Merge tag 'upstream/0.9.6'

Upstream version 0.9.6

8 files changed, 62 insertions, 17 deletions

diff --git a/tests/unilbrk/test-u16-possible-linebreaks.c b/tests/unilbrk/test-u16-possible-linebreaks.c
index cdfa75cf..8d0f0062 100644
--- a/tests/unilbrk/test-u16-possible-linebreaks.c
+++ b/tests/unilbrk/test-u16-possible-linebreaks.c
@@ -1,5 +1,5 @@
 /* Test of line breaking of UTF-16 strings.
-   Copyright (C) 2008-2010 Free Software Foundation, Inc.
+   Copyright (C) 2008-2015 Free Software Foundation, Inc.
 
    This program is free software: you can redistribute it and/or modify
    it under the terms of the GNU General Public License as published by
@@ -51,7 +51,7 @@ main ()
           ASSERT (p[i] == (i == 60 ? UC_BREAK_MANDATORY :
                            i == 5
                            || i == 11 || i == 25
-                           || i == 27 || i == 29 || i == 30 || i == 35
+                           || i == 29 || i == 30
                            || i == 45 || i == 51
                            || i == 52 || i == 53 || i == 55 || i == 56
                            || i == 58 || i == 59 ? UC_BREAK_POSSIBLE :
@@ -70,7 +70,7 @@ main ()
           ASSERT (p[i] == (i == 60 ? UC_BREAK_MANDATORY :
                            i == 5
                            || i == 11 || i == 25
-                           || i == 27 || i == 29 || i == 30 || i == 35
+                           || i == 29 || i == 30
                            || i == 37 || i == 45 || i == 51
                            || i == 52 || i == 53 || i == 55 || i == 56
                            || i == 58 || i == 59 ? UC_BREAK_POSSIBLE :
@@ -80,5 +80,20 @@ main ()
     }
   }
 
+  /* Test that a break is possible after a zero-width space followed by some
+     regular spaces (rule LB8 in Unicode TR#14 revision 26).  */
+  {
+    static const uint16_t input[4] = { 'x', 0x200B, ' ', 'y' };
+    char *p = (char *) malloc (SIZEOF (input));
+    size_t i;
+
+    u16_possible_linebreaks (input, SIZEOF (input), "UTF-8", p);
+    for (i = 0; i < 4; i++)
+      {
+        ASSERT (p[i] == (i == 3 ? UC_BREAK_POSSIBLE : UC_BREAK_PROHIBITED));
+      }
+    free (p);
+  }
+
   return 0;
 }
diff --git a/tests/unilbrk/test-u16-width-linebreaks.c b/tests/unilbrk/test-u16-width-linebreaks.c
index 2523afb6..64aa577c 100644
--- a/tests/unilbrk/test-u16-width-linebreaks.c
+++ b/tests/unilbrk/test-u16-width-linebreaks.c
@@ -1,5 +1,5 @@
 /* Test of line breaking of UTF-16 strings.
-   Copyright (C) 2008-2010 Free Software Foundation, Inc.
+   Copyright (C) 2008-2015 Free Software Foundation, Inc.
 
    This program is free software: you can redistribute it and/or modify
    it under the terms of the GNU General Public License as published by
diff --git a/tests/unilbrk/test-u32-possible-linebreaks.c b/tests/unilbrk/test-u32-possible-linebreaks.c
index d775ebcd..d4335ce4 100644
--- a/tests/unilbrk/test-u32-possible-linebreaks.c
+++ b/tests/unilbrk/test-u32-possible-linebreaks.c
@@ -1,5 +1,5 @@
 /* Test of line breaking of UTF-32 strings.
-   Copyright (C) 2008-2010 Free Software Foundation, Inc.
+   Copyright (C) 2008-2015 Free Software Foundation, Inc.
 
    This program is free software: you can redistribute it and/or modify
    it under the terms of the GNU General Public License as published by
@@ -51,7 +51,7 @@ main ()
           ASSERT (p[i] == (i == 60 ? UC_BREAK_MANDATORY :
                            i == 5
                            || i == 11 || i == 25
-                           || i == 27 || i == 29 || i == 30 || i == 35
+                           || i == 29 || i == 30
                            || i == 45 || i == 51
                            || i == 52 || i == 53 || i == 55 || i == 56
                            || i == 58 || i == 59 ? UC_BREAK_POSSIBLE :
@@ -70,7 +70,7 @@ main ()
           ASSERT (p[i] == (i == 60 ? UC_BREAK_MANDATORY :
                            i == 5
                            || i == 11 || i == 25
-                           || i == 27 || i == 29 || i == 30 || i == 35
+                           || i == 29 || i == 30
                            || i == 37 || i == 45 || i == 51
                            || i == 52 || i == 53 || i == 55 || i == 56
                            || i == 58 || i == 59 ? UC_BREAK_POSSIBLE :
@@ -80,5 +80,20 @@ main ()
     }
   }
 
+  /* Test that a break is possible after a zero-width space followed by some
+     regular spaces (rule LB8 in Unicode TR#14 revision 26).  */
+  {
+    static const uint32_t input[4] = { 'x', 0x200B, ' ', 'y' };
+    char *p = (char *) malloc (SIZEOF (input));
+    size_t i;
+
+    u32_possible_linebreaks (input, SIZEOF (input), "UTF-8", p);
+    for (i = 0; i < 4; i++)
+      {
+        ASSERT (p[i] == (i == 3 ? UC_BREAK_POSSIBLE : UC_BREAK_PROHIBITED));
+      }
+    free (p);
+  }
+
   return 0;
 }
diff --git a/tests/unilbrk/test-u32-width-linebreaks.c b/tests/unilbrk/test-u32-width-linebreaks.c
index 050c2197..d597d4e5 100644
--- a/tests/unilbrk/test-u32-width-linebreaks.c
+++ b/tests/unilbrk/test-u32-width-linebreaks.c
@@ -1,5 +1,5 @@
 /* Test of line breaking of UTF-32 strings.
-   Copyright (C) 2008-2010 Free Software Foundation, Inc.
+   Copyright (C) 2008-2015 Free Software Foundation, Inc.
 
    This program is free software: you can redistribute it and/or modify
    it under the terms of the GNU General Public License as published by
diff --git a/tests/unilbrk/test-u8-possible-linebreaks.c b/tests/unilbrk/test-u8-possible-linebreaks.c
index 555acf62..fe7397cc 100644
--- a/tests/unilbrk/test-u8-possible-linebreaks.c
+++ b/tests/unilbrk/test-u8-possible-linebreaks.c
@@ -1,5 +1,5 @@
 /* Test of line breaking of UTF-8 strings.
-   Copyright (C) 2008-2010 Free Software Foundation, Inc.
+   Copyright (C) 2008-2015 Free Software Foundation, Inc.
 
    This program is free software: you can redistribute it and/or modify
    it under the terms of the GNU General Public License as published by
@@ -45,7 +45,7 @@ main ()
           ASSERT (p[i] == (i == 90 ? UC_BREAK_MANDATORY :
                            i == 7
                            || i == 13 || i == 39
-                           || i == 41 || i == 43 || i == 44 || i == 50
+                           || i == 43 || i == 44
                            || i == 61 || i == 67
                            || i == 70 || i == 73 || i == 77 || i == 80
                            || i == 84 || i == 87 ? UC_BREAK_POSSIBLE :
@@ -64,7 +64,7 @@ main ()
           ASSERT (p[i] == (i == 90 ? UC_BREAK_MANDATORY :
                            i == 7
                            || i == 13 || i == 39
-                           || i == 41 || i == 43 || i == 44 || i == 50
+                           || i == 43 || i == 44
                            || i == 52 || i == 61 || i == 67
                            || i == 70 || i == 73 || i == 77 || i == 80
                            || i == 84 || i == 87 ? UC_BREAK_POSSIBLE :
@@ -74,5 +74,20 @@ main ()
     }
   }
 
+  /* Test that a break is possible after a zero-width space followed by some
+     regular spaces (rule LB8 in Unicode TR#14 revision 26).  */
+  {
+    static const uint8_t input[6] = "x\342\200\213 y";
+    char *p = (char *) malloc (SIZEOF (input));
+    size_t i;
+
+    u8_possible_linebreaks (input, SIZEOF (input), "UTF-8", p);
+    for (i = 0; i < 4; i++)
+      {
+        ASSERT (p[i] == (i == 5 ? UC_BREAK_POSSIBLE : UC_BREAK_PROHIBITED));
+      }
+    free (p);
+  }
+
   return 0;
 }
diff --git a/tests/unilbrk/test-u8-width-linebreaks.c b/tests/unilbrk/test-u8-width-linebreaks.c
index e78b075c..157860a0 100644
--- a/tests/unilbrk/test-u8-width-linebreaks.c
+++ b/tests/unilbrk/test-u8-width-linebreaks.c
@@ -1,5 +1,5 @@
 /* Test of line breaking of UTF-8 strings.
-   Copyright (C) 2008-2010 Free Software Foundation, Inc.
+   Copyright (C) 2008-2015 Free Software Foundation, Inc.
 
    This program is free software: you can redistribute it and/or modify
    it under the terms of the GNU General Public License as published by
diff --git a/tests/unilbrk/test-ulc-possible-linebreaks.c b/tests/unilbrk/test-ulc-possible-linebreaks.c
index c64fade9..d6396b0f 100644
--- a/tests/unilbrk/test-ulc-possible-linebreaks.c
+++ b/tests/unilbrk/test-ulc-possible-linebreaks.c
@@ -1,5 +1,5 @@
 /* Test of line breaking of strings.
-   Copyright (C) 2008-2010 Free Software Foundation, Inc.
+   Copyright (C) 2008-2015 Free Software Foundation, Inc.
 
    This program is free software: you can redistribute it and/or modify
    it under the terms of the GNU General Public License as published by
@@ -42,8 +42,8 @@ main ()
     for (i = 0; i < 36; i++)
       {
         ASSERT (p[i] == (i == 35 ? UC_BREAK_MANDATORY :
-                         i == 5 || i == 11 || i == 13 || i == 15 || i == 16
-                         || i == 21 || i == 31 ? UC_BREAK_POSSIBLE :
+                         i == 5 || i == 11 || i == 15 || i == 16
+                         || i == 31 ? UC_BREAK_POSSIBLE :
                          UC_BREAK_PROHIBITED));
       }
     free (p);
diff --git a/tests/unilbrk/test-ulc-width-linebreaks.c b/tests/unilbrk/test-ulc-width-linebreaks.c
index 875554a8..593b85cc 100644
--- a/tests/unilbrk/test-ulc-width-linebreaks.c
+++ b/tests/unilbrk/test-ulc-width-linebreaks.c
@@ -1,5 +1,5 @@
 /* Test of line breaking of strings.
-   Copyright (C) 2008-2010 Free Software Foundation, Inc.
+   Copyright (C) 2008-2015 Free Software Foundation, Inc.
 
    This program is free software: you can redistribute it and/or modify
    it under the terms of the GNU General Public License as published by
@@ -42,7 +42,7 @@ main ()
     for (i = 0; i < 36; i++)
       {
         ASSERT (p[i] == (i == 35 ? UC_BREAK_MANDATORY :
-                         i == 11 || i == 21 || i == 31 ? UC_BREAK_POSSIBLE :
+                         i == 11 || i == 16 || i == 31 ? UC_BREAK_POSSIBLE :
                          UC_BREAK_PROHIBITED));
       }
     free (p);