From 74a1bb5fae4d39fcd2c1457ffaeef9c7ac07f2d6 Mon Sep 17 00:00:00 2001
From: Simon Cozens <simon@simon-cozens.org>
Date: Mon, 31 Oct 2022 13:33:47 +0000
Subject: [PATCH] Use shaper to check whether glyphs exist, see #7

---
 shaperglot/checks/orthographies.py | 17 +++++++++++++----
 1 file changed, 13 insertions(+), 4 deletions(-)

diff --git a/shaperglot/checks/orthographies.py b/shaperglot/checks/orthographies.py
index 8242307..13bd3a1 100644
--- a/shaperglot/checks/orthographies.py
+++ b/shaperglot/checks/orthographies.py
@@ -1,7 +1,12 @@
+import re
+
 from strictyaml import Map
 
 from .common import ShaperglotCheck, and_join
 
+def parse_bases(bases):
+    return [x[0] or x[1] for x in re.findall(r"\{([^}]+)\}|(\S+)", bases)]
+
 
 class OrthographiesCheck(ShaperglotCheck):
     name = "orthographies"
@@ -11,7 +16,7 @@ class OrthographiesCheck(ShaperglotCheck):
     def __init__(self, lang):
         exemplar_chars = lang.get("exemplarChars", {})
         marks = exemplar_chars.get("marks", "").replace("◌", "").split() or []
-        bases = exemplar_chars.get("base", "").split() or []
+        bases = parse_bases(exemplar_chars.get("base", ""))
         self.all_glyphs = marks + bases
         self.marks = set(marks)
         self.bases = set(bases) - self.marks
@@ -21,22 +26,26 @@ def describe(self):
             f"'{g}'" for g in self.all_glyphs
         )
 
+    def can_shape(self, text, checker):
+        buf = checker.vharfbuzz.shape(text)
+        return all(gi.codepoint != 0 for gi in buf.glyph_infos)
+
     def execute(self, checker):
         if not self.all_glyphs:
             checker.results.warn(
                 f"No glyphs were defined for language {checker.lang['name']}"
             )
             return
-        missing = [x for x in self.bases if ord(x) not in checker.cmap]
+        missing = [x for x in self.bases if not self.can_shape(x, checker)]
         if missing:
             missing = ", ".join(missing)
             checker.results.fail(f"Some base glyphs were missing: {missing}")
         else:
             checker.results.okay("All base glyphs were present in the font")
         if self.marks:
-            missing = [x for x in self.marks if ord(x) not in checker.cmap]
+            missing = [x for x in self.marks if not self.can_shape(x, checker)]
             if missing:
-                missing = ", ".join(missing)
+                missing = ", ".join([chr(0x25cc)+x for x in missing])
                 checker.results.fail(f"Some mark glyphs were missing: {missing}")
             else:
                 checker.results.okay("All mark glyphs were present in the font")