From 9d3d3f815faf2bcb5f4479bb7cc7abe901b9136d Mon Sep 17 00:00:00 2001
From: Daniel Ehrenberg <littledan@Macintosh-103.local>
Date: Mon, 5 Jan 2009 22:19:14 -0600
Subject: [PATCH] Unicode normalization bug fixes (incomplete)

---
 basis/unicode/data/data.factor                 |  3 ++-
 basis/unicode/normalize/normalize-tests.factor |  2 +-
 basis/unicode/normalize/normalize.factor       | 16 ++++++++--------
 3 files changed, 11 insertions(+), 10 deletions(-)

diff --git a/basis/unicode/data/data.factor b/basis/unicode/data/data.factor
index 80cf40fbf1..f86dccb555 100644
--- a/basis/unicode/data/data.factor
+++ b/basis/unicode/data/data.factor
@@ -4,7 +4,7 @@ USING: combinators.short-circuit assocs math kernel sequences
 io.files hashtables quotations splitting grouping arrays
 math.parser hash2 math.order byte-arrays words namespaces words
 compiler.units parser io.encodings.ascii values interval-maps
-ascii sets combinators locals math.ranges sorting ;
+ascii sets combinators locals math.ranges sorting make ;
 IN: unicode.data
 
 VALUE: simple-lower
@@ -102,6 +102,7 @@ VALUE: properties
       "Cc" "Cf" "Cs" "Co" } ;
 
 : num-chars HEX: 2FA1E ;
+
 ! the maximum unicode char in the first 3 planes
 
 : ?set-nth ( val index seq -- )
diff --git a/basis/unicode/normalize/normalize-tests.factor b/basis/unicode/normalize/normalize-tests.factor
index cae1380ab4..6970e1a2b6 100644
--- a/basis/unicode/normalize/normalize-tests.factor
+++ b/basis/unicode/normalize/normalize-tests.factor
@@ -41,4 +41,4 @@ IN: unicode.normalize.tests
         [ { { 5 { 1 2 3 4 5 } } } [ nfkd ] assert= ]
     } cleave ;
 
-! parse-test [ run-line ] each
+parse-test 1000 head [ run-line ] each
diff --git a/basis/unicode/normalize/normalize.factor b/basis/unicode/normalize/normalize.factor
index 0c00f526c7..0e1881785f 100644
--- a/basis/unicode/normalize/normalize.factor
+++ b/basis/unicode/normalize/normalize.factor
@@ -1,7 +1,7 @@
 ! Copyright (C) 2008 Daniel Ehrenberg.
 ! See http://factorcode.org/license.txt for BSD license.
 USING: sequences namespaces make unicode.data kernel math arrays
-locals sorting.insertion accessors ;
+locals sorting.insertion accessors assocs ;
 IN: unicode.normalize
 
 ! Conjoining Jamo behavior
@@ -117,16 +117,17 @@ SYMBOL: char
 : pass-combining ( -- )
     current non-starter? [ current , to pass-combining ] when ;
 
-: try-compose ( last-class char current-class -- )
-    swapd = [ after get push ] [
-        char get over combine-chars
-        [ nip char set ] [ after get push ] if*
+:: try-compose ( last-class new-char current-class -- new-class )
+    last-class current-class = [ new-char after get push last-class ] [
+        char get new-char combine-chars
+        [ char set last-class ]
+        [ new-char after get push current-class ] if*
     ] if ;
 
-: compose-iter ( n -- )
+: compose-iter ( last-class -- )
     current [
         dup combining-class dup
-        [ [ try-compose ] keep to compose-iter ] [ 3drop ] if
+        [ try-compose to compose-iter ] [ 3drop ] if
     ] [ drop ] if* ;
 
 : ?new-after ( -- )
@@ -138,7 +139,6 @@ SYMBOL: char
             char set to ?new-after
             0 compose-iter
             char get , after get %
-            to
         ] if (compose)
     ] when* ;