html.c: fix handling of invalid entities (fix #58, thanks @ccbenny)

rrthomas · Sep 13, 2024 · aecbbe0 · aecbbe0
1 parent 483e401
commit aecbbe0
Show file tree

Hide file tree

Showing 2 changed files with 18 additions and 0 deletions.
diff --git a/src/html.c b/src/html.c
@@ -855,12 +855,19 @@ transform_html_ucs2 (RECODE_SUBTASK subtask)
 		  valid = false;
 	      }
 	  }
+	else /* Not '#' or A-z, so not a valid entity.  */
+	  {
+	    valid = false;
+	    /* Push back '&', but do not output current input_char yet.  */
+	    *cursor = '\0';
+	  }
 
 	if (echo || !valid)
 	  {
 	    recode_put_ucs2 ('&', subtask);
 	    for (cursor = buffer; *cursor; cursor++)
 	      recode_put_ucs2 (*cursor, subtask);
+	    /* input_char is processed again in the next iteration.  */
 	  }
       }
     else

diff --git a/tests/t40_html.py b/tests/t40_html.py
@@ -0,0 +1,11 @@
+# -*- coding: utf-8 -*-
+import common
+from common import setup_module, teardown_module
+
+input = '&amp; S&P, S & P, &'
+output =    '& S&P, S & P, &'
+
+class Test:
+    def test_1(self):
+        common.request('html..')
+        common.validate(input, output)