Slightly modify grammar of ls_format

- From now on literal `%` signs always have to be escaped with `\` (`\%`),
even inside conditional braces.
- `%%` won't work anymore.
- Nonexistent placeholders (for example`%&`) will be expanded to empty
  strings (`''`).

This change fixes also some bugs:
- inappropriate output width when \t was used without %S
  (example: `%p{ }\t%{ }p`)
- curly braces inside conditional braces (example: `%{{}p{}}`)

Minor: recognize literal `\t` in list_format config option and in `-F`.

topydo/lib/ListFormat.py

@@ -68,20 +68,9 @@ def strip_placeholder_braces(p_matchobj):
     into:
         (%B)
     """
-    try:
-        before = p_matchobj.group('before').strip('{}')
-    except AttributeError:
-        before = ''
-
+    before = p_matchobj.group('before') or ''
     placeholder = p_matchobj.group('placeholder')
-
-    try:
-        after = p_matchobj.group('after').strip('{}')
-    except AttributeError:
-        after = ''
-
+    after = p_matchobj.group('after') or ''
     whitespace = p_matchobj.group('whitespace') or ''
-    start = p_matchobj.group('start') or ''
-    end = p_matchobj.group('end') or ''
 
-    return start + before + '%' + placeholder + after + whitespace + end
+    return before + '%' + placeholder + after + whitespace

topydo/lib/PrettyPrinterFilter.py

@@ -174,39 +174,51 @@ class PrettyPrinterFormatFilter(PrettyPrinterFilter):
         # relative completion date
         placeholders['X'] = lambda t: 'x ' + humanize_date(t.completion_date()) if t.is_completed() else ''
 
-        # literal %
-        placeholders['%'] = lambda _: '%'
-
         # text (truncated if necessary)
         placeholders['S'] = lambda t: t.text()
 
-
-        p_todo_str = self.format
-
+        p_todo_str = re.sub(r'\\t', '\t', self.format)
+        p_todo_str_list = re.split(r'(?<!\\)%', p_todo_str)
+        main_pattern = (r'^({{(?P<before>.+?)}})?'
+                        r'(?P<placeholder>{ph}|\[{ph}\])'
+                        r'({{(?P<after>.+?)}})?'
+                        r'(?P<whitespace> *)')
+        truncate = False
+
+        for index, substr in enumerate(p_todo_str_list):
+            if index == 0:
+                continue    # first item in p_todo_str_list is surely not a placeholder
+            if not re.match(main_pattern.format(ph='['+''.join(placeholders.keys()) + ']'), substr):
+                substr = re.sub(main_pattern.format(ph='.'), '', substr) # remove nonexistent placeholder
+                p_todo_str_list[index] = substr
+                continue
             for placeholder, getter in placeholders.items():
                 repl = getter(p_todo)
-            pattern = (r'(?P<start>.*)'
-                       r'%(?P<before>{{.+?}})?'
-                       r'(?P<placeholder>{ph}|\[{ph}\])'
-                       r'(?P<after>{{.+?}})?'
-                       r'(?P<whitespace> *)'
-                       r'(?P<end>.*)').format(ph=placeholder)
-            match = re.match(pattern, p_todo_str)
+                pattern = main_pattern.format(ph=placeholder)
+                match = re.match(pattern, substr)
                 if match:
                     if repl == '':
-                    p_todo_str = re.sub(pattern, match.group('start') + match.group('end'), p_todo_str)
+                        substr = re.sub(pattern, '', substr)
                     else:
-
-                    p_todo_str = re.sub(pattern, strip_placeholder_braces, p_todo_str)
-                    p_todo_str = re.sub(r'%({ph}|\[{ph}\])'.format(ph=placeholder), repl, p_todo_str)
-                    p_todo_str = p_todo_str.rstrip()
+                        substr = re.sub(pattern, strip_placeholder_braces, substr)
+                        substr = re.sub(r'(?<!\\)%({ph}|\[{ph}\])'.format(ph=placeholder), repl, substr)
 
                         if placeholder == 'S':
+                            truncate = True
+                            repl_S = repl # copy for truncating final p_todo_str
+
+                    p_todo_str_list[index] = substr
+                    break
+
+        p_todo_str = ''.join(p_todo_str_list)
+        p_todo_str = re.sub(r'\\%', '%', p_todo_str)
         p_todo_str = re.sub(' *\t *', '\t', p_todo_str)
+
+        if truncate:
             line_width = get_terminal_size().columns
             if len(p_todo_str) >= line_width:
                 text_lim = line_width - len(p_todo_str) - 4
-                            p_todo_str = re.sub(re.escape(repl), repl[:text_lim] + '...', p_todo_str)
+                p_todo_str = re.sub(re.escape(repl_S), repl_S[:text_lim] + '...', p_todo_str)
 
         # cut trailing space left when last placeholder in p_todo_str is empty and its predecessor is not
         return p_todo_str.rstrip()