From 2a96d1c7d6a43d35ed47e260d646d728b955ae51 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Marcel=20Fabian=20Kr=C3=BCger?= <zauguin@gmail.com>
Date: Mon, 19 Apr 2021 17:07:29 +0200
Subject: [PATCH] Normalize some names of combining characters

The list remaps according to the table in MathML-Core, always selecting
the variant which has an XML entity name is such a character exists.
---
 mlist_to_mml.lua |  4 ++++
 remap_comb.lua   | 27 +++++++++++++++++++++++++++
 2 files changed, 31 insertions(+)
 create mode 100644 remap_comb.lua

diff --git a/mlist_to_mml.lua b/mlist_to_mml.lua
index 3eed3b8..e972480 100644
--- a/mlist_to_mml.lua
+++ b/mlist_to_mml.lua
@@ -1,3 +1,5 @@
+local remap_comb = require'remap_comb'
+
 local noad_t, accent_t, style_t, choice_t = node.id'noad', node.id'accent', node.id'style', node.id'choice'
 local radical_t, fraction_t, fence_t = node.id'radical', node.id'fraction', node.id'fence'
 
@@ -102,12 +104,14 @@ local function accent_to_table(accent, sub, cur_style)
     if sub & 1 == 1 then
       top_acc.stretchy = 'false'
     end
+    top_acc[1] = remap_comb[top_acc[1]] or top_acc[1]
   end
   if bot_acc then
     bot_acc[0] = 'mo'
     if sub & 2 == 2 then
       bot_acc.stretchy = 'false'
     end
+    bot_acc[1] = remap_comb[bot_acc[1]] or bot_acc[1]
   end
   return {[0] = top_acc and (bot_acc and 'munderover' or 'mover') or 'munder',
     nucleus,
diff --git a/remap_comb.lua b/remap_comb.lua
new file mode 100644
index 0000000..2083a3f
--- /dev/null
+++ b/remap_comb.lua
@@ -0,0 +1,27 @@
+return {
+  ["\u{0332}"] = "\u{2212}",
+  ["\u{0330}"] = "\u{02DC}",
+  ["\u{0328}"] = "\u{02DB}",
+  ["\u{20EF}"] = "\u{2192}",
+  ["\u{032C}"] = "\u{02C7}",
+  ["\u{032E}"] = "\u{02D8}",
+  ["\u{0306}"] = "\u{02D8}",
+  ["\u{030B}"] = "\u{02DD}",
+  ["\u{0302}"] = "\u{02C6}",
+  ["\u{0324}"] = "\u{00A8}",
+  ["\u{0317}"] = "\u{00B4}",
+  ["\u{031F}"] = "\u{002B}",
+  ["\u{0307}"] = "\u{002E}",
+  ["\u{0305}"] = "\u{2212}",
+  ["\u{0303}"] = "\u{02DC}",
+  ["\u{0316}"] = "\u{0060}",
+  ["\u{0301}"] = "\u{00B4}",
+  ["\u{030C}"] = "\u{02C7}",
+  ["\u{0327}"] = "\u{00B8}",
+  ["\u{0308}"] = "\u{00A8}",
+  ["\u{0300}"] = "\u{0060}",
+  ["\u{0323}"] = "\u{002E}",
+  ["\u{0304}"] = "\u{00AF}",
+  ["\u{032D}"] = "\u{005E}",
+  ["\u{20D7}"] = "\u{2192}",
+}