. This prevents build problems with the official image on DockerHub.
 - Improved `(scheme lazy)` to allow `force` and `make-promise` to accept an argument of any type. Improved representation of promises to more precisely differentiate them from other objects.
+- Updated `(scheme lazy)` such that `force` will recursively force promises.
 - Add type checking to record type accessor functions. We now raise an error if the passed object is of the wrong record type.
 - Fix issues with expanding `cond-expand` expressions in libraries. Previously there would be issues with the expansion if the code needed to be within the context of a `begin`.
-- Updated `(scheme lazy)` such that `force` will recursively force promises.
+- Modified the reader to handle escaped intraline whitespace properly, per R7RS.
 
 ## 0.20 - August 14, 2020
 

runtime.c

@@ -7118,7 +7118,8 @@ static int _read_is_hex_digit(char c)
 static void _read_string(void *data, object cont, port_type *p) 
 {
   char c;
-  int escaped = 0;
+  int escaped = 0, escaped_whitespace = 0, 
+      ewrn = 0; // esc whitespace read newline
   while(1) {
     // Read more data into buffer, if needed
     if (p->buf_idx == p->mem_buf_len) {
@@ -7129,6 +7130,28 @@ static void _read_string(void *data, object cont, port_type *p)
     c = p->mem_buf[p->buf_idx++];
     p->col_num++;
 
+    if (escaped_whitespace) {
+      switch (c) {
+      case '\t':
+      case ' ':
+        p->col_num++;
+        continue;
+        break;
+      case '\n':
+        if (ewrn == 0) {
+          ewrn = 1;
+          p->line_num++;
+          p->col_num = 1;
+          continue;
+        }
+        break;
+      default:
+        escaped_whitespace = 0;
+        ewrn = 0;
+        break;
+      }
+    }
+
     if (escaped) {
       escaped = 0;
       switch (c) {
@@ -7192,6 +7215,18 @@ static void _read_string(void *data, object cont, port_type *p)
         }
         break;
       }
+      case '\t':
+      case ' ':
+        escaped_whitespace = 1;
+        ewrn = 0;
+        p->col_num++;
+        break;
+      case '\n':
+        escaped_whitespace = 1;
+        ewrn = 1;
+        p->line_num++;
+        p->col_num = 1;
+        break;
       default:
         _read_error(data, p, "invalid escape character in string"); // TODO: char
         break;

tests/unit-tests.scm

@@ -165,6 +165,14 @@
     "abcde")
   (assert:equal "string-for-each" v '(101 100 99 98 97)))
 
+(assert:equal "interline whitespace" "1 2 \      
+3 4" "1 2 3 4")
+
+(assert:equal "interline whitespace" "1 2 \
+
+
+3 4" "1 2 \n\n3 4")
+
 ;; UTF-8 / Strings
 (assert:equal "UTF8 string length" (string-length (make-string 1 (integer->char 128))) 1)
 (assert:equal "UTF8 bv length" (bytevector-length (string->utf8 (make-string 1 (integer->char 128)))) 2)