Add win32 surface intended for use with printer DCs; GDI will be used
as much as possible, and the surface will be a paginated surface
that supports fine-grained fallback.
(Original work from Adrian Johnson; additional fixes by me.)
Same for FT_Render_Glyph.
When the user asks us to render a glyph that is not available in the font,
it's mostly an unavoidable kind of error for them, as in, they can't
avoid such a call. So it's not nice to put cairo_t in an error state and
refuse any further drawying.
Many PDF files are created using buggy software and cause such glpyh-not-found
errors for CID 0 for example.
Eventually we should propagate these kind of errors up and return it from
the function call causing it, but that needs API change to add return value
to all text functions, so for now we just ignore these errors.
The finer-grained fallbacks would not work correctly if the page
was set to a larger size.
Add _cairo_paginated_surface_set_size() function that is called
from cairo_ps_surface_set_size() and cairo_pdf_surface_set_size().
The DSC and EPS specifications do not allow the use of initclip.
Instead each page is wrapped in a gsave/restore pair and a
"grestore gsave" is emitted when the clip path is reset.
The cairo_matrix_multiply(out,a,b) docs say that it is equivalent to applying
matrix a first, followed by b. Looking at _cairo_gstate_backend_to_user() we should
apply device_transform_inverse followed by ctm_inverse. That's what we do now.
This patch adds cairo_surface_copy_page and cairo_surface_show_page
as public methods, leaving the previous cairo_show_page variants as
shorthands. copy_page/show_page are specific to the surface, not
to the context, so they need to be surface methods.
The BC (background color) in the smask of the recently added
cairo_mask() support was causing Ghostscript to crash due to the wrong
number of BC values. The BC entry has been removed as BC default color
is already what we want.
Under rare circumstances we may need to extract a surface that
represents a bitmap with width==0 and rows==0. Detect this case at the
start and simply return a zero-sized surface.
Fixes https://bugs.freedesktop.org/show_bug.cgi?id=12284.
The optimization that avoids replaying commands prior to an unclipped
CLEAR operation now starts playback from the first command after the
CLEAR. This avoids the need to handle the unclipped CLEAR in the PDF
surface.
This reverts commit 919bea6dbb.
Sadly as Behdad points out some backends do modify the glyph array and,
for example cairo-xlib-surface, hide this from the compiler with some
evil casts.
Skip the memory duplication of the incoming glyphs if we do not need
to transform them into the backend coordinate system.
As a consequence we need to constify the glyphs passed to the backend
functions.
The PDF surface was adding extra stops at the 0.0 and 1.0 offset when
there was not already stops at these offsets. This has been replaced
with code to move the coordinates of the linear gradient line in to
the position of the first and last offset.
If scaled_font_destroy() is called, a deadlock can result; there's no
reason to call destroy since the initialization failed (and, indeed,
it might not be valid to do so anyway).
Patch from: Robert O'Callahan <roc@ocallahans.org>
Add cairo_win32_font_face_create_for_logfontw_hfont, allow win32
scaled_fonts to rescale themselves properly to the required CTM and
only use the font_face's hfont if we're sure it's appropriate.
The stream handling has been changed to support writing the content to
one or more group objects. Each page has a top level knockout
group. The first operation in the knockout group paints another group
containing the content. Fallback images are painted from the knockout
group. This ensures that fallback images do not composite with any
content under the image.
