Use Pango instead of uniseg to break lines

weasyprint/layout/inlines.py

@@ -14,11 +14,9 @@ from __future__ import division, unicode_literals
 
 import unicodedata
 
-import uniseg.linebreak
-
 from ..css.computed_values import ex_ratio, strut_layout
 from ..formatting_structure import boxes
-from ..text import split_first_line
+from ..text import can_break_text, split_first_line
 from .absolute import AbsolutePlaceholder, absolute_layout
 from .float import avoid_collisions, float_layout
 from .min_max import handle_min_max_height, handle_min_max_width
@@ -703,8 +701,8 @@ def split_inline_box(context, box, position_x, max_x, skip_stack,
         if None in (last_letter, first):
             can_break = True
         else:
-            can_break = bool(list(uniseg.linebreak.line_break_breakables(
-                last_letter + first))[1])
+            can_break = can_break_text(
+                last_letter + first, child.style['lang'])
 
         if can_break:
             children.extend(waiting_children)
@@ -1118,9 +1116,7 @@ def can_break_inside(box):
     if isinstance(box, boxes.AtomicInlineLevelBox):
         return False
     elif isinstance(box, boxes.TextBox):
-        if box.text:
-            return any(uniseg.linebreak.line_break_breakables(box.text[1:]))
+        return can_break_text(box.text, box.style['lang'])
     elif isinstance(box, boxes.ParentBox):
         return any(can_break_inside(child) for child in box.children)
-    else:
-        return False
+    return False

weasyprint/text.py

@@ -126,6 +126,11 @@ ffi.cdef('''
         int height;
     } PangoRectangle;
 
+    typedef struct {
+        guint is_line_break: 1;
+        /* ... */
+    } PangoLogAttr;
+
     int pango_version (void);
 
     double pango_units_to_double (int i);
@@ -210,6 +215,10 @@ ffi.cdef('''
 
     PangoContext * pango_layout_get_context (PangoLayout *layout);
 
+    void pango_get_log_attrs (
+        const char *text, int length, int level, PangoLanguage *language,
+        PangoLogAttr *log_attrs, int attrs_len);
+
 
     // PangoCairo
 
@@ -1168,3 +1177,21 @@ def show_first_line(context, pango_layout, hinting):
     pango.pango_layout_set_width(pango_layout.layout, -1)
     pangocairo.pango_cairo_show_layout_line(
         context, next(pango_layout.iter_lines()))
+
+
+def can_break_text(text, lang):
+    if not text or len(text) < 2:
+        return False
+    if lang:
+        lang_p, lang = unicode_to_char_p(lang)
+    else:
+        lang = None
+        language = pango.pango_language_get_default()
+    if lang:
+        language = pango.pango_language_from_string(lang_p)
+    text_p, bytestring = unicode_to_char_p(text)
+    length = len(bytestring) + 1
+    log_attrs = ffi.new('PangoLogAttr[]', length)
+    pango.pango_get_log_attrs(
+        text_p, len(bytestring), -1, language, log_attrs, length)
+    return any(attr.is_line_break for attr in log_attrs[1:length - 1])