src/scanner.c

@@ -23,6 +23,7 @@ akerr_ErrorContext *akbasic_scanner_zero(akbasic_Runtime *obj)
     obj->current = 0;
     obj->start = 0;
     obj->hasError = false;
+    obj->tokentype = AKBASIC_TOK_UNDEFINED;
     SUCCEED_RETURN(errctx);
 }
 
@@ -409,6 +410,16 @@ akerr_ErrorContext *akbasic_scanner_scan(akbasic_Runtime *obj, const char *line,
     obj->current = 0;
     obj->start = 0;
     obj->hasError = false;
+    /*
+     * The `REM` early-exit below leaves `tokentype` holding AKBASIC_TOK_REM,
+     * and the loop's post-switch check reads it before the first character of
+     * the *next* line has assigned anything. A line whose first character
+     * carries no token of its own -- leading whitespace -- then re-triggered
+     * the REM break and scanned to nothing: every indented line after a REM
+     * was silently skipped. A numbered program never saw it, because the line
+     * number is the first token and overwrites the leftover.
+     */
+    obj->tokentype = AKBASIC_TOK_UNDEFINED;
     /*
      * Cleared here rather than by each caller, so the flag always describes the
      * line this call just scanned. It used to be cleared only in

tests/language/statements/rem_indented_line.bas

@@ -0,0 +1,14 @@
+REM The line after this comment is indented, and it must still run: the
+REM scanner's REM early-exit used to leave TOK_REM armed, and the next
+REM line's leading whitespace re-triggered it -- every indented line
+REM after a REM was silently skipped. Unnumbered on purpose: a numbered
+REM line's first token overwrites the leftover and hides the defect.
+PRINT 1
+REM an indented statement follows
+  PRINT 2
+REM an indented multi-line DEF body, the shape that found it
+DEF F(N#)
+  REM a comment inside the body
+  RETURN N# + 5
+PRINT F(10)
+END

tests/language/statements/rem_indented_line.txt

@@ -0,0 +1,3 @@
+1
+2
+15