delete lang engine

2000-11-30  Havoc Pennington  <hp@pobox.com>

* modules/thai/thai.c: delete lang engine

* modules/tamil/tamil.c: delete lang engine
(tamil_engine_x_new): fix type tag for shape engine

* modules/indic/myanmar.c: delete lang engine
(pango_engine_x_new): fix type tag for shape engine

* modules/indic/gurmukhi.c: delete lang engine
(pango_indic_engine_x_new): fix type tag for shape engine

* modules/indic/gujarati.c: delete lang engine
(pango_indic_engine_x_new): fix type tag for shape engine

* modules/indic/devanagari.c: delete lang engine
(pango_indic_engine_x_new): fix type tag for shape engine

* modules/indic/pango-indic-script.h (SCRIPT_ENGINE_DEFINITION):
delete lang engine

* modules/indic/bengali.c: delete the lang engine
(pango_indic_engine_x_new): fix type tag for shape engine

* modules/hangul/hangul.c: delete the lang engine
(hangul_engine_x_new): fix type tag for shape engine

* modules/basic/basic.c: delete the lang engine
(basic_engine_x_new): fix type tag for shape engine

* modules/basic/basic-win32.c: delete the lang engine
(basic_engine_win32_new): this was a shape engine,
use correct type tag

* modules/basic/basic-ft2.c: delete the lang engine

* modules/arabic/arabic.c: Delete the lang engine
(arabic_engine_x_new): this is a shape
engine, not a lang engine, fix type tag

* pango/pango-layout.c (pango_layout_index_to_line_x): handle
the fact that paragraph delimiters aren't in the layout lines
(pango_layout_index_to_pos): update to handle paragraph
delimiters

* pango/break.c (pango_find_paragraph_boundary): New function
to find paragraph boundaries

* pango/pango-layout.c (get_items_log_attrs): don't separate calls
to pango_break() when directional level changes

* pango/pango-layout.h (struct _PangoLayoutLine): put start index
of the line into the struct

 * pango/pango-layout.c (pango_layout_get_cursor_pos): Fixups to
 reflect the fact that paragraph separators are removed from the
 input text.

 * pango/pango-layout.c (can_break_at): don't
 special-case start of line and whitespace-following-alphabetic
 here, because pango_break() already handles that properly

 * tests/testboundaries.c, tests/Makefile.am, tests/runtests.sh:
 Add directory for test programs, and a script to run them all

 * configure.in: Create Makefile in tests

 * pango/break.c (pango_break): Try for a real implementation of
 the Unicode text boundary algorithms
 (pango_get_log_attrs): Allow length to be -1

 * pango/pango-context.c (pango_itemize): use pango_item_new(),
 assert that items added to the list are sane.

 * pango/pango-layout.c (pango_layout_check_lines): Reimplement
 to honor the paragraph boundaries from pango_break()

 * pango/pango-layout.c (process_item): use pango_item_split() here

 * pango/pango-item.c (pango_item_split): New function to split an
 item into two items

1 files changed, 10 insertions, 3 deletions

diff --git a/pango/pango-context.c b/pango/pango-context.c
index 0e9f7146..c8a7d5c1 100644
--- a/pango/pango-context.c
+++ b/pango/pango-context.c
@@ -510,7 +510,10 @@ pango_context_get_base_dir (PangoContext *context)
  * @cached_iter:      Cached attribute iterator, or NULL
  *
  * Breaks a piece of text into segments with consistent
- * directional level and shaping engine.
+ * directional level and shaping engine. Each byte of @text will
+ * be contained in exactly one of the items in the returned list;
+ * the generated list of items will be in logical order (the start
+ * offsets of the items are ascending).
  *
  * @cached_iter should be an iterator over @attrs currently positioned at a
  * range before or containing @start_index; @cached_iter will be advanced to
@@ -565,7 +568,7 @@ pango_itemize (PangoContext      *context,
   embedding_levels = g_new (guint8, n_chars);
 
   pango_log2vis_get_embedding_levels (text_ucs4, n_chars, &base_dir,
-					embedding_levels);
+                                      embedding_levels);
 
   /* Storing these as ranges would be a lot more efficient,
    * but also more complicated... we take the simple
@@ -603,7 +606,11 @@ pango_itemize (PangoContext      *context,
 	  fonts[i] != fonts[i-1] ||
 	  extra_attr_lists[i] != extra_attr_lists[i-1])
 	{
-	  item = g_new (PangoItem, 1);
+          /* assert that previous item got at least one char */
+          g_assert (item == NULL || item->length > 0);
+          g_assert (item == NULL || item->num_chars > 0);
+          
+	  item = pango_item_new ();
 	  item->offset = p - text;
 	  item->num_chars = 0;
 	  item->analysis.level = embedding_levels[i];