check_layout: Check for unknown keys

This can spot mispelled special keys.

check_layout.output

@@ -26,6 +26,7 @@ latn_bone: Missing important key, missing: loc capslock
 latn_colemak: Some keys contain whitespaces, unexpected:  ́
 latn_dvorak: Missing important key, missing: loc capslock
 latn_neo2: Layout redefines the bottom row but some important keys are missing, missing: loc switch_clipboard
+latn_qwerty_jp: Layout contains unknown keys: accent-macron
 latn_qwerty_se: Duplicate keys: !, ', ,, -, ., ?
 latn_qwerty_tly: Duplicate keys: a, c, j, q
 latn_qwerty_tly: Missing programming keys, missing: loc esc, loc tab

check_layout.py

@@ -1,5 +1,5 @@
 import xml.etree.ElementTree as ET
-import sys, os, glob
+import sys, os, glob, re
 
 layout_file_name = 0
 warnings = []
@@ -15,6 +15,9 @@ KEY_ATTRIBUTES = set([
     "c", "nw", "ne", "sw", "se", "w", "e", "n", "s"
     ])
 
+# Keys defined in KeyValue.java
+known_keys = set()
+
 def warn(msg):
     global warnings
     warnings.append("%s: %s" % (layout_file_name, msg))
@@ -103,6 +106,23 @@ def check_layout(layout):
     if root.get("script") == None:
         warn("Layout doesn't specify a script.")
 
+    keys_without_loc = set(( k.removeprefix("loc ") for k in keys ))
+    # Keys with a len under 3 are often composed characters
+    special_keys = set(( k for k in keys_without_loc if len(k) > 3 and ":" not in k ))
+    unknown = special_keys.difference(known_keys)
+    if len(unknown) > 0:
+        warn("Layout contains unknown keys: %s" % key_list_str(unknown))
+
+# Fill 'known_keys', which is used for some checks
+def parse_known_keys():
+    global known_keys
+    with open("srcs/juloo.keyboard2/KeyValue.java", "r") as f:
+        known_keys = set(
+                ( m.group(1) for m in re.finditer('case "([^"]+)":', f.read()) )
+                )
+
+parse_known_keys()
+
 for fname in sorted(glob.glob("srcs/layouts/*.xml")):
     layout_id, _ = os.path.splitext(os.path.basename(fname))
     if layout_id in KNOWN_NOT_LAYOUT: