name-parser: Sort Widths before Weights

[why]
Usually we want the width before the weight:

ZedMono Nerd Font Extended ExtraBold Italic

but we get:

ZedMono Nerd Font ExtraBold Extended Italic

The reason is that we do not sort widths and weights and handle them as
one list, keeping the original order.

All fonts but Zed have the width before the weight and it never turned
up

[how]
Handle widths and weights on their own and only afterwards concattenate
width (first) to weights (second).

Signed-off-by: Fini Jastrow <ulf.fini.jastrow@desy.de>

bin/scripts/name_parser/FontnameTools.py

@@ -393,8 +393,12 @@ def parse_font_name(name):
  # Weights end up as Typographic Family parts ('after the dash')
  # Styles end up as Family parts (for classic grouping of four)
  # Others also end up in Typographic Family ('before the dash')
+ widths = [ m + s
+ for s in list(FontnameTools.known_widths)
+ for m in list(FontnameTools.known_modifiers) + ['']
+ ]
  weights = [ m + s
- for s in list(FontnameTools.known_weights2) + list(FontnameTools.known_widths)
+ for s in list(FontnameTools.known_weights2)
  for m in list(FontnameTools.known_modifiers) + [''] if m != s
  ] + list(FontnameTools.known_weights1) + list(FontnameTools.known_slopes)
  weights = [ w for w in weights if w not in FontnameTools.known_styles ]
@@ -408,9 +412,11 @@ def parse_font_name(name):
  r'(?:uni-)?1[14]', # GohuFont uni
  ]
 
+ ( style, width_token ) = FontnameTools.get_name_token(style, widths)
  ( style, weight_token ) = FontnameTools.get_name_token(style, weights)
  ( style, style_token ) = FontnameTools.get_name_token(style, FontnameTools.known_styles)
  ( style, other_token ) = FontnameTools.get_name_token(style, other)
+ weight_token = width_token + weight_token
  while 'Regular' in style_token and len(style_token) > 1:
  # Correct situation where "Regular" and something else is given
  style_token.remove('Regular')