[latex3-commits] [git/LaTeX3-latex3-luaotfload] ignorable: Make invisible chars invisible by default (c30fea4)

[Marcel Fabian Krüger]

Repository : https://github.com/latex3/luaotfload
On branch  : ignorable
Link       : https://github.com/latex3/luaotfload/commit/c30fea48d76b183deba2926b814316a3b0a96588

>---------------------------------------------------------------

commit c30fea48d76b183deba2926b814316a3b0a96588
Author: Marcel Fabian Krüger <tex at 2krueger.de>
Date:   Sat Aug 17 16:51:07 2019 +0200

    Make invisible chars invisible by default


>---------------------------------------------------------------

c30fea48d76b183deba2926b814316a3b0a96588
 src/luaotfload-notdef.lua                            |  2 +-
 texmf/tex/latex-dev/luaotfload/luaotfload-notdef.lua | 16 ++++++++++++++--
 2 files changed, 15 insertions(+), 3 deletions(-)

diff --git a/src/luaotfload-notdef.lua b/src/luaotfload-notdef.lua
index 2d97777..29162da 100644
--- a/src/luaotfload-notdef.lua
+++ b/src/luaotfload-notdef.lua
@@ -187,7 +187,7 @@ end
 otfregister {
   name = 'invisible',
   description = 'Remove invisible control characters',
-  default = false,
+  default = true,
   initializers = {
     node = invisibleinitialiser,
   },
diff --git a/texmf/tex/latex-dev/luaotfload/luaotfload-notdef.lua b/texmf/tex/latex-dev/luaotfload/luaotfload-notdef.lua
index e8356bf..29162da 100644
--- a/texmf/tex/latex-dev/luaotfload/luaotfload-notdef.lua
+++ b/texmf/tex/latex-dev/luaotfload/luaotfload-notdef.lua
@@ -30,6 +30,18 @@ local remove             = node.direct.remove
 local setfont            = node.direct.setfont
 local traverse_char      = node.direct.traverse_char
 
+-- According to DerivedCoreProperties.txt, Default_Ignorable_Code_Point
+-- is generated from:
+--    Other_Default_Ignorable_Code_Point
+--  + Cf (Format characters)
+--  + Variation_Selector
+--  - White_Space
+--  - FFF9..FFFB (Interlinear annotation format characters)
+--  - 13430..13438 (Egyptian hieroglyph format characters)
+--  - Prepended_Concatenation_Mark (Exceptional format characters that should be visible)
+-- Based on HarfBuzz, we add the exclusion
+--  - Lo (Letter, Other)
+-- This affects Hangul fillers.
 local ignorable_codepoints do
   local sep = lpeg.P' '^0 * ';' * lpeg.P' '^0
   local codepoint = lpeg.S'0123456789ABCDEF'^4/function(c)return tonumber(c, 16)end
@@ -52,7 +64,7 @@ local ignorable_codepoints do
                                                + 'Variation_Selector' * lpeg.Cc(true)
                                                + 'White_Space' * lpeg.Cc(nil)
                                                + 'Prepended_Concatenation_Mark' * lpeg.Cc(nil)
-                                          ) * lpeg.P' '^0 * '#')^-1 * (1-lpeg.P'\n')^0 * '\n'
+                                          ) * ' # ' * (1-lpeg.P'Lo'))^-1 * (1-lpeg.P'\n')^0 * '\n'
   file = lpeg.Cf(
       lpeg.Carg(1)
     * entry^0
@@ -175,7 +187,7 @@ end
 otfregister {
   name = 'invisible',
   description = 'Remove invisible control characters',
-  default = false,
+  default = true,
   initializers = {
     node = invisibleinitialiser,
   },