Different fix for displaying booleans properly in xhtml vs html

Original patch was rendering input tags as xhtml <input .../> rather
than html4/5 format as <input ...>.

Author: rouilj

CHANGES.txt

@@ -70,9 +70,12 @@ Fixed:
 - French translation gave errors with Python 3 because of ISO-8859-1
   character in .mo file header. (Joseph Myers)
 - Fix representation of boolean html attributes to be 'required'
-  rather than the xhtml form of 'required="required"'. Specifiy
-  attribute value same as attribute name or attribute value of None,
-  to output attribute as boolean. (John Rouillard)
+  rather than the xhtml form of 'required="required"'. Specify
+  (reverted attribute value same as attribute name or) attribute
+  value of None, to output attribute as boolean. (John Rouillard)
+  Reverted (part of) this change. It breaks rendering of non-boolean
+  attributes (like name="name"). So only value of None renders
+  attribute properly as boolean. (Ralf Schlatterbeck)
 - issue2551076 - in responsive template, default searches for bugs and
   tasks sets status=new default should be "don't care". (Report:
   Ludwig Reiter; Fix: John Rouillard)
@@ -85,9 +88,6 @@ Fixed:
   for translations. (Report: John Rouillard; Fix: Christof Meerwald)
 - Cleanup code by linting using flake8. (John Rouillard)
 - Cleanup code by security linting using bandit. (John Rouillard)
-- Revert (part of) the change that optimizes rendering of boolean
-  attributes in HTML-4. This breaks rendering of non-boolean attributes
-  (like name="name"). (Ralf Schlatterbeck)
 
 2019-10-23 2.0.0 alpha 0
 

roundup/cgi/templating.py

@@ -443,11 +443,10 @@ def html4_cgi_escape_attrs(**attrs):
         <input required ..> not <input required="required" ...>
         The latter is xhtml. Recognize booleans by:
           value is None
-          value is the same as the attribute
-        Code can use either method to indicate a pure boolean.
+        Code can use None to indicate a pure boolean.
     '''
     return ' '.join(['%s="%s"'%(k,html_escape(str(v), True)) 
-                         if v != None and k != v else '%s'%(k)
+                         if v != None else '%s'%(k)
                          for k,v in sorted(attrs.items())])
 
 def xhtml_cgi_escape_attrs(**attrs):
@@ -457,8 +456,7 @@ def xhtml_cgi_escape_attrs(**attrs):
            <input required="required" ...> not <input required ..>
         The latter is html4 or 5. Recognize booleans by:
           value is None
-          value is the same as the atribute
-        Code can use either method to indicate a pure boolean.
+        Code can use None to indicate a pure boolean.
     '''
     return ' '.join(['%s="%s"'%(k,html_escape(str(v), True))
                          if v != None else '%s="%s"'%(k,k)
@@ -480,11 +478,12 @@ def __init__(self):
         html_version = 'html4'
         if hasattr(self._client.instance.config, 'HTML_VERSION'):
             html_version = self._client.instance.config.HTML_VERSION
-        # HTML-4 allows attributes like required=required, so for now
-        # revert the change that optimizes this and breaks rendering of
-        # non-boolean attributes (like name="name").
-        self.input = input_xhtml
-        self.cgi_escape_attrs=xhtml_cgi_escape_attrs
+        if html_version == 'xhtml':
+            self.input = input_xhtml
+            self.cgi_escape_attrs=xhtml_cgi_escape_attrs
+        else:
+            self.input = input_html4
+            self.cgi_escape_attrs=html4_cgi_escape_attrs
         # self._context is used for translations.
         # will be initialized by the first call to .gettext()
         self._context = None