Don't split a line when drawing it

To draw a text line, the previous behaviour didn't rely only on the text
actually set on the layout, but also relied on the fact that the line
was cut again when drawn. This change removes the line cutting when
drawing, and thus only relies on the line splitting done during the
layout. This fixes a bug causing some words not being displayed at the
end of a text line drawn with hinting, and the actual drawing size with
hinting was bigger than the size calculated during the layout.

The text included in the drawn layout object was sometimes not cut at
the right position, it was longer but cut when actually drawn. This
commit also fixes this, by always setting the right text in the layout
object.

Fixing this bug enables us to remove a hack introduced to fight against
an "accumulation of floating point errors". I now think that "it wasn't
our war"™. I think that the real reasons of this hack were probably:

- a problem with trailing spaces in the shrink-to-fit functions fixed in
  commit 3a620db, and
- this line-cutting bug while drawing, fixed now.

I've tried hard to reproduce the shink-to-fit problem without this hack,
with no success. I don't see anymore how it can theorically happen with
the current code of the "preferred" module. The only bug fixed by this
hack that I've been able to reproduce is the hinting problem explained
in the first paragraph, and this bug is now really fixed.

Moreover, this hack used to cause another problem: as the maximum size
allowed to an inline block was actually bigger than the real size
available in the line, an inline block whose size was between the real
and the allowed sizes was put on a new line by the split_inline_box
function. This commit fixes #288, the bug reported about this problem.

weasyprint/text.py

@@ -274,13 +274,12 @@ def get_ink_position(line):
 
 def first_line_metrics(first_line, text, layout, resume_at, hyphenated=False):
     length = first_line.length
-    if not hyphenated:
-        first_line_text = utf8_slice(text, slice(length))
-        if first_line_text.endswith(' ') and resume_at:
-            # Remove trailing spaces
-            layout.set_text(first_line_text.rstrip(' '))
-            first_line = next(layout.iter_lines(), None)
-            length = first_line.length if first_line is not None else 0
+    if not hyphenated and resume_at:
+        # Create layout with final text, remove trailing spaces if needed
+        first_line_text = utf8_slice(text, slice(length)).rstrip(' ')
+        layout.set_text(first_line_text)
+        first_line = next(layout.iter_lines(), None)
+        length = first_line.length if first_line is not None else 0
     width, height = get_size(first_line)
     baseline = units_to_double(pango.pango_layout_iter_get_baseline(ffi.gc(
         pango.pango_layout_get_iter(layout.layout),
@@ -418,14 +417,6 @@ def split_first_line(text, style, hinting, max_width, line_width):
     ``baseline``: baseline in pixels of the first line
 
     """
-    # In some cases (shrink-to-fit result being the preferred width)
-    # this value is coming from Pango itself,
-    # but floating point errors have accumulated:
-    #   width2 = (width + X) - X   # in some cases, width2 < width
-    # Increase the value a bit to compensate and not introduce
-    # an unexpected line break.
-    if max_width is not None:
-        max_width += style.font_size * 0.2
     # Step #1: Get a draft layout with the first line
     layout = None
     if max_width:
@@ -467,24 +458,26 @@ def split_first_line(text, style, hinting, max_width, line_width):
         # The first word is longer than the line, try to hyphenize it
         first_part = ''
         second_part = text
-    next_word = second_part.split(' ', 1)[0]
 
-    if not next_word:
+    next_word = second_part.split(' ', 1)[0]
+    if next_word:
+        # next_word might fit without a space afterwards
+        new_first_line = first_part + next_word
+        layout.set_text(new_first_line)
+        lines = layout.iter_lines()
+        first_line = next(lines, None)
+        second_line = next(lines, None)
+        first_line_width, _height = get_size(first_line)
+        if second_line is None and first_line_width <= max_width:
+            # The next word fits in the first line, keep the layout
+            resume_at = len(new_first_line.encode('utf-8')) + 1
+            if resume_at == len(text.encode('utf-8')):
+                resume_at = None
+            return first_line_metrics(first_line, text, layout, resume_at)
+    else:
         # We did not find a word on the next line
-        return first_line_metrics(first_line, text, layout, resume_at)
-
-    # next_word might fit without a space afterwards.
-    # Pango previously counted that space’s advance width.
-    new_first_line = first_part + next_word
-    layout.set_text(new_first_line)
-    lines = layout.iter_lines()
-    first_line = next(lines, None)
-    second_line = next(lines, None)
-    first_line_width, _height = get_size(first_line)
-    if second_line is None and first_line_width <= max_width:
-        # The next word fits in the first line, keep the layout
-        resume_at = len(new_first_line.encode('utf-8')) + 1
-        return first_line_metrics(first_line, text, layout, resume_at)
+        if first_part:
+            return first_line_metrics(first_line, text, layout, resume_at)
 
     # Step #4: Try to hyphenize
     hyphens = style.hyphens
@@ -581,5 +574,9 @@ def show_first_line(context, pango_layout, hinting):
     context = ffi.cast('cairo_t *', context._pointer)
     if hinting:
         pangocairo.pango_cairo_update_layout(context, pango_layout.layout)
+    # Set an infinite width as we don't want to break lines when drawing, the
+    # lines have already been split and the size may differ for example because
+    # of hinting.
+    pango.pango_layout_set_width(pango_layout.layout, -1)
     pangocairo.pango_cairo_show_layout_line(
         context, next(pango_layout.iter_lines()))