summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorMichael Vrhel <michael.vrhel@artifex.com>2015-11-04 15:05:22 -0800
committerTor Andersson <tor.andersson@artifex.com>2016-02-29 15:52:53 +0100
commit927d36d58bf0896c2ab8b470f926e1ed1a736561 (patch)
treef7a0f4fbfeb9c7620cb46ea3c5bcf72ae785aab4
parentbbe65a9910fd5fa6478985351a0fd2690d5a27f5 (diff)
downloadmupdf-927d36d58bf0896c2ab8b470f926e1ed1a736561.tar.xz
Add mutool create tool, and PDF font and image resource creation.
Initial framework for creating pdfs This adds a create option to mutool for us to use in working on the API for creating content as well as adding content to existing documents. mutool create: Get page sizes and add them Start the parsing of the contents.txt file which may have multiple page information. Add the pages at the proper sizes. Further work on mutool create_pdf Remove the calls that were being made to the pdf-write device. Clean up several issues with the reading of the page contents. Get the content streams for each page associated with the page->contents Temp. created a pdf_create_page_contents procedure. I will merge this with pdf_create_page as there is significant overlap. Next is to add in the font and image resources and indirect references. Include pdfcreate in build Merge pdf_create_page_contents and pdf_create_page Add support for images in pdfcreate This adds images to the pdf document using a function stolen from pdf-device (send_image). This was renamed pdf_add_image_res and added to pdf-image. Down the road, send-image will be removed. Prior to that, I need to work on making sure that multiple copies of the same image do not end up in the document. Code was also added to create the page resources to point to the proper image in the document. Next fonts will be added in a similar manner, then I will work on computing the md5 sums of image and fonts to ensure only one copy ends up in the document. Then pdf-write will be reworked to use the same code as opposed to its current list of md5 sums that are stored in a device structure. mutool pdfcreate: support for WinAnsiEncoded fonts Added support for very simple fonts (WinAnsiEncoding). Methods added in pdf-font.c. Added first_width and last_width to fz_font_s and stem_v to pdf_font_desc_s. Ran code through memento with simple test of 4 page document creation including an image and a font. Fixed several leaks as well as buffer corruption issues (main changes in pdfcreate). Thanks to Robin for the help with Memento in finding leaks. Added StemV to pdf names as it was needed for the font descriptor creation. Fix for pdf_write_document rename to pdf_save_document Add resource_ids to pdf document structure The purpose of this structure will be to allow the search and reuse of resources when we attempt to add new ones to the document. Fix name changes from recent updates pdf_create branch updated to work with recent changes in master Initial use of hash table for resources To avoid adding in the same resource this adds a resource_tables member to pdf_document. The resource_tables structure consists of multiple fz_hash_table entries, one for each resource type. When an attempt is made to search for an existing resource, the table will be initialized in a brute force search for existing resources. Currently this is only set up for the image resources and accessed through pdf_add_image_res. If a match is found, the reference object is returned. If no match is found NULL is returned and the ref object created in pdf_add_image_res is added into the hash table. In this case, a command line such as create -o output.pdf -f F0:font.ttf -i Im0:image.jpg -i Im1:image1.jpg \\ -i Im2:image.jpg contents.txt will avoid the insertion of two copies of image.jpg into the output PDF document. CID Identity-H Font added for handing ttf This adds a method for adding a ttf to a PDF as a CID font with Identity-H mapping and a ToUnicode entry that is created using FT_Get_Char_Index This takes much care in the creation of the ToUnicode CMap to ensure that the minimum number of entries are created in that we try to use beginbfrange as much as possible before using beginbfchar. The code makes sure to limit the number of entries in a group to 100 and to not cross first-byte boundaries for the CID values as described in the Adobe Technical note 5411. Add missing file pdf-resources.c pdf-resources.c was missing and should have been committed earlier. Added to windows project file. Not sure where else it needs to be added for the other platforms. Clean up names and spacing Make sure that the visible functions have the proper namespace (e.g. pdf_xxxx) Also make sure we have a blank line prior to comment. Be consistent with static function naming in pdf_resources.c pdfwrite make use of image resource fz_hash_table The pdfwrite device now shares the structure that stores the resource images for pdfcreate. With this fix, pdfwrite now avoids duplicating the writing of the same images that are shared across multiple pages. Add missing file pdf-resources.c Initial work toward having pdfwrite use Identity-H Type0 encoding for fonts Finish of CID type0 Identity-H font for pdfwrite This adds in the proper widths which may have been stored in the source font in the width table (parsed from the W entry in the pdf file) or if the free type structure has its own cmap then we can get the width from free type. Widths are restructured into format described in 5.6.3 of PDF spec. Fix issue from conflict merging and multiple define of structure Clean up warnings and make mutool create use simple font
-rw-r--r--Makefile2
-rw-r--r--include/mupdf/fitz/font.h2
-rw-r--r--include/mupdf/pdf/document.h5
-rw-r--r--include/mupdf/pdf/font.h6
-rw-r--r--include/mupdf/pdf/resource.h42
-rw-r--r--platform/win32/libmupdf.vcproj8
-rw-r--r--platform/win32/mutool.vcproj4
-rw-r--r--platform/x11/pdfapp.c2
-rw-r--r--resources/pdf/names.txt2
-rw-r--r--source/pdf/pdf-device.c444
-rw-r--r--source/pdf/pdf-font.c1003
-rw-r--r--source/pdf/pdf-form.c2
-rw-r--r--source/pdf/pdf-image.c175
-rw-r--r--source/pdf/pdf-page.c21
-rw-r--r--source/pdf/pdf-resources.c295
-rw-r--r--source/pdf/pdf-xref.c3
-rw-r--r--source/tools/mudraw.c2
-rw-r--r--source/tools/mutool.c2
-rw-r--r--source/tools/pdfcreate.c516
19 files changed, 2258 insertions, 278 deletions
diff --git a/Makefile b/Makefile
index b742f7d8..7dfeed61 100644
--- a/Makefile
+++ b/Makefile
@@ -232,7 +232,7 @@ $(OUT)/cmapdump.o : include/mupdf/pdf/cmap.h source/pdf/pdf-cmap.c source/pdf/pd
# --- Tools and Apps ---
MUTOOL := $(addprefix $(OUT)/, mutool)
-MUTOOL_OBJ := $(addprefix $(OUT)/tools/, mutool.o mudraw.o pdfclean.o pdfextract.o pdfinfo.o pdfposter.o pdfshow.o pdfpages.o)
+MUTOOL_OBJ := $(addprefix $(OUT)/tools/, mutool.o mudraw.o pdfclean.o pdfcreate.o pdfextract.o pdfinfo.o pdfposter.o pdfshow.o pdfpages.o)
$(MUTOOL_OBJ): $(FITZ_HDR) $(PDF_HDR)
$(MUTOOL) : $(MUPDF_LIB) $(THIRD_LIB)
$(MUTOOL) : $(MUTOOL_OBJ)
diff --git a/include/mupdf/fitz/font.h b/include/mupdf/fitz/font.h
index 51aa201e..ea992112 100644
--- a/include/mupdf/fitz/font.h
+++ b/include/mupdf/fitz/font.h
@@ -68,6 +68,8 @@ struct fz_font_s
int width_count;
short width_default; /* in 1000 units */
short *width_table; /* in 1000 units */
+ int first_width;
+ int last_width;
/* cached glyph metrics */
float *advance_cache;
diff --git a/include/mupdf/pdf/document.h b/include/mupdf/pdf/document.h
index 525b7b76..7f6a63d9 100644
--- a/include/mupdf/pdf/document.h
+++ b/include/mupdf/pdf/document.h
@@ -12,6 +12,7 @@ typedef struct pdf_annot_s pdf_annot;
typedef struct pdf_widget_s pdf_widget;
typedef struct pdf_hotspot_s pdf_hotspot;
typedef struct pdf_js_s pdf_js;
+typedef struct pdf_resource_tables_s pdf_resource_tables;
enum
{
@@ -255,6 +256,8 @@ struct pdf_document_s
int num_type3_fonts;
int max_type3_fonts;
fz_font **type3_fonts;
+
+ pdf_resource_tables *resources;
};
/*
@@ -266,7 +269,7 @@ struct pdf_document_s
*/
pdf_document *pdf_create_document(fz_context *ctx);
-pdf_page *pdf_create_page(fz_context *ctx, pdf_document *doc, fz_rect rect, int res, int rotate);
+pdf_page *pdf_create_page(fz_context *ctx, pdf_document *doc, fz_rect rect, fz_buffer *buffer, int rotate);
void pdf_insert_page(fz_context *ctx, pdf_document *doc, pdf_page *page, int at);
diff --git a/include/mupdf/pdf/font.h b/include/mupdf/pdf/font.h
index 17902198..f472f188 100644
--- a/include/mupdf/pdf/font.h
+++ b/include/mupdf/pdf/font.h
@@ -63,6 +63,7 @@ struct pdf_font_desc_s
float cap_height;
float x_height;
float missing_width;
+ float stem_v;
/* Encoding (CMap) */
pdf_cmap *encoding;
@@ -121,4 +122,9 @@ float pdf_text_stride(fz_context *ctx, pdf_font_desc *fontdesc, float fontsize,
void pdf_run_glyph(fz_context *ctx, pdf_document *doc, pdf_obj *resources, fz_buffer *contents, fz_device *dev, const fz_matrix *ctm, void *gstate, int nestedDepth);
+pdf_res* pdf_add_simple_font_res(fz_context *ctx, pdf_document *doc, fz_buffer *buffer);
+pdf_res* pdf_add_cid_font_res(fz_context *ctx, pdf_document *doc, fz_buffer *buffer, fz_font *font);
+
+int pdf_font_writing_supported(fz_font *font);
+
#endif
diff --git a/include/mupdf/pdf/resource.h b/include/mupdf/pdf/resource.h
index 868bd168..443123ee 100644
--- a/include/mupdf/pdf/resource.h
+++ b/include/mupdf/pdf/resource.h
@@ -9,6 +9,46 @@ void *pdf_find_item(fz_context *ctx, fz_store_drop_fn *drop, pdf_obj *key);
void pdf_remove_item(fz_context *ctx, fz_store_drop_fn *drop, pdf_obj *key);
/*
+ * Structures used for managing resource locations and avoiding multiple
+ * occurrences when resources are added to the document. The search for existing
+ * resources will be performed when we are first trying to add an item. Object
+ * refs are stored in an fz_hash_table structure using a hash of the md5 sum of
+ * the data, enabling rapid lookup.
+ */
+typedef struct pdf_res_s pdf_res;
+
+struct pdf_res_s
+{
+ int num;
+ pdf_obj *obj;
+};
+
+typedef struct pdf_res_table_s pdf_res_table;
+typedef void* (pdf_res_search_fn)(fz_context *ctx, pdf_document *doc, pdf_res_table *list,
+ void *item, void *val);
+
+struct pdf_res_table_s
+{
+ int count;
+ fz_hash_table *hash;
+ pdf_res_search_fn *search;
+};
+
+struct pdf_resource_tables_s
+{
+ pdf_res_table *image;
+ pdf_res_table *font;
+ pdf_res_table *color;
+ pdf_res_table *pattern;
+ pdf_res_table *shading;
+};
+
+void* pdf_resource_table_search(fz_context *ctx, pdf_document *doc, pdf_res_table *table, void *item, void *md5);
+void pdf_resource_table_init(fz_context *ctx, pdf_document *doc);
+void pdf_resource_table_free(fz_context *ctx, pdf_document *doc);
+void* pdf_resource_table_put(fz_context *ctx, pdf_res_table *table, void *key, pdf_obj *obj);
+
+/*
* Functions, Colorspaces, Shadings and Images
*/
@@ -24,6 +64,8 @@ int pdf_is_jpx_image(fz_context *ctx, pdf_obj *dict);
fz_image *pdf_load_image(fz_context *ctx, pdf_document *doc, pdf_obj *obj);
+pdf_res* pdf_add_image_res(fz_context *ctx, pdf_document *doc, fz_image *image, int mask);
+
/*
* Pattern
*/
diff --git a/platform/win32/libmupdf.vcproj b/platform/win32/libmupdf.vcproj
index 2b444ac6..0fdf568d 100644
--- a/platform/win32/libmupdf.vcproj
+++ b/platform/win32/libmupdf.vcproj
@@ -1238,6 +1238,14 @@
RelativePath="..\..\source\pdf\pdf-repair.c"
>
</File>
+ <File
+ RelativePath="..\..\source\pdf\pdf-resources.c"
+ >
+ </File>
+ <File
+ RelativePath="..\..\source\pdf\pdf-resources.c"
+ >
+ </File>
<File
RelativePath="..\..\source\pdf\pdf-run.c"
>
diff --git a/platform/win32/mutool.vcproj b/platform/win32/mutool.vcproj
index eec5372b..d3a30200 100644
--- a/platform/win32/mutool.vcproj
+++ b/platform/win32/mutool.vcproj
@@ -474,6 +474,10 @@
RelativePath="..\..\source\tools\pdfclean.c"
>
</File>
+ <File
+ RelativePath="..\..\source\tools\pdfcreate.c"
+ >
+ </File>
<File
RelativePath="..\..\source\tools\pdfextract.c"
>
diff --git a/platform/x11/pdfapp.c b/platform/x11/pdfapp.c
index 6b227ab3..0742d66b 100644
--- a/platform/x11/pdfapp.c
+++ b/platform/x11/pdfapp.c
@@ -255,7 +255,7 @@ static int make_fake_doc(pdfapp_t *app)
bounds.x1 = app->winw;
bounds.y1 = app->winh;
- newpage = pdf_create_page(ctx, pdf, bounds, 72, 0);
+ newpage = pdf_create_page(ctx, pdf, bounds, NULL, 0);
dev = pdf_page_write(ctx, pdf, newpage);
diff --git a/resources/pdf/names.txt b/resources/pdf/names.txt
index 1eeb3fa1..a173b6bb 100644
--- a/resources/pdf/names.txt
+++ b/resources/pdf/names.txt
@@ -300,12 +300,14 @@ Squiggly
Stamp
Standard
StdCF
+StemV
StmF
StrF
StrikeOut
SubFilter
Subtype
Subtype2
+Supplement
T
TR
TR2
diff --git a/source/pdf/pdf-device.c b/source/pdf/pdf-device.c
index 85e2f27d..7cc4e5ea 100644
--- a/source/pdf/pdf-device.c
+++ b/source/pdf/pdf-device.c
@@ -36,14 +36,15 @@ struct gstate_s
fz_matrix tm;
};
-typedef struct image_entry_s image_entry;
-
-struct image_entry_s
-{
- char digest[16];
- int id;
- pdf_obj *ref;
-};
+/* The image digest information, object reference, as well as indirect reference
+ * ID are all stored in doc->resources->image, and so they are maintained
+ * through the life of the document not just this page level device. As we
+ * encounter images on a page, we will add to the hash table if they are not
+ * already present. When we have an image on a particular page, the resource
+ * dict will be updated with the proper indirect reference across the document.
+ * We do need to maintain some information as to what image resources we have
+ * already specified for this page which is the purpose of image_indices
+ */
typedef struct alpha_entry_s alpha_entry;
@@ -91,12 +92,17 @@ struct pdf_device_s
int num_imgs;
int max_imgs;
- image_entry *images;
+ int *image_indices;
+
+ int num_cid_fonts;
+ int max_cid_fonts;
+ int *font_indices;
int num_alphas;
int max_alphas;
alpha_entry *alphas;
+ /* Base font information */
int num_fonts;
int max_fonts;
font_entry *fonts;
@@ -109,208 +115,6 @@ struct pdf_device_s
#define CURRENT_GSTATE(pdev) (&(pdev)->gstates[(pdev)->num_gstates-1])
/* Helper functions */
-
-static int
-send_image(fz_context *ctx, pdf_device *pdev, fz_image *image, int mask, int smask)
-{
- fz_pixmap *pixmap = NULL;
- pdf_obj *imobj = NULL;
- pdf_obj *imref = NULL;
- fz_compressed_buffer *cbuffer = NULL;
- fz_compression_params *cp = NULL;
- fz_buffer *buffer = NULL;
- int i, num;
- fz_md5 state;
- unsigned char digest[16];
- fz_colorspace *colorspace = image->colorspace;
- pdf_document *doc = pdev->doc;
-
- /* If we can maintain compression, do so */
- cbuffer = image->buffer;
-
- fz_var(pixmap);
- fz_var(buffer);
- fz_var(imobj);
- fz_var(imref);
-
- fz_try(ctx)
- {
- if (cbuffer != NULL && cbuffer->params.type != FZ_IMAGE_PNG && cbuffer->params.type != FZ_IMAGE_TIFF)
- {
- buffer = fz_keep_buffer(ctx, cbuffer->buffer);
- cp = &cbuffer->params;
- }
- else
- {
- unsigned int size;
- int n;
- /* Currently, set to maintain resolution; should we consider
- * subsampling here according to desired output res? */
- pixmap = fz_get_pixmap_from_image(ctx, image, image->w, image->h);
- colorspace = pixmap->colorspace; /* May be different to image->colorspace! */
- n = (pixmap->n == 1 ? 1 : pixmap->n-1);
- size = image->w * image->h * n;
- buffer = fz_new_buffer(ctx, size);
- buffer->len = size;
- if (pixmap->n == 1)
- {
- memcpy(buffer->data, pixmap->samples, size);
- }
- else
- {
- /* Need to remove the alpha plane */
- unsigned char *d = buffer->data;
- unsigned char *s = pixmap->samples;
- int mod = n;
- while (size--)
- {
- *d++ = *s++;
- mod--;
- if (mod == 0)
- s++, mod = n;
- }
- }
- }
-
- fz_md5_init(&state);
- fz_md5_update(&state, buffer->data, buffer->len);
- fz_md5_final(&state, digest);
- for(i=0; i < pdev->num_imgs; i++)
- {
- if (!memcmp(&digest, pdev->images[i].digest, sizeof(digest)))
- {
- num = i;
- break;
- }
- }
-
- if (i < pdev->num_imgs)
- break;
-
- if (pdev->num_imgs == pdev->max_imgs)
- {
- int newmax = pdev->max_imgs * 2;
- if (newmax == 0)
- newmax = 4;
- pdev->images = fz_resize_array(ctx, pdev->images, newmax, sizeof(*pdev->images));
- pdev->max_imgs = newmax;
- }
- num = pdev->num_imgs++;
- memcpy(pdev->images[num].digest,digest,16);
- pdev->images[num].ref = NULL; /* Will be filled in later */
-
- imobj = pdf_new_dict(ctx, doc, 3);
- pdf_dict_put_drop(ctx, imobj, PDF_NAME_Type, PDF_NAME_XObject);
- pdf_dict_put_drop(ctx, imobj, PDF_NAME_Subtype, PDF_NAME_Image);
- pdf_dict_put_drop(ctx, imobj, PDF_NAME_Width, pdf_new_int(ctx, doc, image->w));
- pdf_dict_put_drop(ctx, imobj, PDF_NAME_Height, pdf_new_int(ctx, doc, image->h));
- if (mask)
- {}
- else if (!colorspace || colorspace->n == 1)
- pdf_dict_put_drop(ctx, imobj, PDF_NAME_ColorSpace, PDF_NAME_DeviceGray);
- else if (colorspace->n == 3)
- pdf_dict_put_drop(ctx, imobj, PDF_NAME_ColorSpace, PDF_NAME_DeviceRGB);
- else if (colorspace->n == 4)
- pdf_dict_put_drop(ctx, imobj, PDF_NAME_ColorSpace, PDF_NAME_DeviceCMYK);
- if (!mask)
- pdf_dict_put_drop(ctx, imobj, PDF_NAME_BitsPerComponent, pdf_new_int(ctx, doc, image->bpc));
- switch (cp ? cp->type : FZ_IMAGE_UNKNOWN)
- {
- case FZ_IMAGE_UNKNOWN: /* Unknown also means raw */
- default:
- break;
- case FZ_IMAGE_JPEG:
- if (cp->u.jpeg.color_transform != -1)
- pdf_dict_put_drop(ctx, imobj, PDF_NAME_ColorTransform, pdf_new_int(ctx, doc, cp->u.jpeg.color_transform));
- pdf_dict_put_drop(ctx, imobj, PDF_NAME_Filter, PDF_NAME_DCTDecode);
- break;
- case FZ_IMAGE_JPX:
- if (cp->u.jpx.smask_in_data)
- pdf_dict_put_drop(ctx, imobj, PDF_NAME_SMaskInData, pdf_new_int(ctx, doc, cp->u.jpx.smask_in_data));
- pdf_dict_put_drop(ctx, imobj, PDF_NAME_Filter, PDF_NAME_JPXDecode);
- break;
- case FZ_IMAGE_FAX:
- if (cp->u.fax.columns)
- pdf_dict_put_drop(ctx, imobj, PDF_NAME_Columns, pdf_new_int(ctx, doc, cp->u.fax.columns));
- if (cp->u.fax.rows)
- pdf_dict_put_drop(ctx, imobj, PDF_NAME_Rows, pdf_new_int(ctx, doc, cp->u.fax.rows));
- if (cp->u.fax.k)
- pdf_dict_put_drop(ctx, imobj, PDF_NAME_K, pdf_new_int(ctx, doc, cp->u.fax.k));
- if (cp->u.fax.end_of_line)
- pdf_dict_put_drop(ctx, imobj, PDF_NAME_EndOfLine, pdf_new_int(ctx, doc, cp->u.fax.end_of_line));
- if (cp->u.fax.encoded_byte_align)
- pdf_dict_put_drop(ctx, imobj, PDF_NAME_EncodedByteAlign, pdf_new_int(ctx, doc, cp->u.fax.encoded_byte_align));
- if (cp->u.fax.end_of_block)
- pdf_dict_put_drop(ctx, imobj, PDF_NAME_EndOfBlock, pdf_new_int(ctx, doc, cp->u.fax.end_of_block));
- if (cp->u.fax.black_is_1)
- pdf_dict_put_drop(ctx, imobj, PDF_NAME_BlackIs1, pdf_new_int(ctx, doc, cp->u.fax.black_is_1));
- if (cp->u.fax.damaged_rows_before_error)
- pdf_dict_put_drop(ctx, imobj, PDF_NAME_DamagedRowsBeforeError, pdf_new_int(ctx, doc, cp->u.fax.damaged_rows_before_error));
- pdf_dict_put_drop(ctx, imobj, PDF_NAME_Filter, PDF_NAME_CCITTFaxDecode);
- break;
- case FZ_IMAGE_JBIG2:
- /* FIXME - jbig2globals */
- cp->type = FZ_IMAGE_UNKNOWN;
- break;
- case FZ_IMAGE_FLATE:
- if (cp->u.flate.columns)
- pdf_dict_put_drop(ctx, imobj, PDF_NAME_Columns, pdf_new_int(ctx, doc, cp->u.flate.columns));
- if (cp->u.flate.colors)
- pdf_dict_put_drop(ctx, imobj, PDF_NAME_Colors, pdf_new_int(ctx, doc, cp->u.flate.colors));
- if (cp->u.flate.predictor)
- pdf_dict_put_drop(ctx, imobj, PDF_NAME_Predictor, pdf_new_int(ctx, doc, cp->u.flate.predictor));
- pdf_dict_put_drop(ctx, imobj, PDF_NAME_Filter, PDF_NAME_FlateDecode);
- pdf_dict_put_drop(ctx, imobj, PDF_NAME_BitsPerComponent, pdf_new_int(ctx, doc, image->bpc));
- break;
- case FZ_IMAGE_LZW:
- if (cp->u.lzw.columns)
- pdf_dict_put_drop(ctx, imobj, PDF_NAME_Columns, pdf_new_int(ctx, doc, cp->u.lzw.columns));
- if (cp->u.lzw.colors)
- pdf_dict_put_drop(ctx, imobj, PDF_NAME_Colors, pdf_new_int(ctx, doc, cp->u.lzw.colors));
- if (cp->u.lzw.predictor)
- pdf_dict_put_drop(ctx, imobj, PDF_NAME_Predictor, pdf_new_int(ctx, doc, cp->u.lzw.predictor));
- if (cp->u.lzw.early_change)
- pdf_dict_put_drop(ctx, imobj, PDF_NAME_EarlyChange, pdf_new_int(ctx, doc, cp->u.lzw.early_change));
- pdf_dict_put_drop(ctx, imobj, PDF_NAME_Filter, PDF_NAME_LZWDecode);
- break;
- case FZ_IMAGE_RLD:
- pdf_dict_put_drop(ctx, imobj, PDF_NAME_Filter, PDF_NAME_RunLengthDecode);
- break;
- }
- if (mask)
- {
- pdf_dict_put_drop(ctx, imobj, PDF_NAME_ImageMask, pdf_new_bool(ctx, doc, 1));
- }
- if (image->mask)
- {
- int smasknum = send_image(ctx, pdev, image->mask, 0, 1);
- pdf_dict_put(ctx, imobj, PDF_NAME_SMask, pdev->images[smasknum].ref);
- }
-
- imref = pdf_new_ref(ctx, doc, imobj);
- pdf_update_stream(ctx, doc, imref, buffer, 1);
-
- {
- char text[32];
- snprintf(text, sizeof(text), "XObject/Img%d", num);
- pdf_dict_putp(ctx, pdev->resources, text, imref);
- }
- pdev->images[num].ref = imref;
- }
- fz_always(ctx)
- {
- fz_drop_buffer(ctx, buffer);
- pdf_drop_obj(ctx, imobj);
- fz_drop_pixmap(ctx, pixmap);
- }
- fz_catch(ctx)
- {
- pdf_drop_obj(ctx, imref);
- fz_rethrow(ctx);
- }
- return num;
-}
-
static void
pdf_dev_stroke_state(fz_context *ctx, pdf_device *pdev, const fz_stroke_state *stroke_state)
{
@@ -550,66 +354,116 @@ pdf_dev_alpha(fz_context *ctx, pdf_device *pdev, float alpha, int stroke)
}
static void
+pdf_dev_add_font_res(fz_context *ctx, fz_device *dev, pdf_res *fres)
+{
+ char text[32];
+ pdf_device *pdev = (pdf_device*)dev;
+ int k;
+ int num;
+
+ /* Check if we already had this one */
+ for (k = 0; k < pdev->num_cid_fonts; k++)
+ {
+ if (pdev->font_indices[k] == fres->num)
+ return;
+ }
+
+ /* Not there so add to resources */
+ snprintf(text, sizeof(text), "Font/F%d", fres->num);
+ pdf_dict_putp(ctx, pdev->resources, text, fres->obj);
+
+ /* And add index to our list for this page */
+ if (pdev->num_cid_fonts == pdev->max_cid_fonts)
+ {
+ int newmax = pdev->max_cid_fonts * 2;
+ if (newmax == 0)
+ newmax = 4;
+ pdev->font_indices = fz_resize_array(ctx, pdev->image_indices, newmax, sizeof(*pdev->font_indices));
+ pdev->max_cid_fonts = newmax;
+ }
+ num = pdev->num_cid_fonts++;
+ pdev->font_indices[num] = fres->num;
+}
+
+static void
pdf_dev_font(fz_context *ctx, pdf_device *pdev, fz_font *font, float size)
{
int i;
pdf_document *doc = pdev->doc;
gstate *gs = CURRENT_GSTATE(pdev);
+ pdf_res *fres;
/* If the font is unchanged, nothing to do */
if (gs->font >= 0 && pdev->fonts[gs->font].font == font)
return;
- if (font->ft_buffer != NULL || font->ft_substitute)
- fz_throw(ctx, FZ_ERROR_GENERIC, "pdf device supports only base 14 fonts currently");
+ if (font->ft_substitute)
+ fz_throw(ctx, FZ_ERROR_GENERIC, "pdf device does not support substitute metrics");
- /* Have we sent such a font before? */
- for (i = 0; i < pdev->num_fonts; i++)
- if (pdev->fonts[i].font == font)
- break;
+ if (font->ft_buffer != NULL && !pdf_font_writing_supported(font))
+ fz_throw(ctx, FZ_ERROR_GENERIC, "pdf device does not support font types found in this file");
- if (i == pdev->num_fonts)
+ if (font->ft_buffer != NULL)
{
- pdf_obj *o;
- pdf_obj *ref = NULL;
+ /* This will add it to the xref if needed */
+ fres = pdf_add_cid_font_res(ctx, doc, font->ft_buffer, font);
+ fz_buffer_printf(ctx, gs->buf, "/F%d %f Tf\n", fres->num, size);
- fz_var(ref);
+ /* Possibly add to page resources */
+ pdf_dev_add_font_res(ctx, (fz_device*) pdev, fres);
+ pdf_drop_obj(ctx, fres->obj);
+ }
+ else
+ {
+ /* Have the device handle the base fonts */
+ /* Have we sent such a font before? */
+ for (i = 0; i < pdev->num_fonts; i++)
+ if (pdev->fonts[i].font == font)
+ break;
- /* No. Need to make a new one */
- if (pdev->num_fonts == pdev->max_fonts)
+ if (i == pdev->num_fonts)
{
- int newmax = pdev->max_fonts * 2;
- if (newmax == 0)
- newmax = 4;
- pdev->fonts = fz_resize_array(ctx, pdev->fonts, newmax, sizeof(*pdev->fonts));
- pdev->max_fonts = newmax;
- }
- pdev->fonts[i].font = fz_keep_font(ctx, font);
+ pdf_obj *o;
+ pdf_obj *ref = NULL;
- o = pdf_new_dict(ctx, doc, 3);
- fz_try(ctx)
- {
- char text[32];
- pdf_dict_put_drop(ctx, o, PDF_NAME_Type, PDF_NAME_Font);
- pdf_dict_put_drop(ctx, o, PDF_NAME_Subtype, PDF_NAME_Type1);
- pdf_dict_put_drop(ctx, o, PDF_NAME_BaseFont, pdf_new_name(ctx, doc, font->name));
- pdf_dict_put_drop(ctx, o, PDF_NAME_Encoding, PDF_NAME_WinAnsiEncoding);
- ref = pdf_new_ref(ctx, doc, o);
- snprintf(text, sizeof(text), "Font/F%d", i);
- pdf_dict_putp(ctx, pdev->resources, text, ref);
- }
- fz_always(ctx)
- {
- pdf_drop_obj(ctx, o);
- pdf_drop_obj(ctx, ref);
- }
- fz_catch(ctx)
- {
- fz_rethrow(ctx);
+ fz_var(ref);
+
+ /* No. Need to make a new one */
+ if (pdev->num_fonts == pdev->max_fonts)
+ {
+ int newmax = pdev->max_fonts * 2;
+ if (newmax == 0)
+ newmax = 4;
+ pdev->fonts = fz_resize_array(ctx, pdev->fonts, newmax, sizeof(*pdev->fonts));
+ pdev->max_fonts = newmax;
+ }
+ pdev->fonts[i].font = fz_keep_font(ctx, font);
+
+ o = pdf_new_dict(ctx, doc, 3);
+ fz_try(ctx)
+ {
+ char text[32];
+ pdf_dict_put_drop(ctx, o, PDF_NAME_Type, PDF_NAME_Font);
+ pdf_dict_put_drop(ctx, o, PDF_NAME_Subtype, PDF_NAME_Type1);
+ pdf_dict_put_drop(ctx, o, PDF_NAME_BaseFont, pdf_new_name(ctx, doc, font->name));
+ pdf_dict_put_drop(ctx, o, PDF_NAME_Encoding, PDF_NAME_WinAnsiEncoding);
+ ref = pdf_new_ref(ctx, doc, o);
+ snprintf(text, sizeof(text), "Font/Fb%d", i);
+ pdf_dict_putp(ctx, pdev->resources, text, ref);
+ }
+ fz_always(ctx)
+ {
+ pdf_drop_obj(ctx, o);
+ pdf_drop_obj(ctx, ref);
+ }
+ fz_catch(ctx)
+ {
+ fz_rethrow(ctx);
+ }
+ pdev->num_fonts++;
}
- pdev->num_fonts++;
+ fz_buffer_printf(ctx, gs->buf, "/Fb%d %f Tf\n", i, size);
}
- fz_buffer_printf(ctx, gs->buf, "/F%d %f Tf\n", i, size);
}
static void
@@ -721,11 +575,21 @@ pdf_dev_text_span(fz_context *ctx, pdf_device *pdev, fz_text_span *span, float s
}
fz_buffer_printf(ctx, gs->buf, "<");
- for (/* i from its current value */; i < j; i++)
+ if (span->font->ft_buffer == NULL)
+ {
+ /* A standard 14 type font */
+ for (/* i from its current value */; i < j; i++)
+ {
+ fz_buffer_printf(ctx, gs->buf, "%02x", span->items[i].ucs);
+ }
+ }
+ else
{
- /* FIXME: should use it->gid, rather than it->ucs, and convert
- * to the correct encoding */
- fz_buffer_printf(ctx, gs->buf, "%02x", span->items[i].ucs);
+ /* Non-standard font. Saved as Type0 Identity-H */
+ for (/* i from its current value */; i < j; i++)
+ {
+ fz_buffer_printf(ctx, gs->buf, "%04x", span->items[i].gid);
+ }
}
fz_buffer_printf(ctx, gs->buf, "> Tj\n");
}
@@ -1040,21 +904,63 @@ pdf_dev_ignore_text(fz_context *ctx, fz_device *dev, const fz_text *text, const
}
static void
-pdf_dev_fill_image(fz_context *ctx, fz_device *dev, fz_image *image, const fz_matrix *ctm, float alpha)
+pdf_dev_add_image_res(fz_context *ctx, fz_device *dev, pdf_res *im_res)
{
+ char text[32];
pdf_device *pdev = (pdf_device*)dev;
+ int k;
int num;
+
+ /* Check if we already had this one */
+ for (k = 0; k < pdev->num_imgs; k++)
+ {
+ if (pdev->image_indices[k] == im_res->num)
+ return;
+ }
+
+ /* Not there so add to resources */
+ snprintf(text, sizeof(text), "XObject/Img%d", im_res->num);
+ pdf_dict_putp(ctx, pdev->resources, text, im_res->obj);
+
+ /* And add index to our list for this page */
+ if (pdev->num_imgs == pdev->max_imgs)
+ {
+ int newmax = pdev->max_imgs * 2;
+ if (newmax == 0)
+ newmax = 4;
+ pdev->image_indices = fz_resize_array(ctx, pdev->image_indices, newmax, sizeof(*pdev->image_indices));
+ pdev->max_imgs = newmax;
+ }
+ num = pdev->num_imgs++;
+ pdev->image_indices[num] = im_res->num;
+}
+
+static void
+pdf_dev_fill_image(fz_context *ctx, fz_device *dev, fz_image *image, const fz_matrix *ctm, float alpha)
+{
+ pdf_device *pdev = (pdf_device*)dev;
+ pdf_res *im_res;
gstate *gs = CURRENT_GSTATE(pdev);
fz_matrix local_ctm = *ctm;
pdf_dev_end_text(ctx, pdev);
- num = send_image(ctx, pdev, image, 0, 0);
+ im_res = pdf_add_image_res(ctx, pdev->doc, image, 0);
+ if (im_res == NULL)
+ {
+ fz_warn(ctx, "pdf_add_image_res: problem adding image resource");
+ return;
+ }
pdf_dev_alpha(ctx, pdev, alpha, 0);
+
/* PDF images are upside down, so fiddle the ctm */
fz_pre_scale(&local_ctm, 1, -1);
fz_pre_translate(&local_ctm, 0, -1);
pdf_dev_ctm(ctx, pdev, &local_ctm);
- fz_buffer_printf(ctx, gs->buf, "/Img%d Do\n", num);
+ fz_buffer_printf(ctx, gs->buf, "/Img%d Do\n", im_res->num);
+
+ /* Possibly add to page resources */
+ pdf_dev_add_image_res(ctx, dev, im_res);
+ pdf_drop_obj(ctx, im_res->obj);
}
static void
@@ -1071,20 +977,26 @@ pdf_dev_fill_image_mask(fz_context *ctx, fz_device *dev, fz_image *image, const
fz_colorspace *colorspace, const float *color, float alpha)
{
pdf_device *pdev = (pdf_device*)dev;
+ pdf_res* im_res = NULL;
gstate *gs = CURRENT_GSTATE(pdev);
- int num;
fz_matrix local_ctm = *ctm;
pdf_dev_end_text(ctx, pdev);
- num = send_image(ctx, pdev, image, 1, 0);
+ im_res = pdf_add_image_res(ctx, pdev->doc, image, 1);
+ if (im_res == NULL)
+ {
+ fz_warn(ctx, "pdf_add_image_res: problem adding image resource");
+ return;
+ }
fz_buffer_printf(ctx, gs->buf, "q\n");
pdf_dev_alpha(ctx, pdev, alpha, 0);
pdf_dev_color(ctx, pdev, colorspace, color, 0);
+
/* PDF images are upside down, so fiddle the ctm */
fz_pre_scale(&local_ctm, 1, -1);
fz_pre_translate(&local_ctm, 0, -1);
pdf_dev_ctm(ctx, pdev, &local_ctm);
- fz_buffer_printf(ctx, gs->buf, "/Img%d Do Q\n", num);
+ fz_buffer_printf(ctx, gs->buf, "/Img%d Do Q\n", im_res->num);
}
static void
@@ -1283,11 +1195,6 @@ pdf_dev_drop_imp(fz_context *ctx, fz_device *dev)
fz_drop_font(ctx, pdev->fonts[i].font);
}
- for (i = pdev->num_imgs-1; i >= 0; i--)
- {
- pdf_drop_obj(ctx, pdev->images[i].ref);
- }
-
for (i = pdev->num_groups - 1; i >= 0; i--)
{
pdf_drop_obj(ctx, pdev->groups[i].ref);
@@ -1306,9 +1213,10 @@ pdf_dev_drop_imp(fz_context *ctx, fz_device *dev)
pdf_drop_obj(ctx, pdev->resources);
+ fz_free(ctx, pdev->font_indices);
+ fz_free(ctx, pdev->image_indices);
fz_free(ctx, pdev->groups);
fz_free(ctx, pdev->fonts);
- fz_free(ctx, pdev->images);
fz_free(ctx, pdev->alphas);
fz_free(ctx, pdev->gstates);
}
diff --git a/source/pdf/pdf-font.c b/source/pdf/pdf-font.c
index f516468d..5d8cc392 100644
--- a/source/pdf/pdf-font.c
+++ b/source/pdf/pdf-font.c
@@ -1,4 +1,5 @@
#include "mupdf/pdf.h"
+#include <zlib.h>
#include <ft2build.h>
#include FT_FREETYPE_H
@@ -1366,3 +1367,1005 @@ float pdf_text_stride(fz_context *ctx, pdf_font_desc *fontdesc, float fontsize,
return x;
}
+
+/* Populate font description. According to spec, required for Latin fonts are
+ * FontName, Flags, FontBBox, ItalicAngle, Ascent, Descent, CapHeight (Latin),
+ * StemV */
+static void
+pdf_fontdesc_init(fz_context *ctx, pdf_font_desc *fontdesc)
+{
+ FT_Face face;
+ int fterr;
+ int row_pos;
+ unsigned char *ptr;
+ unsigned int k;
+ int count = 0;
+ int width;
+
+ if (fontdesc->font == NULL || fontdesc->font->ft_face == NULL)
+ fz_throw(ctx, FZ_ERROR_GENERIC, "cannot create font description");
+
+ face = fontdesc->font->ft_face;
+ fontdesc->ascent = face->ascender * 1000.0f / face->units_per_EM;
+ fontdesc->descent = face->descender * 1000.0f / face->units_per_EM;
+ fontdesc->italic_angle = 0; /* 0 for now */
+
+ /* Get the cap height and stem thickness from capital O. Obviously an
+ .* issue if this is not a latin font */
+ fterr = FT_Load_Char(fontdesc->font->ft_face, 'O',
+ FT_LOAD_NO_HINTING | FT_LOAD_IGNORE_TRANSFORM);
+ if (fterr)
+ {
+ fz_warn(ctx, "freetype load char O: %s", ft_error_string(fterr));
+ return;
+ }
+ fontdesc->cap_height = face->glyph->metrics.height;
+ width = face->glyph->metrics.width;
+
+ fterr = FT_Set_Char_Size(fontdesc->font->ft_face, 1000, 1000, 300, 300);
+ if (fterr)
+ {
+ fz_warn(ctx, "freetype set char size: %s", ft_error_string(fterr));
+ return;
+ }
+ fterr = FT_Load_Char(fontdesc->font->ft_face, 'O', 0);
+ if (fterr)
+ {
+ fz_warn(ctx, "freetype load char O: %s", ft_error_string(fterr));
+ return;
+ }
+ fterr = FT_Render_Glyph(face->glyph, FT_RENDER_MODE_NORMAL);
+ if (fterr)
+ {
+ fz_warn(ctx, "freetype render glyph O: %s", ft_error_string(fterr));
+ return;
+ }
+ row_pos = floor(face->glyph->bitmap.rows / 2.0 + 0.5);
+ ptr = face->glyph->bitmap.buffer + row_pos * face->glyph->bitmap.pitch;
+ for (k = 0; k < face->glyph->bitmap.width; k++)
+ {
+ count += 1;
+ if (ptr[k] == 0)
+ break;
+ }
+ count -= 1;
+ fontdesc->stem_v = width * (float)count / (float)face->glyph->bitmap.width;
+ fontdesc->flags = PDF_FD_NONSYMBOLIC; /* ToDo: FixMe. Set non-symbolic always for now */
+}
+
+static void ft_width_for_simple_table(fz_context *ctx, pdf_font_desc *fontdesc,
+ int *firstcode)
+{
+ FT_ULong charcode;
+ FT_UInt gindex;
+ int size;
+ int fterr;
+
+ *firstcode = 0;
+ fterr = FT_Select_Charmap(fontdesc->font->ft_face, ft_encoding_unicode);
+ if (fterr)
+ {
+ fz_warn(ctx, "freetype select char map: %s", ft_error_string(fterr));
+ return;
+ }
+ fterr = FT_Set_Char_Size(fontdesc->font->ft_face, 1000, 1000, 72, 72);
+ if (fterr)
+ {
+ fz_warn(ctx, "freetype set char size: %s", ft_error_string(fterr));
+ return;
+ }
+ charcode = FT_Get_First_Char(fontdesc->font->ft_face, &gindex);
+ *firstcode = charcode;
+ fontdesc->font->first_width = charcode;
+ while (gindex != 0)
+ {
+ fterr = FT_Load_Char(fontdesc->font->ft_face, charcode,
+ FT_LOAD_NO_HINTING | FT_LOAD_NO_BITMAP | FT_LOAD_IGNORE_TRANSFORM);
+ if (fterr)
+ {
+ fz_warn(ctx, "freetype load char (charcode %ld): %s", charcode, ft_error_string(fterr));
+ return;
+ }
+ size = ((FT_Face)fontdesc->font->ft_face)->glyph->metrics.horiAdvance;
+ if (charcode < 256)
+ {
+ fontdesc->font->width_table[charcode] = size;
+ fontdesc->font->last_width = charcode;
+ }
+ charcode = FT_Get_Next_Char(fontdesc->font->ft_face, charcode, &gindex);
+ }
+}
+
+/* This needs to go away and we need to denote when the buffer is or is not
+* deflated */
+static fz_buffer *deflatebuf(fz_context *ctx, unsigned char *p, int n)
+{
+ fz_buffer *buf;
+ uLongf csize;
+ int t;
+
+ buf = fz_new_buffer(ctx, compressBound(n));
+ csize = buf->cap;
+ t = compress(buf->data, &csize, p, n);
+ if (t != Z_OK)
+ {
+ fz_drop_buffer(ctx, buf);
+ fz_throw(ctx, FZ_ERROR_GENERIC, "cannot deflate buffer");
+ }
+ buf->len = csize;
+ return buf;
+}
+
+static pdf_obj*
+pdf_font_stream_ref(fz_context *ctx, pdf_document *doc, fz_buffer *buf)
+{
+ pdf_obj *obj = NULL;
+ pdf_obj *newlen = NULL;
+ pdf_obj *currlen = NULL;
+ pdf_obj *ref = NULL;
+ fz_buffer *d_buf = NULL;
+
+ fz_var(obj);
+ fz_var(newlen);
+ fz_var(currlen);
+ fz_var(ref);
+
+ fz_try(ctx)
+ {
+ obj = pdf_new_dict(ctx, doc, 3);
+ pdf_dict_put_drop(ctx, obj, PDF_NAME_Filter, PDF_NAME_FlateDecode);
+ currlen = pdf_new_int(ctx, doc, buf->len);
+ d_buf = deflatebuf(ctx, buf->data, buf->len);
+ newlen = pdf_new_int(ctx, doc, d_buf->len);
+ pdf_dict_put_drop(ctx, obj, PDF_NAME_Length, newlen);
+ pdf_dict_put_drop(ctx, obj, PDF_NAME_Length1, currlen);
+ ref = pdf_new_ref(ctx, doc, obj);
+ pdf_update_stream(ctx, doc, ref, d_buf, 1);
+ }
+ fz_always(ctx)
+ {
+ fz_drop_buffer(ctx, d_buf);
+ pdf_drop_obj(ctx, obj);
+ }
+ fz_catch(ctx)
+ {
+ pdf_drop_obj(ctx, newlen);
+ pdf_drop_obj(ctx, currlen);
+ pdf_drop_obj(ctx, ref);
+ fz_rethrow(ctx);
+ }
+ return ref;
+}
+
+static pdf_obj*
+pdf_font_desc_ref(fz_context *ctx, pdf_document *doc, pdf_font_desc *fontdesc, pdf_obj *fileref)
+{
+ pdf_obj *ref = NULL;
+ pdf_obj *fdobj = NULL;
+ pdf_obj *bbox = NULL;
+ FT_Face face;
+
+ if (fontdesc->font == NULL || fontdesc->font->ft_face == NULL)
+ fz_throw(ctx, FZ_ERROR_GENERIC, "cannot create font description");
+ face = fontdesc->font->ft_face;
+
+ fz_try(ctx)
+ {
+ fdobj = pdf_new_dict(ctx, doc, 10);
+ pdf_dict_put_drop(ctx, fdobj, PDF_NAME_Type, PDF_NAME_FontDescriptor);
+ pdf_dict_put_drop(ctx, fdobj, PDF_NAME_FontName, pdf_new_name(ctx, doc, fontdesc->font->name));
+
+ bbox = pdf_new_array(ctx, doc, 4);
+ pdf_array_push_drop(ctx, bbox, pdf_new_real(ctx, doc, 1000.0f * fontdesc->font->bbox.x0));
+ pdf_array_push_drop(ctx, bbox, pdf_new_real(ctx, doc, 1000.0f * fontdesc->font->bbox.y0));
+ pdf_array_push_drop(ctx, bbox, pdf_new_real(ctx, doc, 1000.0f * fontdesc->font->bbox.x1));
+ pdf_array_push_drop(ctx, bbox, pdf_new_real(ctx, doc, 1000.0f * fontdesc->font->bbox.y1));
+ pdf_dict_put_drop(ctx, fdobj, PDF_NAME_FontBBox, bbox);
+ bbox = NULL;
+ pdf_dict_put_drop(ctx, fdobj, PDF_NAME_ItalicAngle, pdf_new_real(ctx, doc, fontdesc->italic_angle));
+ pdf_dict_put_drop(ctx, fdobj, PDF_NAME_Ascent, pdf_new_real(ctx, doc, fontdesc->ascent));
+ pdf_dict_put_drop(ctx, fdobj, PDF_NAME_Descent, pdf_new_real(ctx, doc, fontdesc->descent));
+ pdf_dict_put_drop(ctx, fdobj, PDF_NAME_CapHeight, pdf_new_real(ctx, doc, fontdesc->cap_height));
+ pdf_dict_put_drop(ctx, fdobj, PDF_NAME_StemV, pdf_new_real(ctx, doc, fontdesc->stem_v));
+ pdf_dict_put_drop(ctx, fdobj, PDF_NAME_Flags, pdf_new_real(ctx, doc, fontdesc->flags));
+
+ switch (ft_kind(face))
+ {
+ case TYPE1:
+ pdf_dict_put_drop(ctx, fdobj, PDF_NAME_FontFile, fileref);
+ break;
+ case TRUETYPE:
+ pdf_dict_put_drop(ctx, fdobj, PDF_NAME_FontFile2, fileref);
+ break;
+ case UNKNOWN:
+ fz_throw(ctx, FZ_ERROR_GENERIC, "Unknown font subtype");
+ break;
+ }
+ ref = pdf_new_ref(ctx, doc, fdobj);
+ }
+ fz_always(ctx)
+ {
+ pdf_drop_obj(ctx, fdobj);
+ }
+ fz_catch(ctx)
+ {
+ pdf_drop_obj(ctx, ref);
+ pdf_drop_obj(ctx, bbox);
+ fz_rethrow(ctx);
+ }
+ return ref;
+}
+
+static pdf_obj*
+pdf_font_widths_ref(fz_context *ctx, pdf_document *doc, pdf_font_desc *fontdesc)
+{
+ pdf_obj *ref = NULL;
+ pdf_obj *arr = NULL;
+ int k;
+
+ fz_var(arr);
+ fz_var(ref);
+
+ fz_try(ctx)
+ {
+ arr = pdf_new_array(ctx, doc, 256);
+ for (k = fontdesc->font->first_width; k < fontdesc->font->last_width; k++)
+ pdf_array_push_drop(ctx, arr, pdf_new_int(ctx, doc, fontdesc->font->width_table[k]));
+ ref = pdf_new_ref(ctx, doc, arr);
+ }
+ fz_always(ctx)
+ {
+ pdf_drop_obj(ctx, arr);
+ }
+ fz_catch(ctx)
+ {
+ pdf_drop_obj(ctx, ref);
+ fz_rethrow(ctx);
+ }
+ return ref;
+}
+
+static pdf_obj*
+pdf_add_cid_system_info(fz_context *ctx, pdf_document *doc)
+{
+ pdf_obj *fobj = NULL;
+ pdf_obj *ref = NULL;
+
+ fz_var(fobj);
+
+ fz_try(ctx)
+ {
+ fobj = pdf_new_dict(ctx, doc, 3);
+ pdf_dict_put_drop(ctx, fobj, PDF_NAME_Ordering, pdf_new_string(ctx, doc, "Identity", strlen("Identity")));
+ pdf_dict_put_drop(ctx, fobj, PDF_NAME_Registry, pdf_new_string(ctx, doc, "Adobe", strlen("Adobe")));
+ pdf_dict_put_drop(ctx, fobj, PDF_NAME_Supplement, pdf_new_int(ctx, doc, 0));
+ ref = pdf_new_ref(ctx, doc, fobj);
+ }
+ fz_always(ctx)
+ {
+ pdf_drop_obj(ctx, fobj);
+ }
+ fz_catch(ctx)
+ {
+ pdf_drop_obj(ctx, ref);
+ fz_rethrow(ctx);
+ }
+ return ref;
+}
+
+/* Different states of starting, same width as last, or consecutive glyph */
+enum { FW_START, FW_SAME, FW_RUN };
+
+static void
+pdf_publish_cid_widths(fz_context *ctx, pdf_document *doc, pdf_obj *fwobj, pdf_obj *run_obj, int state, int first_code, int prev_code, int prev_size)
+{
+ pdf_obj *temp_array = NULL;
+
+ fz_var(temp_array);
+
+ fz_try(ctx)
+ {
+ switch (state)
+ {
+ case FW_SAME:
+ /* Add three entries. First cid, last cid and width */
+ pdf_array_push_drop(ctx, fwobj, pdf_new_int(ctx, doc, first_code));
+ pdf_array_push_drop(ctx, fwobj, pdf_new_int(ctx, doc, prev_code));
+ pdf_array_push_drop(ctx, fwobj, pdf_new_int(ctx, doc, prev_size));
+ break;
+ case FW_RUN:
+ if (pdf_array_len(ctx, run_obj) > 0)
+ {
+ pdf_array_push_drop(ctx, fwobj, pdf_new_int(ctx, doc, first_code));
+ pdf_array_push_drop(ctx, fwobj, run_obj);
+ }
+ else
+ {
+ pdf_drop_obj(ctx, run_obj);
+ }
+ run_obj = NULL;
+ break;
+ case FW_START:
+ /* Lone wolf. Not part of a consecutive run */
+ pdf_array_push_drop(ctx, fwobj, pdf_new_int(ctx, doc, prev_code));
+ temp_array = pdf_new_array(ctx, doc, 1);
+ pdf_array_push_drop(ctx, temp_array, pdf_new_int(ctx, doc, prev_size));
+ pdf_array_push_drop(ctx, fwobj, temp_array);
+ temp_array = NULL;
+ break;
+ }
+ }
+ fz_catch(ctx)
+ {
+ pdf_drop_obj(ctx, temp_array);
+ fz_rethrow(ctx);
+ }
+}
+
+/* ToDo: Ignore the default sized characters */
+static pdf_obj*
+pdf_create_cid_widths(fz_context *ctx, pdf_document *doc, pdf_font_desc *fontdesc, fz_font *source_font)
+{
+ pdf_obj *fwobj = NULL;
+ pdf_obj *run_obj = NULL;
+ FT_ULong curr_code;
+ FT_ULong prev_code;
+ FT_UInt gindex;
+ int curr_size;
+ int prev_size;
+ int first_code;
+ int new_first_code;
+ int fterr;
+ int state = FW_START;
+ int new_state = FW_START;
+ int publish = 0;
+
+ if (fontdesc->font == NULL || fontdesc->font->ft_face == NULL || fontdesc->wmode)
+ {
+ fz_warn(ctx, "cannot compute cid widths description");
+ return NULL;
+ }
+
+ if (source_font->width_table == NULL)
+ {
+ fterr = FT_Select_Charmap(fontdesc->font->ft_face, ft_encoding_unicode);
+ if (fterr)
+ {
+ fz_warn(ctx, "freetype select char map: %s", ft_error_string(fterr));
+ return NULL;
+ }
+ fterr = FT_Set_Char_Size(fontdesc->font->ft_face, 1000, 1000, 72, 72);
+ if (fterr)
+ {
+ fz_warn(ctx, "freetype set char size: %s", ft_error_string(fterr));
+ return NULL;
+ }
+
+ /* Prime the pump. */
+ prev_code = FT_Get_First_Char(fontdesc->font->ft_face, &gindex);
+ if (gindex == 0)
+ {
+ fz_warn(ctx, "freetype FT_Get_First_Char fail: %s", ft_error_string(fterr));
+ return NULL;
+ }
+ prev_size = ((FT_Face)fontdesc->font->ft_face)->glyph->metrics.horiAdvance;
+ first_code = prev_code;
+ }
+ else
+ {
+ /* Prime the pump. */
+ prev_code = 0;
+ prev_size = source_font->width_table[0];
+ first_code = prev_code;
+ gindex = 1;
+ }
+
+ fz_var(fwobj);
+ fz_var(run_obj);
+
+ fz_try(ctx)
+ {
+ fwobj = pdf_new_array(ctx, doc, 10);
+ while (gindex != 0)
+ {
+ if (source_font->width_table == NULL)
+ {
+ curr_code = FT_Get_Next_Char(fontdesc->font->ft_face, prev_code, &gindex);
+ curr_size = ((FT_Face)fontdesc->font->ft_face)->glyph->metrics.horiAdvance;
+ }
+ else
+ {
+ curr_code = prev_code + 1;
+ if (curr_code == source_font->width_count)
+ {
+ gindex = 0;
+ curr_size = -1;
+ }
+ else
+ curr_size = source_font->width_table[curr_code];
+ }
+
+ /* Check if we need to publish or keep collecting */
+ if (prev_code == curr_code - 1)
+ {
+ /* A concecutive code. */
+ switch (state)
+ {
+ case FW_SAME:
+ if (curr_size != prev_size)
+ {
+ /* End of same widths for consecutive ids. Current will
+ * be pushed as prev. below during next iteration */
+ publish = 1;
+ run_obj = pdf_new_array(ctx, doc, 10);
+ new_state = FW_RUN;
+ /* And the new first code is our current code */
+ new_first_code = curr_code;
+ }
+ break;
+ case FW_RUN:
+ if (curr_size == prev_size)
+ {
+ /* Same width, so start a new same entry starting with
+ * the previous code. i.e. the prev size is not put
+ * in the run */
+ publish = 1;
+ new_state = FW_SAME;
+ new_first_code = prev_code;
+ }
+ else
+ {
+ /* Add prev size to run_obj */
+ pdf_array_push_drop(ctx, run_obj, pdf_new_int(ctx, doc, prev_size));
+ }
+ break;
+ case FW_START:
+ /* Starting fresh. Determine our state */
+ if (curr_size == prev_size)
+ {
+ state = FW_SAME;
+ }
+ else
+ {
+ run_obj = pdf_new_array(ctx, doc, 10);
+ pdf_array_push_drop(ctx, run_obj, pdf_new_int(ctx, doc, prev_size));
+ state = FW_RUN;
+ }
+ new_first_code = prev_code;
+ break;
+ }
+ }
+ else
+ {
+ /* Non conscecutive code. Restart */
+ if (state == FW_RUN)
+ {
+ pdf_array_push_drop(ctx, run_obj, pdf_new_int(ctx, doc, prev_size));
+ }
+ new_state = FW_START;
+ publish = 1;
+ }
+
+ if (publish)
+ {
+ pdf_publish_cid_widths(ctx, doc, fwobj, run_obj, state, first_code, prev_code, prev_size);
+ state = new_state;
+ publish = 0;
+ first_code = new_first_code;
+ }
+ prev_size = curr_size;
+ prev_code = curr_code;
+
+ /* See if we need to flush */
+ if (gindex == 0)
+ pdf_publish_cid_widths(ctx, doc, fwobj, run_obj, state, first_code, prev_code, prev_size);
+ }
+ }
+ fz_catch(ctx)
+ {
+ pdf_drop_obj(ctx, fwobj);
+ pdf_drop_obj(ctx, run_obj);
+ fz_rethrow(ctx);
+ }
+ return fwobj;
+}
+
+/* Descendant font construction used for CID font creation from ttf or Adobe type1 */
+static pdf_obj*
+pdf_add_descendant_font(fz_context *ctx, pdf_document *doc, fz_buffer *buffer, pdf_font_desc *fontdesc, fz_font *source_font)
+{
+ pdf_obj *fobj = NULL;
+ pdf_obj *fref = NULL;
+ pdf_obj *fstr_ref = NULL;
+ pdf_obj *fsys_ref = NULL;
+ pdf_obj *fdes_ref = NULL;
+ pdf_obj *fw = NULL;
+
+ const char *ps_name;
+ FT_Face face = fontdesc->font->ft_face;
+
+ fz_var(fobj);
+ fz_var(fref);
+ fz_var(fstr_ref);
+ fz_var(fsys_ref);
+ fz_var(fw);
+
+ fz_try(ctx)
+ {
+ /* refs */
+ fstr_ref = pdf_font_stream_ref(ctx, doc, buffer);
+ fdes_ref = pdf_font_desc_ref(ctx, doc, fontdesc, fstr_ref);
+ fsys_ref = pdf_add_cid_system_info(ctx, doc);
+
+ /* We may have a cid font already with width info in source font and no
+ * cmap in the ft face */
+ fw = pdf_create_cid_widths(ctx, doc, fontdesc, source_font);
+
+ /* And now the font */
+ fobj = pdf_new_dict(ctx, doc, 3);
+ pdf_dict_put_drop(ctx, fobj, PDF_NAME_Type, PDF_NAME_Font);
+ pdf_dict_put_drop(ctx, fobj, PDF_NAME_BaseFont, pdf_new_name(ctx, doc, fontdesc->font->name));
+ pdf_dict_put_drop(ctx, fobj, PDF_NAME_FontDescriptor, fdes_ref);
+ if (fw != NULL)
+ pdf_dict_put_drop(ctx, fobj, PDF_NAME_W, fw);
+ if (source_font != NULL && source_font->width_table != NULL)
+ pdf_dict_put_drop(ctx, fobj, PDF_NAME_DW, pdf_new_int(ctx, doc, source_font->width_default));
+ if ((ps_name = FT_Get_Postscript_Name(face)) != NULL)
+ pdf_dict_put_drop(ctx, fobj, PDF_NAME_BaseFont, pdf_new_string(ctx, doc, ps_name, strlen(ps_name)));
+ switch (ft_kind(face))
+ {
+ case TYPE1:
+ pdf_dict_put_drop(ctx, fobj, PDF_NAME_Subtype, PDF_NAME_CIDFontType0);
+ break;
+ case TRUETYPE:
+ pdf_dict_put_drop(ctx, fobj, PDF_NAME_Subtype, PDF_NAME_CIDFontType2);
+ break;
+ case UNKNOWN:
+ fz_throw(ctx, FZ_ERROR_GENERIC, "Unknown font subtype");
+ break;
+ }
+ pdf_dict_put_drop(ctx, fobj, PDF_NAME_CIDSystemInfo, fsys_ref);
+ fref = pdf_new_ref(ctx, doc, fobj);
+ }
+ fz_always(ctx)
+ {
+ pdf_drop_obj(ctx, fobj);
+ }
+ fz_catch(ctx)
+ {
+ pdf_drop_obj(ctx, fstr_ref);
+ pdf_drop_obj(ctx, fsys_ref);
+ pdf_drop_obj(ctx, fdes_ref);
+ pdf_drop_obj(ctx, fw);
+ pdf_drop_obj(ctx, fref);
+ fz_rethrow(ctx);
+ }
+ return fref;
+}
+
+/* Create the ToUnicode CMap. */
+static pdf_obj*
+pdf_add_cid_to_unicode(fz_context *ctx, pdf_document *doc, pdf_font_desc *fontdesc)
+{
+ fz_buffer *fzbuf = NULL;
+ pdf_obj *fref = NULL;
+ pdf_obj *fobj = NULL;
+ FT_Face face = fontdesc->font->ft_face;
+ FT_UInt glyph_index;
+ char hex_glyph[7];
+ char hex_unicode[7];
+ char listinfo[19];
+ char entry[22];
+ unsigned short *table, *seq;
+ int k;
+ int has_lock = 0;
+ int count;
+ int temp_count;
+ int pos;
+
+ fz_var(has_lock);
+ fz_var(fzbuf);
+ fz_var(fref);
+ fz_var(fobj);
+
+ fz_try(ctx)
+ {
+ fzbuf = fz_new_buffer(ctx, 0);
+
+ /* Boiler plate */
+ fz_write_buffer(ctx, fzbuf, "/CIDInit /ProcSet findresource begin\n", strlen("/CIDInit /ProcSet findresource begin\n"));
+ fz_write_buffer(ctx, fzbuf, "12 dict begin\n", strlen("12 dict begin\n"));
+ fz_write_buffer(ctx, fzbuf, "begincmap\n", strlen("begincmap\n"));
+ fz_write_buffer(ctx, fzbuf, "/CIDSystemInfo\n", strlen("/CIDSystemInfo\n"));
+ fz_write_buffer(ctx, fzbuf, "<</Registry(Adobe)\n", strlen("<</Registry(Adobe)\n"));
+ fz_write_buffer(ctx, fzbuf, "/Ordering(UCS) /Supplement 0>> def\n", strlen("/Ordering(UCS) /Supplement 0>> def\n"));
+ fz_write_buffer(ctx, fzbuf, "/CMapName /Adobe-Identity-UCS def\n", strlen("/CMapName /Adobe-Identity-UCS def\n"));
+ fz_write_buffer(ctx, fzbuf, "/CMapType 2 def\n", strlen("/CMapType 2 def\n"));
+ fz_write_buffer(ctx, fzbuf, "1 begincodespacerange\n", strlen("1 begincodespacerange\n"));
+ fz_write_buffer(ctx, fzbuf, "<0000> <FFFF>\n", strlen("<0000> <FFFF>\n"));
+ fz_write_buffer(ctx, fzbuf, "endcodespacerange\n", strlen("endcodespacerange\n"));
+
+ /* Sort via populating */
+ table = fz_calloc(ctx, 65536, sizeof(unsigned short));
+ count = 0;
+ for (k = 0; k < 65536; k++)
+ {
+ fz_lock(ctx, FZ_LOCK_FREETYPE);
+ has_lock = 1;
+ glyph_index = FT_Get_Char_Index(face, k);
+ fz_unlock(ctx, FZ_LOCK_FREETYPE);
+ has_lock = 0;
+ if (glyph_index > 0 && glyph_index < 65536)
+ {
+ if (table[glyph_index] == 0)
+ {
+ count++;
+ table[glyph_index] = k;
+ }
+ }
+ }
+
+ /* Now output non-zero entries. */
+ /* Note to have a valid CMap, the number of entries in table set can
+ * not exceed 100, so we have to break into multipe tables. Also, note
+ * that to reduce the file size we should be looking for sequential
+ * ranges. Per Adobe technical note #5411, we can't have a range
+ * cross a boundary where the high order byte changes */
+ /* First the ranges */
+ {
+ seq = fz_calloc(ctx, 65536, sizeof(unsigned short));
+ int num_seq = 0;
+ int k_start = -1;
+ int k_end = -1;
+ int match = 0;
+ char hex_start[7];
+ char hex_end[7];
+ char hex_value_start[7];
+ int j;
+
+ k = 0;
+
+ /* First find the ranges */
+ while (1)
+ {
+ if (k == 65535)
+ break;
+ if (table[k] + 1 == table[k + 1])
+ match = 1;
+ else
+ match = 0;
+
+ /* End any sequences across upper byte boundary changes */
+ if ((k & 0xff00) != ((k + 1) & 0xff00))
+ match = 0;
+
+ /* Start of a sequence */
+ if (k_start == -1 && match)
+ {
+ k_start = k;
+ k_end = k + 1;
+ }
+
+ /* In a sequence */
+ if (k_start != -1 && match)
+ {
+ k_end = k + 1;
+ }
+
+ /* Done with a sequence */
+ if (k_start != -1 && !match)
+ {
+ seq[num_seq * 2] = k_start;
+ seq[num_seq * 2 + 1] = k_end;
+ num_seq = num_seq + 1;
+ k_start = -1;
+ }
+ k = k + 1;
+ }
+
+ /* Now output the ranges, with the 100 max limit enforced */
+ if (num_seq > 0)
+ {
+ pos = 0;
+ while (num_seq > 0)
+ {
+ if (num_seq > 100)
+ temp_count = 100;
+ else
+ temp_count = num_seq;
+ num_seq = num_seq - temp_count;
+ sprintf(&listinfo[0], "%d beginbfrange\n", temp_count);
+ fz_write_buffer(ctx, fzbuf, listinfo, strlen(listinfo));
+ k = 0;
+ while (k < temp_count)
+ {
+ k = k + 1;
+ k_start = seq[pos * 2];
+ k_end = seq[pos * 2 + 1];
+ sprintf(&hex_start[0], "%04x", k_start);
+ sprintf(&hex_end[0], "%04x", k_end);
+ sprintf(&hex_value_start[0], "%04x", table[k_start]);
+ sprintf(&entry[0], "<%s> <%s> <%s>\n", hex_start, hex_end, hex_value_start);
+ fz_write_buffer(ctx, fzbuf, entry, strlen(entry));
+
+ /* Clear out these values from the table so they are not
+ * used in the single entry */
+ count = count - (k_end - k_start + 1);
+ for (j = k_start; j < k_end + 1; j++)
+ table[j] = 0;
+ pos = pos + 1;
+ }
+ fz_write_buffer(ctx, fzbuf, "endbfrange\n", strlen("endbfrange\n"));
+ }
+ }
+ }
+
+ /* The rest of the values need to be output as individuals */
+ pos = 0;
+ while (count > 0)
+ {
+ if (count > 100)
+ temp_count = 100;
+ else
+ temp_count = count;
+ count = count - temp_count;
+ sprintf(&listinfo[0], "%d beginbfchar\n", temp_count);
+ fz_write_buffer(ctx, fzbuf, listinfo, strlen(listinfo));
+ k = 0;
+ while (k < temp_count)
+ {
+ if (table[pos] > 0)
+ {
+ k = k + 1;
+ sprintf(&hex_glyph[0], "%04x", pos);
+ sprintf(&hex_unicode[0], "%04x", table[pos]);
+ sprintf(&entry[0], "<%s> <%s>\n", hex_glyph, hex_unicode);
+ fz_write_buffer(ctx, fzbuf, entry, strlen(entry));
+ }
+ pos = pos + 1;
+ }
+ fz_write_buffer(ctx, fzbuf, "endbfchar\n", strlen("endbfchar\n"));
+ }
+ fz_write_buffer(ctx, fzbuf, "endcmap\n", strlen("endcmap\n"));
+ fz_write_buffer(ctx, fzbuf, "CMapName currentdict /CMap defineresource pop\n", strlen("CMapName currentdict /CMap defineresource pop\n"));
+ fz_write_buffer(ctx, fzbuf, "end\nend\n", strlen("end\nend\n"));
+
+ fobj = pdf_new_dict(ctx, doc, 3);
+ fref = pdf_new_ref(ctx, doc, fobj);
+ pdf_update_stream(ctx, doc, fref, fzbuf, 0);
+ }
+ fz_always(ctx)
+ {
+ fz_free(ctx, table);
+ fz_free(ctx, seq);
+ fz_drop_buffer(ctx, fzbuf);
+ pdf_drop_obj(ctx, fobj);
+ }
+ fz_catch(ctx)
+ {
+ if (has_lock)
+ fz_unlock(ctx, FZ_LOCK_FREETYPE);
+ pdf_drop_obj(ctx, fref);
+ fz_rethrow(ctx);
+ }
+ return fref;
+}
+
+/* Creates CID font with Identity-H CMap and a ToUnicode CMap that is created by
+ * using the TTF cmap table "backwards" to go from the GID to a Unicode value.
+ * If this is coming from a source file, we have source_font so that we can
+ * possibly get any width information that may have been embedded in the PDF
+ * W name tag (or W2 if vertical text) */
+pdf_res*
+pdf_add_cid_font_res(fz_context *ctx, pdf_document *doc, fz_buffer *buffer, fz_font *source_font)
+{
+ pdf_obj *fobj = NULL;
+ pdf_obj *fref = NULL;
+ pdf_res *fres = NULL;
+ pdf_obj *obj_desc_ref = NULL;
+ pdf_obj *obj_tounicode_ref = NULL;
+ pdf_obj *obj_array = NULL;
+ pdf_font_desc *fontdesc = NULL;
+ fz_font *font = NULL;
+ FT_Face face;
+ FT_Error fterr;
+ int has_lock = 0;
+ unsigned char digest[16];
+
+ fz_var(fobj);
+ fz_var(fref);
+ fz_var(fres);
+ fz_var(obj_desc_ref);
+ fz_var(obj_tounicode_ref);
+ fz_var(fontdesc);
+ fz_var(font);
+ fz_var(obj_array);
+ fz_var(has_lock);
+
+ fz_try(ctx)
+ {
+ /* Before we add this font as a resource check if the same font
+ * already exists in our resources for this doc. If yes, then
+ * hand back that reference */
+ fres = pdf_resource_table_search(ctx, doc, doc->resources->font,
+ (void*)buffer, (void*)&(digest[0]));
+ if (fres == NULL)
+ {
+ /* Set up desc, width, and font file */
+ font = fz_new_font_from_memory(ctx, NULL, buffer->data, buffer->len, 0, 1);
+ fontdesc = pdf_new_font_desc(ctx);
+ fontdesc->font = font;
+ face = font->ft_face;
+ fz_lock(ctx, FZ_LOCK_FREETYPE);
+ has_lock = 1;
+ fterr = FT_Set_Char_Size(face, 1000, 1000, 72, 72);
+ if (fterr)
+ fz_warn(ctx, "pdf_add_cid_font_res (FT_Set_Char_Size): %s", ft_error_string(fterr));
+ pdf_fontdesc_init(ctx, fontdesc);
+ fz_unlock(ctx, FZ_LOCK_FREETYPE);
+ has_lock = 0;
+
+ /* Get the descendant font and the tounicode references */
+ obj_desc_ref = pdf_add_descendant_font(ctx, doc, buffer, fontdesc, source_font);
+ obj_tounicode_ref = pdf_add_cid_to_unicode(ctx, doc, fontdesc);
+
+ /* And now the font */
+ fobj = pdf_new_dict(ctx, doc, 3);
+ pdf_dict_put_drop(ctx, fobj, PDF_NAME_Type, PDF_NAME_Font);
+ pdf_dict_put_drop(ctx, fobj, PDF_NAME_Subtype, PDF_NAME_Type0);
+ pdf_dict_put_drop(ctx, fobj, PDF_NAME_BaseFont, pdf_new_name(ctx, doc, fontdesc->font->name));
+ pdf_dict_put_drop(ctx, fobj, PDF_NAME_Encoding, PDF_NAME_Identity_H);
+
+ obj_array = pdf_new_array(ctx, doc, 3);
+ pdf_array_insert_drop(ctx, obj_array, obj_desc_ref, 0);
+ pdf_dict_put_drop(ctx, fobj, PDF_NAME_DescendantFonts, obj_array);
+ pdf_dict_put_drop(ctx, fobj, PDF_NAME_ToUnicode, obj_tounicode_ref);
+ fref = pdf_new_ref(ctx, doc, fobj);
+
+ /* Add ref to our font resource hash table. */
+ fres = pdf_resource_table_put(ctx, doc->resources->font, digest, fref);
+ }
+ }
+ fz_always(ctx)
+ {
+ fz_drop_font(ctx, font);
+ if (fontdesc != NULL)
+ fontdesc->font = NULL;
+ pdf_drop_font(ctx, fontdesc);
+ pdf_drop_obj(ctx, fobj);
+ }
+ fz_catch(ctx)
+ {
+ if (has_lock)
+ fz_unlock(ctx, FZ_LOCK_FREETYPE);
+ pdf_drop_obj(ctx, obj_desc_ref);
+ pdf_drop_obj(ctx, obj_array);
+ pdf_drop_obj(ctx, obj_tounicode_ref);
+ pdf_drop_obj(ctx, fref);
+ fz_rethrow(ctx);
+ }
+ return fres;
+}
+
+/* Creates simple font */
+pdf_res*
+pdf_add_simple_font_res(fz_context *ctx, pdf_document *doc, fz_buffer *buffer)
+{
+ pdf_obj *fobj = NULL;
+ pdf_obj *fref = NULL;
+ pdf_res *fres = NULL;
+ pdf_obj *fstr_ref = NULL;
+ pdf_obj *fdes_ref = NULL;
+ pdf_obj *fwidth_ref = NULL;
+ pdf_font_desc *fontdesc;
+ fz_font *font;
+ FT_Face face;
+ FT_Error fterr;
+ const char *ps_name;
+ int firstcode;
+ int has_lock = 0;
+ unsigned char digest[16];
+
+ fz_var(fobj);
+ fz_var(fref);
+ fz_var(fres);
+ fz_var(fstr_ref);
+ fz_var(fdes_ref);
+ fz_var(fwidth_ref);
+ fz_var(fontdesc);
+ fz_var(font);
+ fz_var(has_lock);
+
+ fz_try(ctx)
+ {
+ /* Before we add this font as a resource check if the same font
+ * already exists in our resources for this doc. If yes, then
+ * hand back that reference */
+ fres = pdf_resource_table_search(ctx, doc, doc->resources->font,
+ (void*)buffer, (void*)&(digest[0]));
+ if (fres == NULL)
+ {
+ /* Set up desc, width, and font file */
+ fobj = pdf_new_dict(ctx, doc, 3);
+ font = fz_new_font_from_memory(ctx, NULL, buffer->data, buffer->len, 0, 1);
+ font->width_count = 256;
+ font->width_table = fz_calloc(ctx, font->width_count, sizeof(int));
+ fontdesc = pdf_new_font_desc(ctx);
+ fontdesc->font = font;
+ face = font->ft_face;
+ fz_lock(ctx, FZ_LOCK_FREETYPE);
+ has_lock = 1;
+ fterr = FT_Set_Char_Size(face, 1000, 1000, 72, 72);
+ if (fterr)
+ fz_warn(ctx, "pdf_add_simple_font_res: %s", ft_error_string(fterr));
+ ft_width_for_simple_table(ctx, fontdesc, &firstcode);
+ pdf_fontdesc_init(ctx, fontdesc);
+ fz_unlock(ctx, FZ_LOCK_FREETYPE);
+ has_lock = 0;
+
+ /* refs */
+ fstr_ref = pdf_font_stream_ref(ctx, doc, buffer);
+ fdes_ref = pdf_font_desc_ref(ctx, doc, fontdesc, fstr_ref);
+ fwidth_ref = pdf_font_widths_ref(ctx, doc, fontdesc);
+
+ /* And now the font */
+ pdf_dict_put_drop(ctx, fobj, PDF_NAME_Type, PDF_NAME_Font);
+ pdf_dict_put_drop(ctx, fobj, PDF_NAME_BaseFont, pdf_new_name(ctx, doc, fontdesc->font->name));
+ pdf_dict_put_drop(ctx, fobj, PDF_NAME_Encoding, PDF_NAME_WinAnsiEncoding);
+ pdf_dict_put_drop(ctx, fobj, PDF_NAME_FirstChar, pdf_new_int(ctx, doc, fontdesc->font->first_width));
+ pdf_dict_put_drop(ctx, fobj, PDF_NAME_LastChar, pdf_new_int(ctx, doc, fontdesc->font->last_width));
+ pdf_dict_put_drop(ctx, fobj, PDF_NAME_Widths, fwidth_ref);
+ pdf_dict_put_drop(ctx, fobj, PDF_NAME_FontDescriptor, fdes_ref);
+ if ((ps_name = FT_Get_Postscript_Name(face)) != NULL)
+ pdf_dict_put_drop(ctx, fobj, PDF_NAME_BaseFont, pdf_new_string(ctx, doc, ps_name, strlen(ps_name)));
+ switch (ft_kind(face))
+ {
+ case TYPE1:
+ pdf_dict_put_drop(ctx, fobj, PDF_NAME_Subtype, PDF_NAME_Type1);
+ break;
+ case TRUETYPE:
+ pdf_dict_put_drop(ctx, fobj, PDF_NAME_Subtype, PDF_NAME_TrueType);
+ break;
+ case UNKNOWN:
+ fz_throw(ctx, FZ_ERROR_GENERIC, "Unknown font subtype");
+ break;
+ }
+ fref = pdf_new_ref(ctx, doc, fobj);
+
+ /* Add ref to our font resource hash table. */
+ fres = pdf_resource_table_put(ctx, doc->resources->font, digest, fref);
+ }
+ }
+ fz_always(ctx)
+ {
+ fz_drop_font(ctx, font);
+ if (fontdesc != NULL)
+ fontdesc->font = NULL;
+ pdf_drop_font(ctx, fontdesc);
+ pdf_drop_obj(ctx, fobj);
+ }
+ fz_catch(ctx)
+ {
+ if (has_lock)
+ fz_unlock(ctx, FZ_LOCK_FREETYPE);
+ pdf_drop_obj(ctx, fstr_ref);
+ pdf_drop_obj(ctx, fdes_ref);
+ pdf_drop_obj(ctx, fwidth_ref);
+ pdf_drop_obj(ctx, fref);
+ fz_rethrow(ctx);
+ }
+ return fres;
+}
+
+int
+pdf_font_writing_supported(fz_font *font)
+{
+ if (font->ft_face == NULL)
+ return 0;
+
+ if (ft_kind(font->ft_face) == TYPE1 || ft_kind(font->ft_face) == TRUETYPE)
+ {
+ return 1;
+ }
+ return 0;
+}
diff --git a/source/pdf/pdf-form.c b/source/pdf/pdf-form.c
index 78adf542..e62c954d 100644
--- a/source/pdf/pdf-form.c
+++ b/source/pdf/pdf-form.c
@@ -23,7 +23,7 @@ static int pdf_field_dirties_document(fz_context *ctx, pdf_document *doc, pdf_ob
return 1;
}
-/* Find the point in a field hierarchy where all descendents
+/* Find the point in a field hierarchy where all descendants
* share the same name */
static pdf_obj *find_head_of_field_group(fz_context *ctx, pdf_obj *obj)
{
diff --git a/source/pdf/pdf-image.c b/source/pdf/pdf-image.c
index b48a0ba5..78828776 100644
--- a/source/pdf/pdf-image.c
+++ b/source/pdf/pdf-image.c
@@ -289,3 +289,178 @@ pdf_load_image(fz_context *ctx, pdf_document *doc, pdf_obj *dict)
pdf_store_item(ctx, dict, image, fz_image_size(ctx, image));
return image;
}
+
+pdf_res*
+pdf_add_image_res(fz_context *ctx, pdf_document *doc, fz_image *image, int mask)
+{
+ fz_pixmap *pixmap = NULL;
+ pdf_obj *imobj = NULL;
+ pdf_obj *imref = NULL;
+ pdf_res *imres = NULL;
+ fz_compressed_buffer *cbuffer = NULL;
+ fz_compression_params *cp = NULL;
+ fz_buffer *buffer = NULL;
+ fz_colorspace *colorspace = image->colorspace;
+ unsigned char digest[16];
+
+ /* If we can maintain compression, do so */
+ cbuffer = image->buffer;
+
+ fz_var(pixmap);
+ fz_var(buffer);
+ fz_var(imobj);
+ fz_var(imref);
+ fz_var(imres);
+
+ fz_try(ctx)
+ {
+ /* Before we add this image as a resource check if the same image
+ * already exists in our resources for this doc. If yes, then
+ * hand back that reference */
+ imres = pdf_resource_table_search(ctx, doc, doc->resources->image,
+ (void*)image, (void*) &(digest[0]));
+ if (imres == NULL)
+ {
+ if (cbuffer != NULL && cbuffer->params.type != FZ_IMAGE_PNG && cbuffer->params.type != FZ_IMAGE_TIFF)
+ {
+ buffer = fz_keep_buffer(ctx, cbuffer->buffer);
+ cp = &cbuffer->params;
+ }
+ else
+ {
+ unsigned int size;
+ int n;
+
+ /* Currently, set to maintain resolution; should we consider
+ * subsampling here according to desired output res? */
+ pixmap = fz_get_pixmap_from_image(ctx, image, image->w, image->h);
+ colorspace = pixmap->colorspace; /* May be different to image->colorspace! */
+ n = (pixmap->n == 1 ? 1 : pixmap->n - 1);
+ size = image->w * image->h * n;
+ buffer = fz_new_buffer(ctx, size);
+ buffer->len = size;
+ if (pixmap->n == 1)
+ {
+ memcpy(buffer->data, pixmap->samples, size);
+ }
+ else
+ {
+ /* Need to remove the alpha plane */
+ unsigned char *d = buffer->data;
+ unsigned char *s = pixmap->samples;
+ int mod = n;
+ while (size--)
+ {
+ *d++ = *s++;
+ mod--;
+ if (mod == 0)
+ s++, mod = n;
+ }
+ }
+ }
+
+ imobj = pdf_new_dict(ctx, doc, 3);
+ pdf_dict_put_drop(ctx, imobj, PDF_NAME_Type, PDF_NAME_XObject);
+ pdf_dict_put_drop(ctx, imobj, PDF_NAME_Subtype, PDF_NAME_Image);
+ pdf_dict_put_drop(ctx, imobj, PDF_NAME_Width, pdf_new_int(ctx, doc, image->w));
+ pdf_dict_put_drop(ctx, imobj, PDF_NAME_Height, pdf_new_int(ctx, doc, image->h));
+ if (mask)
+ {
+ }
+ else if (!colorspace || colorspace->n == 1)
+ pdf_dict_put_drop(ctx, imobj, PDF_NAME_ColorSpace, PDF_NAME_DeviceGray);
+ else if (colorspace->n == 3)
+ pdf_dict_put_drop(ctx, imobj, PDF_NAME_ColorSpace, PDF_NAME_DeviceRGB);
+ else if (colorspace->n == 4)
+ pdf_dict_put_drop(ctx, imobj, PDF_NAME_ColorSpace, PDF_NAME_DeviceCMYK);
+ if (!mask)
+ pdf_dict_put_drop(ctx, imobj, PDF_NAME_BitsPerComponent, pdf_new_int(ctx, doc, image->bpc));
+ switch (cp ? cp->type : FZ_IMAGE_UNKNOWN)
+ {
+ case FZ_IMAGE_UNKNOWN: /* Unknown also means raw */
+ default:
+ break;
+ case FZ_IMAGE_JPEG:
+ if (cp->u.jpeg.color_transform != -1)
+ pdf_dict_put_drop(ctx, imobj, PDF_NAME_ColorTransform, pdf_new_int(ctx, doc, cp->u.jpeg.color_transform));
+ pdf_dict_put_drop(ctx, imobj, PDF_NAME_Filter, PDF_NAME_DCTDecode);
+ break;
+ case FZ_IMAGE_JPX:
+ if (cp->u.jpx.smask_in_data)
+ pdf_dict_put_drop(ctx, imobj, PDF_NAME_SMaskInData, pdf_new_int(ctx, doc, cp->u.jpx.smask_in_data));
+ pdf_dict_put_drop(ctx, imobj, PDF_NAME_Filter, PDF_NAME_JPXDecode);
+ break;
+ case FZ_IMAGE_FAX:
+ if (cp->u.fax.columns)
+ pdf_dict_put_drop(ctx, imobj, PDF_NAME_Columns, pdf_new_int(ctx, doc, cp->u.fax.columns));
+ if (cp->u.fax.rows)
+ pdf_dict_put_drop(ctx, imobj, PDF_NAME_Rows, pdf_new_int(ctx, doc, cp->u.fax.rows));
+ if (cp->u.fax.k)
+ pdf_dict_put_drop(ctx, imobj, PDF_NAME_K, pdf_new_int(ctx, doc, cp->u.fax.k));
+ if (cp->u.fax.end_of_line)
+ pdf_dict_put_drop(ctx, imobj, PDF_NAME_EndOfLine, pdf_new_int(ctx, doc, cp->u.fax.end_of_line));
+ if (cp->u.fax.encoded_byte_align)
+ pdf_dict_put_drop(ctx, imobj, PDF_NAME_EncodedByteAlign, pdf_new_int(ctx, doc, cp->u.fax.encoded_byte_align));
+ if (cp->u.fax.end_of_block)
+ pdf_dict_put_drop(ctx, imobj, PDF_NAME_EndOfBlock, pdf_new_int(ctx, doc, cp->u.fax.end_of_block));
+ if (cp->u.fax.black_is_1)
+ pdf_dict_put_drop(ctx, imobj, PDF_NAME_BlackIs1, pdf_new_int(ctx, doc, cp->u.fax.black_is_1));
+ if (cp->u.fax.damaged_rows_before_error)
+ pdf_dict_put_drop(ctx, imobj, PDF_NAME_DamagedRowsBeforeError, pdf_new_int(ctx, doc, cp->u.fax.damaged_rows_before_error));
+ pdf_dict_put_drop(ctx, imobj, PDF_NAME_Filter, PDF_NAME_CCITTFaxDecode);
+ break;
+ case FZ_IMAGE_JBIG2:
+ /* FIXME - jbig2globals */
+ cp->type = FZ_IMAGE_UNKNOWN;
+ break;
+ case FZ_IMAGE_FLATE:
+ if (cp->u.flate.columns)
+ pdf_dict_put_drop(ctx, imobj, PDF_NAME_Columns, pdf_new_int(ctx, doc, cp->u.flate.columns));
+ if (cp->u.flate.colors)
+ pdf_dict_put_drop(ctx, imobj, PDF_NAME_Colors, pdf_new_int(ctx, doc, cp->u.flate.colors));
+ if (cp->u.flate.predictor)
+ pdf_dict_put_drop(ctx, imobj, PDF_NAME_Predictor, pdf_new_int(ctx, doc, cp->u.flate.predictor));
+ pdf_dict_put_drop(ctx, imobj, PDF_NAME_Filter, PDF_NAME_FlateDecode);
+ pdf_dict_put_drop(ctx, imobj, PDF_NAME_BitsPerComponent, pdf_new_int(ctx, doc, image->bpc));
+ break;
+ case FZ_IMAGE_LZW:
+ if (cp->u.lzw.columns)
+ pdf_dict_put_drop(ctx, imobj, PDF_NAME_Columns, pdf_new_int(ctx, doc, cp->u.lzw.columns));
+ if (cp->u.lzw.colors)
+ pdf_dict_put_drop(ctx, imobj, PDF_NAME_Colors, pdf_new_int(ctx, doc, cp->u.lzw.colors));
+ if (cp->u.lzw.predictor)
+ pdf_dict_put_drop(ctx, imobj, PDF_NAME_Predictor, pdf_new_int(ctx, doc, cp->u.lzw.predictor));
+ if (cp->u.lzw.early_change)
+ pdf_dict_put_drop(ctx, imobj, PDF_NAME_EarlyChange, pdf_new_int(ctx, doc, cp->u.lzw.early_change));
+ pdf_dict_put_drop(ctx, imobj, PDF_NAME_Filter, PDF_NAME_LZWDecode);
+ break;
+ case FZ_IMAGE_RLD:
+ pdf_dict_put_drop(ctx, imobj, PDF_NAME_Filter, PDF_NAME_RunLengthDecode);
+ break;
+ }
+ if (mask)
+ {
+ pdf_dict_put_drop(ctx, imobj, PDF_NAME_ImageMask, pdf_new_bool(ctx, doc, 1));
+ }
+ if (image->mask)
+ pdf_add_image_res(ctx, doc, image->mask, 0);
+ imref = pdf_new_ref(ctx, doc, imobj);
+ pdf_update_stream(ctx, doc, imref, buffer, 1);
+
+ /* Add ref to our image resource hash table. */
+ imres = pdf_resource_table_put(ctx, doc->resources->image, digest, imref);
+ }
+ }
+ fz_always(ctx)
+ {
+ fz_drop_buffer(ctx, buffer);
+ pdf_drop_obj(ctx, imobj);
+ fz_drop_pixmap(ctx, pixmap);
+ }
+ fz_catch(ctx)
+ {
+ pdf_drop_obj(ctx, imref);
+ fz_rethrow(ctx);
+ }
+ return imres;
+}
diff --git a/source/pdf/pdf-page.c b/source/pdf/pdf-page.c
index d893d204..30cbb4e9 100644
--- a/source/pdf/pdf-page.c
+++ b/source/pdf/pdf-page.c
@@ -697,15 +697,17 @@ pdf_delete_page_range(fz_context *ctx, pdf_document *doc, int start, int end)
}
pdf_page *
-pdf_create_page(fz_context *ctx, pdf_document *doc, fz_rect mediabox, int res, int rotate)
+pdf_create_page(fz_context *ctx, pdf_document *doc, fz_rect mediabox, fz_buffer *buffer, int rotate)
{
pdf_page *page = NULL;
- pdf_obj *pageobj;
+ pdf_obj *pageobj, *obj;
float userunit = 1;
fz_matrix ctm, tmp;
fz_rect realbox;
page = pdf_new_page(ctx, doc);
+ obj = NULL;
+ fz_var(obj);
fz_try(ctx)
{
@@ -724,7 +726,7 @@ pdf_create_page(fz_context *ctx, pdf_document *doc, fz_rect mediabox, int res, i
page->rotate = 360 - ((-page->rotate) % 360);
if (page->rotate >= 360)
page->rotate = page->rotate % 360;
- page->rotate = 90*((page->rotate + 45)/90);
+ page->rotate = 90 * ((page->rotate + 45) / 90);
if (page->rotate > 360)
page->rotate = 0;
pdf_dict_put_drop(ctx, pageobj, PDF_NAME_Rotate, pdf_new_int(ctx, doc, page->rotate));
@@ -735,12 +737,21 @@ pdf_create_page(fz_context *ctx, pdf_document *doc, fz_rect mediabox, int res, i
fz_pre_scale(fz_translate(&tmp, -realbox.x0, -realbox.y0), userunit, userunit);
fz_concat(&ctm, &ctm, &tmp);
page->ctm = ctm;
- /* Do not create a Contents, as an empty Contents dict is not
- * valid. See Bug 694712 */
+
+ if (buffer != NULL)
+ {
+ obj = pdf_new_dict(ctx, doc, 4);
+ page->contents = pdf_new_ref(ctx, doc, obj);
+ pdf_update_stream(ctx, doc, page->contents, buffer, 0);
+ pdf_drop_obj(ctx, obj);
+ obj = NULL;
+ pdf_dict_puts(ctx, pageobj, "Contents", page->contents);
+ }
}
fz_catch(ctx)
{
pdf_drop_obj(ctx, page->me);
+ pdf_drop_obj(ctx, obj);
fz_free(ctx, page);
fz_rethrow_message(ctx, "Failed to create page");
}
diff --git a/source/pdf/pdf-resources.c b/source/pdf/pdf-resources.c
new file mode 100644
index 00000000..31812b3c
--- /dev/null
+++ b/source/pdf/pdf-resources.c
@@ -0,0 +1,295 @@
+#include "mupdf/pdf.h"
+
+static void
+res_table_free(fz_context *ctx, pdf_res_table *table)
+{
+ int i, n;
+ pdf_res *res;
+
+ if (table == NULL)
+ return;
+ if (table->hash != NULL)
+ {
+ n = fz_hash_len(ctx, table->hash);
+ for (i = 0; i < n; i++)
+ {
+ void *v = fz_hash_get_val(ctx, table->hash, i);
+ if (v)
+ {
+ res = (pdf_res*)v;
+ pdf_drop_obj(ctx, res->obj);
+ fz_free(ctx, res);
+ }
+ }
+ fz_drop_hash(ctx, table->hash);
+ }
+ fz_free(ctx, table);
+}
+
+static void
+res_image_get_md5(fz_context *ctx, fz_image *image, unsigned char *digest)
+{
+ fz_pixmap *pixmap = NULL;
+ int n, size;
+ fz_buffer *buffer = NULL;
+ fz_md5 state;
+
+ fz_var(pixmap);
+ fz_var(buffer);
+
+ fz_try(ctx)
+ {
+ pixmap = fz_get_pixmap_from_image(ctx, image, 0, 0);
+ n = (pixmap->n == 1 ? 1 : pixmap->n - 1);
+ size = image->w * image->h * n;
+ buffer = fz_new_buffer(ctx, size);
+ buffer->len = size;
+ if (pixmap->n == 1)
+ {
+ memcpy(buffer->data, pixmap->samples, size);
+ }
+ else
+ {
+ /* Need to remove the alpha plane */
+ unsigned char *d = buffer->data;
+ unsigned char *s = pixmap->samples;
+ int mod = n;
+ while (size--)
+ {
+ *d++ = *s++;
+ mod--;
+ if (mod == 0)
+ s++, mod = n;
+ }
+ }
+ fz_md5_init(&state);
+ fz_md5_update(&state, buffer->data, buffer->len);
+ fz_md5_final(&state, digest);
+ }
+ fz_always(ctx)
+ {
+ fz_drop_pixmap(ctx, pixmap);
+ fz_drop_buffer(ctx, buffer);
+ }
+ fz_catch(ctx)
+ {
+ fz_rethrow_message(ctx, "image md5 calculation failed");
+ }
+}
+
+/* Image specific methods */
+static void
+res_image_init(fz_context *ctx, pdf_document *doc, pdf_res_table *table)
+{
+ int len, k;
+ pdf_obj *obj;
+ pdf_obj *type;
+ pdf_res *res = NULL;
+ fz_image *image = NULL;
+ unsigned char digest[16];
+ int num = 0;
+
+ fz_var(obj);
+ fz_var(image);
+ fz_var(res);
+
+ fz_try(ctx)
+ {
+ table->hash = fz_new_hash_table(ctx, 4096, 16, -1);
+ len = pdf_count_objects(ctx, doc);
+ for (k = 1; k < len; k++)
+ {
+ obj = pdf_load_object(ctx, doc, k, 0);
+ type = pdf_dict_get(ctx, obj, PDF_NAME_Subtype);
+ if (pdf_name_eq(ctx, type, PDF_NAME_Image))
+ {
+ image = pdf_load_image(ctx, doc, obj);
+ res_image_get_md5(ctx, image, digest);
+ fz_drop_image(ctx, image);
+ image = NULL;
+
+ /* Don't allow overwrites. Number the resources for pdfwrite */
+ if (fz_hash_find(ctx, table->hash, (void *)digest) == NULL)
+ {
+ res = fz_malloc(ctx, sizeof(pdf_res));
+ res->num = num;
+ res->obj = obj;
+ num = num + 1;
+ fz_hash_insert(ctx, table->hash, (void *)digest, obj);
+ }
+ }
+ else
+ {
+ pdf_drop_obj(ctx, obj);
+ }
+ obj = NULL;
+ }
+ }
+ fz_always(ctx)
+ {
+ table->count = num;
+ fz_drop_image(ctx, image);
+ pdf_drop_obj(ctx, obj);
+ }
+ fz_catch(ctx)
+ {
+ res_table_free(ctx, table);
+ fz_rethrow_message(ctx, "image resources table failed to initialize");
+ }
+}
+
+static void*
+res_image_search(fz_context *ctx, pdf_document *doc, pdf_res_table *table, void *item,
+ void *md5)
+{
+ unsigned char digest[16];
+
+ fz_image *image = (fz_image*)item;
+ fz_hash_table *hash = table->hash;
+ pdf_res *res;
+
+ if (hash == NULL)
+ res_image_init(ctx, doc, doc->resources->image);
+ hash = doc->resources->image->hash;
+
+ /* Create md5 and see if we have the item in our table */
+ res_image_get_md5(ctx, image, digest);
+ res = fz_hash_find(ctx, hash, (void*)digest);
+
+ /* Return the digest value so that we can avoid having to recompute it when
+ * we come back to add the new resource reference */
+ if (res == NULL)
+ memcpy(md5, digest, 16);
+ else
+ pdf_keep_obj(ctx, res->obj);
+ return (void*) res;
+}
+
+/* Font specific methods */
+
+/* We do need to come up with an effective way to see what is already in the
+ * file to avoid adding to what is already there. This is avoided for pdfwrite
+ * as we check as we add each font. For adding text to an existing file though
+ * it may be more problematic */
+static void
+res_font_init(fz_context *ctx, pdf_document *doc, pdf_res_table *table)
+{
+ table->hash = fz_new_hash_table(ctx, 4096, 16, -1);
+}
+
+static void
+res_font_get_md5(fz_context *ctx, fz_buffer *buffer, unsigned char *digest)
+{
+ fz_md5 state;
+
+ fz_md5_init(&state);
+ fz_md5_update(&state, buffer->data, buffer->len);
+ fz_md5_final(&state, digest);
+}
+
+static void*
+res_font_search(fz_context *ctx, pdf_document *doc, pdf_res_table *table, void *item,
+ void *md5)
+{
+ unsigned char digest[16];
+ fz_buffer *buffer = (fz_buffer*)item;
+ fz_hash_table *hash = table->hash;
+ pdf_res *res;
+
+ if (hash == NULL)
+ res_font_init(ctx, doc, doc->resources->font);
+ hash = doc->resources->font->hash;
+
+ /* Create md5 and see if we have the item in our table */
+ res_font_get_md5(ctx, buffer, digest);
+ res = fz_hash_find(ctx, hash, (void*)digest);
+
+ /* Return the digest value so that we can avoid having to recompute it when
+ * we come back to add the new resource reference */
+ if (res == NULL)
+ memcpy(md5, digest, 16);
+ else
+ pdf_keep_obj(ctx, res->obj);
+ return (void*)res;
+}
+
+/* Accessible methods */
+void*
+pdf_resource_table_search(fz_context *ctx, pdf_document *doc, pdf_res_table *table,
+ void *item, void *md5)
+{
+ return table->search(ctx, doc, table, item, md5);
+}
+
+void*
+pdf_resource_table_put(fz_context *ctx, pdf_res_table *table, void *key, pdf_obj *obj)
+{
+ void *result;
+ pdf_res *res = NULL;
+
+ fz_var(res);
+
+ fz_try(ctx)
+ {
+ res = fz_malloc(ctx, sizeof(pdf_res));
+ res->num = table->count + 1;
+ res->obj = obj;
+ result = fz_hash_insert(ctx, table->hash, key, (void*)res);
+ if (result != NULL)
+ {
+ fz_free(ctx, res);
+ fz_warn(ctx, "warning: hash already present");
+ }
+ else
+ {
+ table->count = table->count + 1;
+ pdf_keep_obj(ctx, obj);
+ result = res;
+ }
+ }
+ fz_catch(ctx)
+ {
+ fz_free(ctx, res);
+ fz_rethrow(ctx);
+ }
+ return result;
+}
+
+void
+pdf_resource_table_free(fz_context *ctx, pdf_document *doc)
+{
+ if (doc->resources == NULL)
+ return;
+ res_table_free(ctx, doc->resources->color);
+ res_table_free(ctx, doc->resources->font);
+ res_table_free(ctx, doc->resources->image);
+ res_table_free(ctx, doc->resources->pattern);
+ res_table_free(ctx, doc->resources->shading);
+ fz_free(ctx, doc->resources);
+ doc->resources = NULL;
+}
+
+void
+pdf_resource_table_init(fz_context *ctx, pdf_document *doc)
+{
+ fz_var(doc);
+ fz_try(ctx)
+ {
+ doc->resources = fz_calloc(ctx, 1, sizeof(pdf_resource_tables));
+ doc->resources->image = fz_calloc(ctx, 1, sizeof(pdf_res_table));
+ doc->resources->image->search = res_image_search;
+ doc->resources->font = fz_calloc(ctx, 1, sizeof(pdf_res_table));
+ doc->resources->font->search = res_font_search;
+ }
+ fz_catch(ctx)
+ {
+ if (doc->resources != NULL)
+ {
+ fz_free(ctx, doc->resources->color);
+ fz_free(ctx, doc->resources->font);
+ fz_free(ctx, doc->resources);
+ doc->resources = NULL;
+ }
+ fz_rethrow_message(ctx, "resources failed to allocate");
+ }
+}
diff --git a/source/pdf/pdf-xref.c b/source/pdf/pdf-xref.c
index 3bcf5d79..32b1d6a2 100644
--- a/source/pdf/pdf-xref.c
+++ b/source/pdf/pdf-xref.c
@@ -1626,6 +1626,8 @@ pdf_close_document(fz_context *ctx, pdf_document *doc)
pdf_lexbuf_fin(ctx, &doc->lexbuf.base);
+ pdf_resource_table_free(ctx, doc);
+
fz_free(ctx, doc);
}
@@ -2744,6 +2746,7 @@ pdf_document *pdf_create_document(fz_context *ctx)
doc->num_incremental_sections = 0;
doc->xref_base = 0;
doc->disallow_new_increments = 0;
+ pdf_resource_table_init(ctx, doc);
pdf_get_populating_xref_entry(ctx, doc, 0);
trailer = pdf_new_dict(ctx, doc, 2);
pdf_dict_put_drop(ctx, trailer, PDF_NAME_Size, pdf_new_int(ctx, doc, 3));
diff --git a/source/tools/mudraw.c b/source/tools/mudraw.c
index 47aab044..d63f936c 100644
--- a/source/tools/mudraw.c
+++ b/source/tools/mudraw.c
@@ -387,7 +387,7 @@ static void drawpage(fz_context *ctx, fz_document *doc, int pagenum)
tbounds = bounds;
fz_transform_rect(&tbounds, &ctm);
- newpage = pdf_create_page(ctx, pdfout, bounds, 72, 0);
+ newpage = pdf_create_page(ctx, pdfout, bounds, NULL, 0);
fz_try(ctx)
{
diff --git a/source/tools/mutool.c b/source/tools/mutool.c
index 6b61dab4..c662ffef 100644
--- a/source/tools/mutool.c
+++ b/source/tools/mutool.c
@@ -15,6 +15,7 @@ int pdfinfo_main(int argc, char *argv[]);
int pdfposter_main(int argc, char *argv[]);
int pdfshow_main(int argc, char *argv[]);
int pdfpages_main(int argc, char *argv[]);
+int pdfcreate_main(int argc, char *argv[]);
static struct {
int (*func)(int argc, char *argv[]);
@@ -28,6 +29,7 @@ static struct {
{ pdfpages_main, "pages", "show information about pdf pages" },
{ pdfposter_main, "poster", "split large page into many tiles" },
{ pdfshow_main, "show", "show internal pdf objects" },
+ { pdfcreate_main, "create", "create pdf document" },
};
static int
diff --git a/source/tools/pdfcreate.c b/source/tools/pdfcreate.c
new file mode 100644
index 00000000..4e4a622e
--- /dev/null
+++ b/source/tools/pdfcreate.c
@@ -0,0 +1,516 @@
+/*
+ * PDF creation tool: Tool for creating pdf content.
+ *
+ * Simple test bed to work with adding content and creating PDFs
+ *
+ */
+
+#include "mupdf/pdf.h"
+
+#define MAX_IMAGES 10
+#define MAX_FONTS 10
+#define MAX_REF_NAME 32
+
+enum { RES_FONT, RES_XOBJECT };
+
+static void usage(void)
+{
+ fprintf(stderr,
+ "usage: mutool create [-o output.pdf] [fonts][images] contents\n"
+ "\t-f\tfont label:font file\n"
+ "\t-i\timage label:image file\n"
+ "\tcontents file, defines page size, graphics, references fonts and images\n"
+ );
+ exit(1);
+}
+
+/* Simple structures to hold and manage contents */
+typedef struct doc_content_s doc_content;
+typedef struct resources_s resources;
+typedef struct page_resource_xref_s page_resource_xref;
+
+struct page_resource_xref_s
+{
+ char *res_name;
+ char res_ref[MAX_REF_NAME];
+ pdf_obj *obj;
+ int type;
+};
+
+struct resources_s
+{
+ char *name;
+ int ref;
+};
+
+struct doc_content_s
+{
+ fz_stream *stm;
+ int num_pages;
+ fz_point *page_sizes;
+ fz_off_t *content_offsets;
+ int *content_lengths;
+ resources *fonts;
+ int num_fonts;
+ resources *images;
+ int num_images;
+ page_resource_xref **ref_im_resources;
+ page_resource_xref **ref_font_resources;
+ int *num_page_im_res;
+ int *num_page_font_res;
+};
+
+/* Look for the presence of image or font refererences in the content */
+static void check_for_reference(fz_context *ctx, resources *resource, int num_res, char *buffer, int res_type, doc_content *doc, int page_num)
+{
+ int i;
+ char * pch;
+ int length;
+ char name[MAX_REF_NAME];
+ int res_num;
+
+ for (i = 0; i < num_res; i++)
+ {
+ pch = strchr(resource[i].name, ':');
+ if (pch != NULL)
+ {
+ length = pch - resource[i].name;
+ if (length < sizeof(name))
+ {
+ memcpy(name, resource[i].name, length);
+ name[length] = 0;
+ pch = strstr(buffer, name);
+ if (pch != NULL)
+ {
+ /* Resource reference is in the content for this document. Mark
+ * it as such so that we know to add it to the document */
+ resource[i].ref = 1;
+
+ /* Also note this page has a reference to this name so
+ * we can add it to the page resource list. Select from
+ * font or image */
+ if (res_type == RES_FONT)
+ {
+ res_num = doc->num_page_font_res[page_num];
+ doc->ref_font_resources[page_num][res_num].obj = NULL; /* Set later */
+ doc->ref_font_resources[page_num][res_num].res_name = resource[i].name;
+ doc->ref_font_resources[page_num][res_num].type = res_type;
+ memcpy(doc->ref_font_resources[page_num][res_num].res_ref, name, sizeof(name));
+ doc->num_page_font_res[page_num] += 1;
+ }
+ else
+ {
+ res_num = doc->num_page_im_res[page_num];
+ doc->ref_im_resources[page_num][res_num].obj = NULL; /* Set later */
+ doc->ref_im_resources[page_num][res_num].res_name = resource[i].name;
+ doc->ref_im_resources[page_num][res_num].type = res_type;
+ memcpy(doc->ref_im_resources[page_num][res_num].res_ref, name, sizeof(name));
+ doc->num_page_im_res[page_num] += 1;
+ }
+ }
+ }
+ else
+ fz_throw(ctx, FZ_ERROR_GENERIC, "Image/Font indirect name too long");
+ }
+ else
+ fz_throw(ctx, FZ_ERROR_GENERIC, "Internal parsing error");
+ }
+}
+
+static void drop_ref_objs(fz_context *ctx, page_resource_xref *refs, int num_refs)
+{
+ int i;
+
+ for (i = 0; i < num_refs; i++)
+ pdf_drop_obj(ctx, refs[i].obj);
+}
+
+static void free_contents(fz_context *ctx, doc_content *content)
+{
+ int i;
+
+ fz_free(ctx, content->content_offsets);
+ fz_free(ctx, content->content_lengths);
+ fz_free(ctx, content->page_sizes);
+ for (i = 0; i < content->num_pages; i++)
+ {
+ drop_ref_objs(ctx, content->ref_font_resources[i], content->num_page_font_res[i]);
+ fz_free(ctx, content->ref_font_resources[i]);
+ }
+ fz_free(ctx, content->ref_font_resources);
+ for (i = 0; i < content->num_pages; i++)
+ {
+ drop_ref_objs(ctx, content->ref_im_resources[i], content->num_page_im_res[i]);
+ fz_free(ctx, content->ref_im_resources[i]);
+ }
+ fz_free(ctx, content->num_page_im_res);
+ fz_free(ctx, content->num_page_font_res);
+ fz_free(ctx, content->ref_im_resources);
+ content->content_offsets = NULL;
+ content->page_sizes = NULL;
+ content->content_lengths = NULL;
+ fz_drop_stream(ctx, content->stm);
+}
+
+/* This is a VERY simple format to give us something to play with
+ * in terms of defining pages and content. Here we parse the contents
+ * defining our page sizes and the command locations for each page.
+ * The format is as follows:
+ * 1) Comment lines are preceded by %
+ * 2) Number of pages is indicated at the begining with /Pages #
+ * 3) Each page is indicated by /Page # [X Y] (# is zero based)
+ * 4) The content is in the form of simple PDF content stream that
+ * may included various drawing commands and reference the
+ * image and font resources.
+ * */
+static void init_parse_contents(fz_context *ctx, char *content_fn, doc_content *content)
+{
+ fz_stream *stm;
+ char buf[1024];
+ int page_count = 0;
+ fz_off_t pre_off;
+ int i;
+
+ fz_var(stm);
+
+ fz_try(ctx)
+ {
+ stm = fz_open_file(ctx, content_fn);
+ while (1)
+ {
+ pre_off = fz_tell(ctx, stm);
+ fz_read_line(ctx, stm, buf, sizeof buf);
+ if (buf[0] == '\0')
+ {
+ if (content->num_pages > (page_count + 1))
+ fz_throw(ctx, FZ_ERROR_GENERIC, "Missing defined pages");
+ else
+ {
+ content->content_lengths[page_count - 1] =
+ fz_tell(ctx, stm) - content->content_offsets[page_count - 1];
+ }
+ break;
+ }
+ if (buf[0] != '%')
+ {
+ if (strncmp(buf, "/Pages", strlen("/Pages")) == 0)
+ {
+ if ((content->num_pages = atoi(&(buf[strlen("/Pages")]))) <= 0)
+ fz_throw(ctx, FZ_ERROR_GENERIC, "Page count invalid");
+ content->content_offsets = fz_malloc_array(ctx, content->num_pages, sizeof(fz_off_t));
+ content->content_lengths = fz_malloc_array(ctx, content->num_pages, sizeof(int));
+ content->page_sizes = fz_malloc_array(ctx, content->num_pages, sizeof(fz_point));
+ content->num_page_im_res = fz_calloc(ctx, content->num_pages, sizeof(int));
+ content->num_page_font_res = fz_calloc(ctx, content->num_pages, sizeof(int));
+ content->ref_font_resources = fz_malloc_array(ctx, content->num_pages, sizeof(page_resource_xref*));
+ content->ref_im_resources = fz_malloc_array(ctx, content->num_pages, sizeof(page_resource_xref*));
+ for (i = 0; i < content->num_pages; i++)
+ {
+ content->ref_font_resources[i] = fz_malloc_array(ctx, MAX_FONTS, sizeof(page_resource_xref));
+ content->ref_im_resources[i] = fz_malloc_array(ctx, MAX_FONTS, sizeof(page_resource_xref));
+ }
+ }
+ else if (strncmp(buf, "/Page", strlen("/Page")) == 0)
+ {
+ int page_num;
+ if (page_count > 0)
+ {
+ content->content_lengths[page_count - 1] =
+ (int)(pre_off - content->content_offsets[page_count - 1]);
+ }
+ if (sscanf(&(buf[strlen("/Page")]), "%d [%f %f]", &page_num,
+ &(content->page_sizes[page_count].x), &(content->page_sizes[page_count].y)) != 3)
+ fz_throw(ctx, FZ_ERROR_GENERIC, "Page size invalid");
+ if (page_num < 0 || page_num >= content->num_pages)
+ fz_throw(ctx, FZ_ERROR_GENERIC, "Page value invalid");
+ if (content->page_sizes[page_count].x < 0 || content->page_sizes[page_count].y < 0)
+ fz_throw(ctx, FZ_ERROR_GENERIC, "Page dimensions invalid");
+ content->content_offsets[page_count] = fz_tell(ctx, stm);
+ page_count++;
+ }
+ else
+ {
+ check_for_reference(ctx, content->images, content->num_images, buf, RES_XOBJECT, content, page_count - 1);
+ check_for_reference(ctx, content->fonts, content->num_fonts, buf, RES_FONT, content, page_count - 1);
+ }
+ }
+ }
+ }
+ fz_catch(ctx)
+ {
+ free_contents(ctx, content);
+ fz_drop_stream(ctx, stm);
+ fz_rethrow(ctx);
+ }
+ content->stm = stm;
+}
+
+/* Get the page contents */
+static int get_page_contents(fz_context *ctx, int page_num, doc_content *content,
+ unsigned char *buffer)
+{
+ int size = content->content_lengths[page_num];
+
+ if (buffer == NULL)
+ return size;
+ fz_seek(ctx, content->stm, content->content_offsets[page_num], SEEK_SET);
+ fz_read(ctx, content->stm, buffer, size);
+ return size;
+}
+
+static void update_res(fz_context *ctx, int num_pages, int *num_res, page_resource_xref **page_resource,
+ char *res_name, pdf_obj *obj)
+{
+ int j, i;
+
+ for (j = 0; j < num_pages; j++)
+ for (i = 0; i < num_res[j]; i++)
+ if (strcmp(res_name, page_resource[j][i].res_name) == 0)
+ page_resource[j][i].obj = obj;
+}
+
+static pdf_obj* create_page_res_dict(fz_context *ctx, pdf_document *pdf,
+ page_resource_xref *ref_res, const char type[], int num_items)
+{
+ pdf_obj *dict = NULL;
+ int i;
+
+ if (num_items <= 0)
+ return NULL;
+
+ fz_var(dict);
+
+ fz_try(ctx)
+ {
+ dict = pdf_new_dict(ctx, pdf, num_items);
+ for (i = 0; i < num_items; i++)
+ {
+ char text[32];
+ snprintf(text, sizeof(text), "%s/%s", type, ref_res[i].res_ref);
+ pdf_dict_putp(ctx, dict, text, ref_res[i].obj);
+ }
+ }
+ fz_catch(ctx)
+ {
+ pdf_drop_obj(ctx, dict);
+ fz_rethrow(ctx);
+ }
+ return dict;
+}
+
+static int create_pdf(fz_context *ctx, char *output, resources fonts[], int num_fonts,
+ resources images[], int num_images, char *contents)
+{
+ fz_rect bounds;
+ pdf_document *pdf = NULL;
+ pdf_page *newpage = NULL;
+ unsigned char *buffer = NULL;
+ fz_buffer *fz_buf = NULL;
+ fz_buffer *im_font_buff = NULL;
+ fz_image *image = NULL;
+ pdf_obj *font_dict = NULL;
+ pdf_obj *im_dict = NULL;
+ pdf_res *im_res = NULL;
+ pdf_res *font_res = NULL;
+ pdf_write_options opts = { 0 };
+ doc_content content = { 0 };
+ int k;
+ int length;
+ char *pch;
+
+ fz_var(pdf);
+ fz_var(newpage);
+ fz_var(buffer);
+ fz_var(fz_buf);
+ fz_var(im_font_buff);
+ fz_var(image);
+ fz_var(font_dict);
+ fz_var(im_dict);
+ fz_var(im_res);
+ fz_var(font_res);
+
+ fz_try(ctx)
+ {
+ pdf = pdf_create_document(ctx);
+ content.num_fonts = num_fonts;
+ content.num_images = num_images;
+ content.fonts = fonts;
+ content.images = images;
+ init_parse_contents(ctx, contents, &content);
+
+ /* Add the resources, getting the reference numbers in the process. */
+ for (k = 0; k < content.num_images; k++)
+ {
+ if (content.images[k].ref)
+ {
+ /* Get the fz_image */
+ pch = strchr(content.images[k].name, ':');
+ if (pch != NULL)
+ {
+ im_font_buff = fz_read_file(ctx, &(pch[1]));
+ image = fz_new_image_from_buffer(ctx, im_font_buff);
+ fz_drop_buffer(ctx, im_font_buff);
+ im_font_buff = NULL;
+ im_res = pdf_add_image_res(ctx, pdf, image, 0);
+ fz_drop_image(ctx, image);
+ image = NULL;
+
+ /* Look through our image page resources and update the
+ * indirect reference number. Here we don't use the numbers
+ * set by the doc resource holder (i.e im_res->num) since we
+ * are using our own content specified for pdfcreate */
+ update_res(ctx, content.num_pages, content.num_page_im_res,
+ content.ref_im_resources, content.images[k].name, im_res->obj);
+ }
+ else
+ fz_throw(ctx, FZ_ERROR_GENERIC, "Image indirect name too long");
+ }
+ }
+ for (k = 0; k < content.num_fonts; k++)
+ {
+ if (content.fonts[k].ref)
+ {
+ pch = strchr(content.fonts[k].name, ':');
+ if (pch != NULL)
+ {
+ im_font_buff = fz_read_file(ctx, &(pch[1]));
+ font_res = pdf_add_simple_font_res(ctx, pdf, im_font_buff);
+ fz_drop_buffer(ctx, im_font_buff);
+ im_font_buff = NULL;
+
+ /* Look through our font page resources and update the indirect
+ * reference number */
+ update_res(ctx, content.num_pages, content.num_page_font_res,
+ content.ref_font_resources, content.fonts[k].name, font_res->obj);
+ }
+ else
+ fz_throw(ctx, FZ_ERROR_GENERIC, "Font indirect name too long");
+ }
+ }
+
+ /* Now the page contents */
+ for (k = 0; k < content.num_pages; k++)
+ {
+ bounds.x0 = 0;
+ bounds.y0 = 0;
+ bounds.x1 = content.page_sizes[k].x;
+ bounds.y1 = content.page_sizes[k].y;
+
+ length = get_page_contents(ctx, k, &content, NULL);
+ if (length > 0)
+ {
+ buffer = fz_malloc(ctx, length);
+ length = get_page_contents(ctx, k, &content, buffer);
+ fz_buf = fz_new_buffer_from_data(ctx, buffer, length);
+ buffer = NULL;
+ newpage = pdf_create_page(ctx, pdf, bounds, fz_buf, 0);
+ /* Create the dicts for the page resources */
+ font_dict = create_page_res_dict(ctx, pdf, content.ref_font_resources[k],
+ "Font", content.num_page_font_res[k]);
+ im_dict = create_page_res_dict(ctx, pdf, content.ref_im_resources[k],
+ "XObject", content.num_page_im_res[k]);
+ if (im_dict != NULL)
+ {
+ pdf_dict_puts(ctx, newpage->me, "Resources", im_dict);
+ pdf_drop_obj(ctx, im_dict);
+ }
+ if (font_dict != NULL)
+ {
+ pdf_dict_puts(ctx, newpage->me, "Resources", font_dict);
+ pdf_drop_obj(ctx, font_dict);
+ }
+ fz_drop_buffer(ctx, fz_buf);
+ fz_buf = NULL;
+ }
+ else
+ {
+ newpage = pdf_create_page(ctx, pdf, bounds, NULL, 0);
+ }
+ pdf_insert_page(ctx, pdf, newpage, INT_MAX);
+ pdf_drop_page(ctx, newpage);
+ newpage = NULL;
+ }
+ pdf_save_document(ctx, pdf, output, &opts);
+ }
+ fz_always(ctx)
+ {
+ pdf_drop_page(ctx, newpage);
+ pdf_close_document(ctx, pdf);
+ free_contents(ctx, &content);
+ fz_free(ctx, buffer);
+ fz_drop_buffer(ctx, fz_buf);
+ fz_drop_buffer(ctx, im_font_buff);
+ fz_drop_image(ctx, image);
+ }
+ fz_catch(ctx)
+ {
+ fz_rethrow(ctx);
+ }
+ return 0;
+}
+
+int pdfcreate_main(int argc, char **argv)
+{
+ char *outfile = "out.pdf";
+ resources fonts[MAX_FONTS];
+ resources images[MAX_IMAGES];
+ char *contents = "";
+ int nfonts = 0;
+ int nimages = 0;
+ int c;
+ int errors = 0;
+ fz_context *ctx;
+
+ while ((c = fz_getopt(argc, argv, "f:i:o:")) != -1)
+ {
+ switch (c)
+ {
+ case 'f':
+ if (nfonts == MAX_FONTS)
+ {
+ fprintf(stderr, "max number of fonts exceeded\n");
+ exit(1);
+ }
+ fonts[nfonts].name = fz_optarg;
+ fonts[nfonts].ref = 0;
+ nfonts++;
+ break;
+ case 'i':
+ if (nimages == MAX_IMAGES)
+ {
+ fprintf(stderr, "max number of images exceeded\n");
+ exit(1);
+ }
+ images[nimages].name = fz_optarg;
+ images[nimages].ref = 0;
+ nimages++;
+ break;
+ case 'o': outfile = fz_optarg; break;
+ default: usage(); break;
+ }
+ }
+
+ if (fz_optind == argc)
+ usage();
+
+ contents = argv[fz_optind++];
+
+ ctx = fz_new_context(NULL, NULL, FZ_STORE_UNLIMITED);
+ if (!ctx)
+ {
+ fprintf(stderr, "cannot initialise context\n");
+ exit(1);
+ }
+
+ fz_try(ctx)
+ {
+ create_pdf(ctx, outfile, fonts, nfonts, images, nimages, contents);
+ }
+ fz_catch(ctx)
+ {
+ errors++;
+ }
+ fz_drop_context(ctx);
+
+ return errors != 0;
+}