static pdf_font_desc *
load_cid_font(pdf_document *xref, pdf_obj *dict, pdf_obj *encoding, pdf_obj *to_unicode)
{
	pdf_obj *widths;
	pdf_obj *descriptor;
	pdf_font_desc *fontdesc = NULL;
	FT_Face face;
	int kind;
	char collection[256];
	char *basefont;
	int i, k, fterr;
	pdf_obj *obj;
	int dw;
	fz_context *ctx = xref->ctx;

	fz_var(fontdesc);

	fz_try(ctx)
	{
		/* Get font name and CID collection */

		basefont = pdf_to_name(pdf_dict_gets(dict, "BaseFont"));

		{
			pdf_obj *cidinfo;
			char tmpstr[64];
			int tmplen;

			cidinfo = pdf_dict_gets(dict, "CIDSystemInfo");
			if (!cidinfo)
				fz_throw(ctx, "cid font is missing info");

			obj = pdf_dict_gets(cidinfo, "Registry");
			tmplen = fz_mini(sizeof tmpstr - 1, pdf_to_str_len(obj));
			memcpy(tmpstr, pdf_to_str_buf(obj), tmplen);
			tmpstr[tmplen] = '\0';
			fz_strlcpy(collection, tmpstr, sizeof collection);

			fz_strlcat(collection, "-", sizeof collection);

			obj = pdf_dict_gets(cidinfo, "Ordering");
			tmplen = fz_mini(sizeof tmpstr - 1, pdf_to_str_len(obj));
			memcpy(tmpstr, pdf_to_str_buf(obj), tmplen);
			tmpstr[tmplen] = '\0';
			fz_strlcat(collection, tmpstr, sizeof collection);
		}

		/* Load font file */

		fontdesc = pdf_new_font_desc(ctx);

		descriptor = pdf_dict_gets(dict, "FontDescriptor");
		if (!descriptor)
			fz_throw(ctx, "syntaxerror: missing font descriptor");
		pdf_load_font_descriptor(fontdesc, xref, descriptor, collection, basefont);

		face = fontdesc->font->ft_face;
		kind = ft_kind(face);

		/* Encoding */

		if (pdf_is_name(encoding))
		{
			if (!strcmp(pdf_to_name(encoding), "Identity-H"))
				fontdesc->encoding = pdf_new_identity_cmap(ctx, 0, 2);
			else if (!strcmp(pdf_to_name(encoding), "Identity-V"))
				fontdesc->encoding = pdf_new_identity_cmap(ctx, 1, 2);
			else
				fontdesc->encoding = pdf_load_system_cmap(ctx, pdf_to_name(encoding));
		}
		else if (pdf_is_indirect(encoding))
		{
			fontdesc->encoding = pdf_load_embedded_cmap(xref, encoding);
		}
		else
		{
			fz_throw(ctx, "syntaxerror: font missing encoding");
		}
		fontdesc->size += pdf_cmap_size(ctx, fontdesc->encoding);

		pdf_set_font_wmode(ctx, fontdesc, pdf_cmap_wmode(ctx, fontdesc->encoding));

		if (kind == TRUETYPE)
		{
			pdf_obj *cidtogidmap;

			cidtogidmap = pdf_dict_gets(dict, "CIDToGIDMap");
			if (pdf_is_indirect(cidtogidmap))
			{
				fz_buffer *buf;

				buf = pdf_load_stream(xref, pdf_to_num(cidtogidmap), pdf_to_gen(cidtogidmap));

				fontdesc->cid_to_gid_len = (buf->len) / 2;
				fontdesc->cid_to_gid = fz_malloc_array(ctx, fontdesc->cid_to_gid_len, sizeof(unsigned short));
				fontdesc->size += fontdesc->cid_to_gid_len * sizeof(unsigned short);
				for (i = 0; i < fontdesc->cid_to_gid_len; i++)
					fontdesc->cid_to_gid[i] = (buf->data[i * 2] << 8) + buf->data[i * 2 + 1];

				fz_drop_buffer(ctx, buf);
			}

			/* if truetype font is external, cidtogidmap should not be identity */
			/* so we map from cid to unicode and then map that through the (3 1) */
			/* unicode cmap to get a glyph id */
			else if (fontdesc->font->ft_substitute)
			{
				fterr = FT_Select_Charmap(face, ft_encoding_unicode);
				if (fterr)
				{
					fz_throw(ctx, "fonterror: no unicode cmap when emulating CID font: %s", ft_error_string(fterr));
				}

				if (!strcmp(collection, "Adobe-CNS1"))
					fontdesc->to_ttf_cmap = pdf_load_system_cmap(ctx, "Adobe-CNS1-UCS2");
				else if (!strcmp(collection, "Adobe-GB1"))
					fontdesc->to_ttf_cmap = pdf_load_system_cmap(ctx, "Adobe-GB1-UCS2");
				else if (!strcmp(collection, "Adobe-Japan1"))
					fontdesc->to_ttf_cmap = pdf_load_system_cmap(ctx, "Adobe-Japan1-UCS2");
				else if (!strcmp(collection, "Adobe-Japan2"))
					fontdesc->to_ttf_cmap = pdf_load_system_cmap(ctx, "Adobe-Japan2-UCS2");
				else if (!strcmp(collection, "Adobe-Korea1"))
					fontdesc->to_ttf_cmap = pdf_load_system_cmap(ctx, "Adobe-Korea1-UCS2");
			}
		}

		pdf_load_to_unicode(xref, fontdesc, NULL, collection, to_unicode);

		/* Horizontal */

		dw = 1000;
		obj = pdf_dict_gets(dict, "DW");
		if (obj)
			dw = pdf_to_int(obj);
		pdf_set_default_hmtx(ctx, fontdesc, dw);

		widths = pdf_dict_gets(dict, "W");
		if (widths)
		{
			int c0, c1, w, n, m;

			n = pdf_array_len(widths);
			for (i = 0; i < n; )
			{
				c0 = pdf_to_int(pdf_array_get(widths, i));
				obj = pdf_array_get(widths, i + 1);
				if (pdf_is_array(obj))
				{
					m = pdf_array_len(obj);
					for (k = 0; k < m; k++)
					{
						w = pdf_to_int(pdf_array_get(obj, k));
						pdf_add_hmtx(ctx, fontdesc, c0 + k, c0 + k, w);
					}
					i += 2;
				}
				else
				{
					c1 = pdf_to_int(obj);
					w = pdf_to_int(pdf_array_get(widths, i + 2));
					pdf_add_hmtx(ctx, fontdesc, c0, c1, w);
					i += 3;
				}
			}
		}

		pdf_end_hmtx(ctx, fontdesc);

		/* Vertical */

		if (pdf_cmap_wmode(ctx, fontdesc->encoding) == 1)
		{
			int dw2y = 880;
			int dw2w = -1000;

			obj = pdf_dict_gets(dict, "DW2");
			if (obj)
			{
				dw2y = pdf_to_int(pdf_array_get(obj, 0));
				dw2w = pdf_to_int(pdf_array_get(obj, 1));
			}

			pdf_set_default_vmtx(ctx, fontdesc, dw2y, dw2w);

			widths = pdf_dict_gets(dict, "W2");
			if (widths)
			{
				int c0, c1, w, x, y, n;

				n = pdf_array_len(widths);
				for (i = 0; i < n; )
				{
					c0 = pdf_to_int(pdf_array_get(widths, i));
					obj = pdf_array_get(widths, i + 1);
					if (pdf_is_array(obj))
					{
						int m = pdf_array_len(obj);
						for (k = 0; k * 3 < m; k ++)
						{
							w = pdf_to_int(pdf_array_get(obj, k * 3 + 0));
							x = pdf_to_int(pdf_array_get(obj, k * 3 + 1));
							y = pdf_to_int(pdf_array_get(obj, k * 3 + 2));
							pdf_add_vmtx(ctx, fontdesc, c0 + k, c0 + k, x, y, w);
						}
						i += 2;
					}
					else
					{
						c1 = pdf_to_int(obj);
						w = pdf_to_int(pdf_array_get(widths, i + 2));
						x = pdf_to_int(pdf_array_get(widths, i + 3));
						y = pdf_to_int(pdf_array_get(widths, i + 4));
						pdf_add_vmtx(ctx, fontdesc, c0, c1, x, y, w);
						i += 5;
					}
				}
			}

			pdf_end_vmtx(ctx, fontdesc);
		}
	}
	fz_catch(ctx)
	{
		pdf_drop_font(ctx, fontdesc);
		fz_throw(ctx, "cannot load cid font (%d %d R)", pdf_to_num(dict), pdf_to_gen(dict));
	}

	return fontdesc;
}
Esempio n. 2
0
static pdf_font_desc *
load_cid_font(pdf_document *doc, pdf_obj *dict, pdf_obj *encoding, pdf_obj *to_unicode)
{
	pdf_obj *widths;
	pdf_obj *descriptor;
	pdf_font_desc *fontdesc = NULL;
	FT_Face face;
	int kind;
	char collection[256];
	char *basefont;
	int i, k, fterr;
	pdf_obj *obj;
	int dw;
	fz_context *ctx = doc->ctx;

	fz_var(fontdesc);

	fz_try(ctx)
	{
		/* Get font name and CID collection */

		basefont = pdf_to_name(pdf_dict_gets(dict, "BaseFont"));

		{
			pdf_obj *cidinfo;
			char tmpstr[64];
			int tmplen;

			cidinfo = pdf_dict_gets(dict, "CIDSystemInfo");
			if (!cidinfo)
				fz_throw(ctx, FZ_ERROR_GENERIC, "cid font is missing info");

			obj = pdf_dict_gets(cidinfo, "Registry");
			tmplen = fz_mini(sizeof tmpstr - 1, pdf_to_str_len(obj));
			memcpy(tmpstr, pdf_to_str_buf(obj), tmplen);
			tmpstr[tmplen] = '\0';
			fz_strlcpy(collection, tmpstr, sizeof collection);

			fz_strlcat(collection, "-", sizeof collection);

			obj = pdf_dict_gets(cidinfo, "Ordering");
			tmplen = fz_mini(sizeof tmpstr - 1, pdf_to_str_len(obj));
			memcpy(tmpstr, pdf_to_str_buf(obj), tmplen);
			tmpstr[tmplen] = '\0';
			fz_strlcat(collection, tmpstr, sizeof collection);
		}

		/* Load font file */

		fontdesc = pdf_new_font_desc(ctx);

		descriptor = pdf_dict_gets(dict, "FontDescriptor");
		if (!descriptor)
			fz_throw(ctx, FZ_ERROR_GENERIC, "syntaxerror: missing font descriptor");
		pdf_load_font_descriptor(fontdesc, doc, descriptor, collection, basefont, 1, 1);

		face = fontdesc->font->ft_face;
		kind = ft_kind(face);

		/* Encoding */

		if (pdf_is_name(encoding))
		{
			if (!strcmp(pdf_to_name(encoding), "Identity-H"))
				fontdesc->encoding = pdf_new_identity_cmap(ctx, 0, 2);
			else if (!strcmp(pdf_to_name(encoding), "Identity-V"))
				fontdesc->encoding = pdf_new_identity_cmap(ctx, 1, 2);
			else
				fontdesc->encoding = pdf_load_system_cmap(ctx, pdf_to_name(encoding));
		}
		else if (pdf_is_indirect(encoding))
		{
			fontdesc->encoding = pdf_load_embedded_cmap(doc, encoding);
		}
		else
		{
			fz_throw(ctx, FZ_ERROR_GENERIC, "syntaxerror: font missing encoding");
		}
		fontdesc->size += pdf_cmap_size(ctx, fontdesc->encoding);

		pdf_set_font_wmode(ctx, fontdesc, pdf_cmap_wmode(ctx, fontdesc->encoding));

		if (kind == TRUETYPE ||
			/* cf. http://code.google.com/p/sumatrapdf/issues/detail?id=1565 */
			!strcmp(pdf_to_name(pdf_dict_gets(dict, "Subtype")), "CIDFontType2") ||
			/* cf. http://code.google.com/p/sumatrapdf/issues/detail?id=1997 */
			pdf_is_indirect(pdf_dict_gets(dict, "CIDToGIDMap")))
		{
			pdf_obj *cidtogidmap;

			cidtogidmap = pdf_dict_gets(dict, "CIDToGIDMap");
			if (pdf_is_indirect(cidtogidmap))
			{
				fz_buffer *buf;

				buf = pdf_load_stream(doc, pdf_to_num(cidtogidmap), pdf_to_gen(cidtogidmap));

				fontdesc->cid_to_gid_len = (buf->len) / 2;
				fontdesc->cid_to_gid = fz_malloc_array(ctx, fontdesc->cid_to_gid_len, sizeof(unsigned short));
				fontdesc->size += fontdesc->cid_to_gid_len * sizeof(unsigned short);
				for (i = 0; i < fontdesc->cid_to_gid_len; i++)
					fontdesc->cid_to_gid[i] = (buf->data[i * 2] << 8) + buf->data[i * 2 + 1];

				fz_drop_buffer(ctx, buf);
			}

			/* if truetype font is external, cidtogidmap should not be identity */
			/* so we map from cid to unicode and then map that through the (3 1) */
			/* unicode cmap to get a glyph id */
			else if (fontdesc->font->ft_substitute)
			{
				fterr = FT_Select_Charmap(face, ft_encoding_unicode);
				if (fterr)
				{
					fz_throw(ctx, FZ_ERROR_GENERIC, "fonterror: no unicode cmap when emulating CID font: %s", ft_error_string(fterr));
				}

				if (!strcmp(collection, "Adobe-CNS1"))
					fontdesc->to_ttf_cmap = pdf_load_system_cmap(ctx, "Adobe-CNS1-UCS2");
				else if (!strcmp(collection, "Adobe-GB1"))
					fontdesc->to_ttf_cmap = pdf_load_system_cmap(ctx, "Adobe-GB1-UCS2");
				else if (!strcmp(collection, "Adobe-Japan1"))
					fontdesc->to_ttf_cmap = pdf_load_system_cmap(ctx, "Adobe-Japan1-UCS2");
				else if (!strcmp(collection, "Adobe-Japan2"))
					fontdesc->to_ttf_cmap = pdf_load_system_cmap(ctx, "Adobe-Japan2-UCS2");
				else if (!strcmp(collection, "Adobe-Korea1"))
					fontdesc->to_ttf_cmap = pdf_load_system_cmap(ctx, "Adobe-Korea1-UCS2");
				/* cf. http://code.google.com/p/sumatrapdf/issues/detail?id=2318 */
				else if (!strcmp(collection, "Adobe-Identity") && fontdesc->font->ft_file)
					fontdesc->font->ft_substitute = 0;
			}
		}

		/* cf. http://code.google.com/p/sumatrapdf/issues/detail?id=1961 */
		fz_try(ctx)
		{

		pdf_load_to_unicode(doc, fontdesc, NULL, collection, to_unicode);

		}
		fz_catch(ctx)
		{
			fz_warn(ctx, "cannot load ToUnicode CMap");
		}

		/* If we have an identity encoding, we're supposed to use the glyph ids directly.
		 * If we only have a substitute font, that won't work.
		 * Make a last ditch attempt by using
		 * the ToUnicode table if it exists to map via the substitute font's cmap. */
		if (strstr(fontdesc->encoding->cmap_name, "Identity-") && fontdesc->font->ft_substitute)
		{
			fz_warn(ctx, "non-embedded font using identity encoding: %s", basefont);
			if (fontdesc->to_unicode && !fontdesc->to_ttf_cmap)
				fontdesc->to_ttf_cmap = pdf_keep_cmap(ctx, fontdesc->to_unicode);
		}

		/* Horizontal */

		dw = 1000;
		obj = pdf_dict_gets(dict, "DW");
		if (obj)
			dw = pdf_to_int(obj);
		pdf_set_default_hmtx(ctx, fontdesc, dw);

		widths = pdf_dict_gets(dict, "W");
		if (widths)
		{
			int c0, c1, w, n, m;

			n = pdf_array_len(widths);
			for (i = 0; i < n; )
			{
				c0 = pdf_to_int(pdf_array_get(widths, i));
				obj = pdf_array_get(widths, i + 1);
				if (pdf_is_array(obj))
				{
					m = pdf_array_len(obj);
					for (k = 0; k < m; k++)
					{
						w = pdf_to_int(pdf_array_get(obj, k));
						pdf_add_hmtx(ctx, fontdesc, c0 + k, c0 + k, w);
					}
					i += 2;
				}
				else
				{
					c1 = pdf_to_int(obj);
					w = pdf_to_int(pdf_array_get(widths, i + 2));
					pdf_add_hmtx(ctx, fontdesc, c0, c1, w);
					i += 3;
				}
			}
		}

		pdf_end_hmtx(ctx, fontdesc);

		/* Vertical */

		if (pdf_cmap_wmode(ctx, fontdesc->encoding) == 1)
		{
			int dw2y = 880;
			int dw2w = -1000;

			obj = pdf_dict_gets(dict, "DW2");
			if (obj)
			{
				dw2y = pdf_to_int(pdf_array_get(obj, 0));
				dw2w = pdf_to_int(pdf_array_get(obj, 1));
			}

			pdf_set_default_vmtx(ctx, fontdesc, dw2y, dw2w);

			widths = pdf_dict_gets(dict, "W2");
			if (widths)
			{
				int c0, c1, w, x, y, n;

				n = pdf_array_len(widths);
				for (i = 0; i < n; )
				{
					c0 = pdf_to_int(pdf_array_get(widths, i));
					obj = pdf_array_get(widths, i + 1);
					if (pdf_is_array(obj))
					{
						int m = pdf_array_len(obj);
						for (k = 0; k * 3 < m; k ++)
						{
							w = pdf_to_int(pdf_array_get(obj, k * 3 + 0));
							x = pdf_to_int(pdf_array_get(obj, k * 3 + 1));
							y = pdf_to_int(pdf_array_get(obj, k * 3 + 2));
							pdf_add_vmtx(ctx, fontdesc, c0 + k, c0 + k, x, y, w);
						}
						i += 2;
					}
					else
					{
						c1 = pdf_to_int(obj);
						w = pdf_to_int(pdf_array_get(widths, i + 2));
						x = pdf_to_int(pdf_array_get(widths, i + 3));
						y = pdf_to_int(pdf_array_get(widths, i + 4));
						pdf_add_vmtx(ctx, fontdesc, c0, c1, x, y, w);
						i += 5;
					}
				}
			}

			pdf_end_vmtx(ctx, fontdesc);
		}
	}
	fz_catch(ctx)
	{
		pdf_drop_font(ctx, fontdesc);
		fz_rethrow_message(ctx, "cannot load cid font (%d %d R)", pdf_to_num(dict), pdf_to_gen(dict));
	}

	return fontdesc;
}