static int
pdf_patternusesblending(fz_obj *dict)
{
fz_obj *obj;
obj = fz_dictgets(dict, "Resources");
if (fz_isdict(obj) && pdf_resourcesuseblending(obj))
return 1;
obj = fz_dictgets(dict, "ExtGState");
if (fz_isdict(obj) && pdf_extgstateusesblending(obj))
return 1;
return 0;
}
static fz_obj *
resolvedest(pdf_xref *xref, fz_obj *dest)
{
if (fz_isname(dest) || fz_isstring(dest))
{
dest = pdf_lookupdest(xref, dest);
return resolvedest(xref, dest);
}
else if (fz_isarray(dest))
{
return fz_arrayget(dest, 0);
}
else if (fz_isdict(dest))
{
dest = fz_dictgets(dest, "D");
return resolvedest(xref, dest);
}
else if (fz_isindirect(dest))
return dest;
return nil;
}
static void
gatherglobalinfo(void)
{
info = fz_malloc(sizeof (struct info));
if (!info)
die(fz_throw("out of memory"));
info->page = -1;
info->pageobj = nil;
info->ref = nil;
info->u.info.obj = nil;
info->u.info.obj = fz_dictgets(xref->trailer, "Info");
if (!fz_isindirect(info->u.info.obj))
die(fz_throw("not an indirect info object"));
cryptinfo = fz_malloc(sizeof (struct info));
cryptinfo->page = -1;
cryptinfo->pageobj = nil;
cryptinfo->ref = nil;
cryptinfo->u.crypt.obj = nil;
if (xref->crypt)
{
cryptinfo->ref = fz_dictgets(xref->trailer, "Encrypt");
if (!fz_isdict(cryptinfo->ref) && !fz_isindirect(cryptinfo->ref))
die(fz_throw("not an indirect crypt object"));
// XXX cryptinfo->u.crypt.obj = xref->crypt->encrypt;
}
}
static fz_obj *
resolvedest(pdf_xref *xref, fz_obj *dest)
{
if (fz_isname(dest) || fz_isstring(dest))
{
dest = pdf_lookupdest(xref, dest);
return resolvedest(xref, dest);
}
else if (fz_isarray(dest))
{
return dest; /* cf. http://code.google.com/p/sumatrapdf/issues/detail?id=275 */
}
else if (fz_isdict(dest))
{
dest = fz_dictgets(dest, "D");
return resolvedest(xref, dest);
}
else if (fz_isindirect(dest))
return dest;
return nil;
}
fz_error
pdf_loadpagetree(pdf_xref *xref)
{
struct info info;
fz_obj *catalog = fz_dictgets(xref->trailer, "Root");
fz_obj *pages = fz_dictgets(catalog, "Pages");
fz_obj *count = fz_dictgets(pages, "Count");
if (!fz_isdict(pages))
return fz_throw("missing page tree");
if (!fz_isint(count))
return fz_throw("missing page count");
xref->pagecap = fz_toint(count);
xref->pagelen = 0;
xref->pagerefs = fz_malloc(sizeof(fz_obj*) * xref->pagecap);
xref->pageobjs = fz_malloc(sizeof(fz_obj*) * xref->pagecap);
info.resources = nil;
info.mediabox = nil;
info.cropbox = nil;
info.rotate = nil;
pdf_loadpagetreenode(xref, pages, info);
return fz_okay;
}
static fz_obj *
resolvedest(pdf_xref *xref, fz_obj *dest)
{
if (fz_isname(dest))
{
dest = fz_dictget(xref->dests, dest);
if (dest)
pdf_resolve(&dest, xref); /* XXX */
return resolvedest(xref, dest);
}
else if (fz_isstring(dest))
{
dest = fz_dictget(xref->dests, dest);
if (dest)
pdf_resolve(&dest, xref); /* XXX */
return resolvedest(xref, dest);
}
else if (fz_isarray(dest))
{
return fz_arrayget(dest, 0);
}
else if (fz_isdict(dest))
{
dest = fz_dictgets(dest, "D");
return resolvedest(xref, dest);
}
else if (fz_isindirect(dest))
return dest;
return nil;
}
static fz_error
gatherpsobjs(int page, fz_obj *pageobj, fz_obj *dict)
{
int i;
for (i = 0; i < fz_dictlen(dict); i++)
{
fz_obj *ref;
fz_obj *xobjdict;
fz_obj *type;
fz_obj *subtype;
int k;
xobjdict = ref = fz_dictgetval(dict, i);
if (!fz_isdict(xobjdict))
return fz_throw("not a xobject dict (%d %d R)", fz_tonum(ref), fz_togen(ref));
type = fz_dictgets(xobjdict, "Subtype");
if (!fz_isname(type))
return fz_throw("not a xobject type (%d %d R)", fz_tonum(ref), fz_togen(ref));
if (strcmp(fz_toname(type), "Form"))
continue;
subtype = fz_dictgets(xobjdict, "Subtype2");
if (subtype && !fz_isname(subtype))
return fz_throw("not a xobject subtype (%d %d R)", fz_tonum(ref), fz_togen(ref));
if (strcmp(fz_toname(type), "PS") &&
(strcmp(fz_toname(type), "Form") || strcmp(fz_toname(subtype), "PS")))
continue;
for (k = 0; k < psobjs; k++)
if (fz_tonum(psobj[k]->ref) == fz_tonum(ref) &&
fz_togen(psobj[k]->ref) == fz_togen(ref))
break;
if (k < psobjs)
continue;
psobjs++;
psobj = fz_realloc(psobj, psobjs * sizeof (struct info *));
if (!psobj)
return fz_throw("out of memory");
psobj[psobjs - 1] = fz_malloc(sizeof (struct info));
if (!psobj[psobjs - 1])
return fz_throw("out of memory");
psobj[psobjs - 1]->page = page;
psobj[psobjs - 1]->pageobj = pageobj;
psobj[psobjs - 1]->ref = ref;
}
return fz_okay;
}
static int
pdf_xobjectusesblending(fz_obj *dict)
{
fz_obj *obj;
obj = fz_dictgets(dict, "Resources");
if (fz_isdict(obj) && pdf_resourcesuseblending(obj))
return 1;
return 0;
}
int
epdf_index_item_page_get(const Epdf_Document *doc, const Epdf_Index_Item *item)
{
fz_obj *dest;
int p;
int n;
int g;
if (!item || !item->link)
return -1;
if (PDF_LGOTO != item->link->kind)
return -1;
dest = item->link->dest;
p = 0;
if (fz_isint(dest))
{
p = fz_toint(dest);
return p;
}
if (fz_isdict(dest))
{
/* The destination is linked from a Go-To action's D array */
fz_obj *D;
D = fz_dictgets(dest, "D");
if (D && fz_isarray(D))
dest = fz_arrayget(D, 0);
}
n = fz_tonum(dest);
g = fz_togen(dest);
for (p = 1; p <= epdf_document_page_count_get(doc); p++)
{
fz_obj *page;
int np;
int gp;
page = pdf_getpageobject(doc->xref, p);
if (!page)
continue;
np = fz_tonum(page);
gp = fz_togen(page);
if (n == np && g == gp)
return p-1;
}
return 0;
}
static fz_error
gathershadings(int page, fz_obj *pageobj, fz_obj *dict)
{
int i;
for (i = 0; i < fz_dictlen(dict); i++)
{
fz_obj *ref;
fz_obj *shade;
fz_obj *type;
int k;
shade = ref = fz_dictgetval(dict, i);
if (!fz_isdict(shade))
return fz_throw("not a shading dict (%d %d R)", fz_tonum(ref), fz_togen(ref));
type = fz_dictgets(shade, "ShadingType");
if (!fz_isint(type) || fz_toint(type) < 1 || fz_toint(type) > 7)
{
fz_warn("not a shading type (%d %d R)", fz_tonum(ref), fz_togen(ref));
type = nil;
}
for (k = 0; k < shadings; k++)
if (fz_tonum(shading[k]->ref) == fz_tonum(ref) &&
fz_togen(shading[k]->ref) == fz_togen(ref))
break;
if (k < shadings)
continue;
shadings++;
shading = fz_realloc(shading, shadings * sizeof (struct info *));
if (!shading)
return fz_throw("out of memory");
shading[shadings - 1] = fz_malloc(sizeof (struct info));
if (!shading[shadings - 1])
return fz_throw("out of memory");
shading[shadings - 1]->page = page;
shading[shadings - 1]->pageobj = pageobj;
shading[shadings - 1]->ref = ref;
shading[shadings - 1]->u.shading.type = type;
}
return fz_okay;
}
static void sweepobj(fz_obj *obj)
{
int i;
if (fz_isindirect(obj))
sweepref(obj);
else if (fz_isdict(obj))
for (i = 0; i < fz_dictlen(obj); i++)
sweepobj(fz_dictgetval(obj, i));
else if (fz_isarray(obj))
for (i = 0; i < fz_arraylen(obj); i++)
sweepobj(fz_arrayget(obj, i));
}
static void addhexfilter(fz_obj *dict)
{
fz_obj *f, *dp, *newf, *newdp;
fz_obj *ahx, *nullobj;
ahx = fz_newname("ASCIIHexDecode");
nullobj = fz_newnull();
newf = newdp = nil;
f = fz_dictgets(dict, "Filter");
dp = fz_dictgets(dict, "DecodeParms");
if (fz_isname(f))
{
newf = fz_newarray(2);
fz_arraypush(newf, ahx);
fz_arraypush(newf, f);
f = newf;
if (fz_isdict(dp))
{
newdp = fz_newarray(2);
fz_arraypush(newdp, nullobj);
fz_arraypush(newdp, dp);
dp = newdp;
}
}
else if (fz_isarray(f))
{
fz_arrayinsert(f, ahx);
if (fz_isarray(dp))
fz_arrayinsert(dp, nullobj);
}
else
f = ahx;
fz_dictputs(dict, "Filter", f);
if (dp)
fz_dictputs(dict, "DecodeParms", dp);
fz_dropobj(ahx);
fz_dropobj(nullobj);
if (newf)
fz_dropobj(newf);
if (newdp)
fz_dropobj(newdp);
}
static void writeobject(int num, int gen)
{
fz_error error;
fz_obj *obj;
fz_obj *type;
error = pdf_loadobject(&obj, xref, num, gen);
if (error)
die(error);
/* skip ObjStm and XRef objects */
if (fz_isdict(obj))
{
type = fz_dictgets(obj, "Type");
if (fz_isname(type) && !strcmp(fz_toname(type), "ObjStm"))
{
uselist[num] = 0;
fz_dropobj(obj);
return;
}
if (fz_isname(type) && !strcmp(fz_toname(type), "XRef"))
{
uselist[num] = 0;
fz_dropobj(obj);
return;
}
}
if (!pdf_isstream(xref, num, gen))
{
fprintf(out, "%d %d obj\n", num, gen);
fz_fprintobj(out, obj, !doexpand);
fprintf(out, "endobj\n\n");
}
else
{
if (doexpand && !pdf_isjpximage(obj))
expandstream(obj, num, gen);
else
copystream(obj, num, gen);
}
fz_dropobj(obj);
}
static void renumberobj(fz_obj *obj)
{
int i;
if (fz_isdict(obj))
{
for (i = 0; i < fz_dictlen(obj); i++)
{
fz_obj *key = fz_dictgetkey(obj, i);
fz_obj *val = fz_dictgetval(obj, i);
if (fz_isindirect(val))
{
val = fz_newindirect(renumbermap[fz_tonum(val)], 0, xref);
fz_dictput(obj, key, val);
fz_dropobj(val);
}
else
{
renumberobj(val);
}
}
}
else if (fz_isarray(obj))
{
for (i = 0; i < fz_arraylen(obj); i++)
{
fz_obj *val = fz_arrayget(obj, i);
if (fz_isindirect(val))
{
val = fz_newindirect(renumbermap[fz_tonum(val)], 0, xref);
fz_arrayput(obj, i, val);
fz_dropobj(val);
}
else
{
renumberobj(val);
}
}
}
}
static fz_error
gatherimages(int page, fz_obj *pageobj, fz_obj *dict)
{
int i;
for (i = 0; i < fz_dictlen(dict); i++)
{
fz_obj *ref;
fz_obj *imagedict;
fz_obj *type;
fz_obj *width;
fz_obj *height;
fz_obj *bpc = nil;
fz_obj *filter = nil;
fz_obj *mask;
fz_obj *cs = nil;
fz_obj *altcs;
int k;
imagedict = ref = fz_dictgetval(dict, i);
if (!fz_isdict(imagedict))
return fz_throw("not an image dict (%d %d R)", fz_tonum(ref), fz_togen(ref));
type = fz_dictgets(imagedict, "Subtype");
if (!fz_isname(type))
return fz_throw("not an image subtype (%d %d R)", fz_tonum(ref), fz_togen(ref));
if (strcmp(fz_toname(type), "Image"))
continue;
filter = fz_dictgets(imagedict, "Filter");
if (filter && !fz_isname(filter) && !fz_isarray(filter))
return fz_throw("not an image filter (%d %d R)", fz_tonum(ref), fz_togen(ref));
mask = fz_dictgets(imagedict, "ImageMask");
altcs = nil;
cs = fz_dictgets(imagedict, "ColorSpace");
if (fz_isarray(cs))
{
fz_obj *cses = cs;
cs = fz_arrayget(cses, 0);
if (fz_isname(cs) && (!strcmp(fz_toname(cs), "DeviceN") || !strcmp(fz_toname(cs), "Separation")))
{
altcs = fz_arrayget(cses, 2);
if (fz_isarray(altcs))
altcs = fz_arrayget(altcs, 0);
}
}
if (fz_isbool(mask) && fz_tobool(mask))
{
if (cs)
fz_warn("image mask (%d %d R) may not have colorspace", fz_tonum(ref), fz_togen(ref));
}
if (cs && !fz_isname(cs))
return fz_throw("not an image colorspace (%d %d R)", fz_tonum(ref), fz_togen(ref));
if (altcs && !fz_isname(altcs))
return fz_throw("not an image alternate colorspace (%d %d R)", fz_tonum(ref), fz_togen(ref));
width = fz_dictgets(imagedict, "Width");
if (!fz_isint(width))
return fz_throw("not an image width (%d %d R)", fz_tonum(ref), fz_togen(ref));
height = fz_dictgets(imagedict, "Height");
if (!fz_isint(height))
return fz_throw("not an image height (%d %d R)", fz_tonum(ref), fz_togen(ref));
bpc = fz_dictgets(imagedict, "BitsPerComponent");
if (!fz_tobool(mask) && !fz_isint(bpc))
return fz_throw("not an image bits per component (%d %d R)", fz_tonum(ref), fz_togen(ref));
if (fz_tobool(mask) && fz_isint(bpc) && fz_toint(bpc) != 1)
return fz_throw("not an image mask bits per component (%d %d R)", fz_tonum(ref), fz_togen(ref));
for (k = 0; k < images; k++)
if (fz_tonum(image[k]->ref) == fz_tonum(ref) &&
fz_togen(image[k]->ref) == fz_togen(ref))
break;
if (k < images)
continue;
images++;
image = fz_realloc(image, images * sizeof (struct info *));
if (!image)
return fz_throw("out of memory");
image[images - 1] = fz_malloc(sizeof (struct info));
if (!image[images - 1])
return fz_throw("out of memory");
image[images - 1]->page = page;
image[images - 1]->pageobj = pageobj;
image[images - 1]->ref = ref;
image[images - 1]->u.image.width = width;
image[images - 1]->u.image.height = height;
image[images - 1]->u.image.bpc = bpc;
image[images - 1]->u.image.filter = filter;
image[images - 1]->u.image.cs = cs;
image[images - 1]->u.image.altcs = altcs;
}
return fz_okay;
}
fz_error *
pdf_loadtype3font(pdf_font **fontp, pdf_xref *xref, fz_obj *dict, fz_obj *ref)
{
fz_error *error;
char buf[256];
char *estrings[256];
pdf_font *font;
fz_obj *encoding;
fz_obj *widths;
fz_obj *resources;
fz_obj *charprocs;
fz_obj *obj;
int first, last;
int i, k, n;
fz_rect bbox;
obj = fz_dictgets(dict, "Name");
if (obj)
strlcpy(buf, fz_toname(obj), sizeof buf);
else
sprintf(buf, "Unnamed-T3");
font = pdf_newfont(buf);
if (!font)
return fz_throw("outofmem: font struct");
pdf_logfont("load type3 font %d %d (%p) {\n", fz_tonum(ref), fz_togen(ref), font);
pdf_logfont("name %s\n", buf);
font->super.render = t3render;
font->super.drop = (void(*)(fz_font*)) t3dropfont;
obj = fz_dictgets(dict, "FontMatrix");
font->matrix = pdf_tomatrix(obj);
pdf_logfont("matrix [%g %g %g %g %g %g]\n",
font->matrix.a, font->matrix.b,
font->matrix.c, font->matrix.d,
font->matrix.e, font->matrix.f);
obj = fz_dictgets(dict, "FontBBox");
bbox = pdf_torect(obj);
pdf_logfont("bbox [%g %g %g %g]\n",
bbox.x0, bbox.y0,
bbox.x1, bbox.y1);
bbox = fz_transformaabb(font->matrix, bbox);
bbox.x0 = fz_floor(bbox.x0 * 1000);
bbox.y0 = fz_floor(bbox.y0 * 1000);
bbox.x1 = fz_ceil(bbox.x1 * 1000);
bbox.y1 = fz_ceil(bbox.y1 * 1000);
fz_setfontbbox((fz_font*)font, bbox.x0, bbox.y0, bbox.x1, bbox.y1);
/*
* Encoding
*/
for (i = 0; i < 256; i++)
estrings[i] = nil;
encoding = fz_dictgets(dict, "Encoding");
if (!encoding) {
error = fz_throw("syntaxerror: Type3 font missing Encoding");
goto cleanup;
}
error = pdf_resolve(&encoding, xref);
if (error)
goto cleanup;
if (fz_isname(obj))
pdf_loadencoding(estrings, fz_toname(encoding));
if (fz_isdict(encoding))
{
fz_obj *base, *diff, *item;
base = fz_dictgets(encoding, "BaseEncoding");
if (fz_isname(base))
pdf_loadencoding(estrings, fz_toname(base));
diff = fz_dictgets(encoding, "Differences");
if (fz_isarray(diff))
{
n = fz_arraylen(diff);
k = 0;
for (i = 0; i < n; i++)
{
item = fz_arrayget(diff, i);
if (fz_isint(item))
k = fz_toint(item);
if (fz_isname(item))
estrings[k++] = fz_toname(item);
if (k < 0) k = 0;
if (k > 255) k = 255;
}
}
}
fz_dropobj(encoding);
error = pdf_newidentitycmap(&font->encoding, 0, 1);
if (error)
goto cleanup;
error = pdf_loadtounicode(font, xref,
estrings, nil, fz_dictgets(dict, "ToUnicode"));
if (error)
goto cleanup;
/*
* Widths
*/
fz_setdefaulthmtx((fz_font*)font, 0);
first = fz_toint(fz_dictgets(dict, "FirstChar"));
last = fz_toint(fz_dictgets(dict, "LastChar"));
widths = fz_dictgets(dict, "Widths");
if (!widths) {
error = fz_throw("syntaxerror: Type3 font missing Widths");
goto cleanup;
}
error = pdf_resolve(&widths, xref);
if (error)
goto cleanup;
for (i = first; i <= last; i++)
{
float w = fz_toreal(fz_arrayget(widths, i - first));
w = font->matrix.a * w * 1000.0;
error = fz_addhmtx((fz_font*)font, i, i, w);
if (error) {
fz_dropobj(widths);
goto cleanup;
}
}
fz_dropobj(widths);
error = fz_endhmtx((fz_font*)font);
if (error)
goto cleanup;
/*
* Resources
*/
resources = nil;
obj = fz_dictgets(dict, "Resources");
if (obj)
{
error = pdf_resolve(&obj, xref);
if (error)
goto cleanup;
error = pdf_loadresources(&resources, xref, obj);
fz_dropobj(obj);
if (error)
goto cleanup;
}
else
pdf_logfont("no resource dict!\n");
/*
* CharProcs
*/
charprocs = fz_dictgets(dict, "CharProcs");
if (!charprocs)
{
error = fz_throw("syntaxerror: Type3 font missing CharProcs");
goto cleanup2;
}
error = pdf_resolve(&charprocs, xref);
if (error)
goto cleanup2;
for (i = 0; i < 256; i++)
{
if (estrings[i])
{
obj = fz_dictgets(charprocs, estrings[i]);
if (obj)
{
pdf_logfont("load charproc %s {\n", estrings[i]);
error = loadcharproc(&font->charprocs[i], xref, resources, obj);
if (error)
goto cleanup2;
error = fz_optimizetree(font->charprocs[i]);
if (error)
goto cleanup2;
pdf_logfont("}\n");
}
}
}
fz_dropobj(charprocs);
if (resources)
fz_dropobj(resources);
pdf_logfont("}\n");
*fontp = font;
return fz_okay;
cleanup2:
if (resources)
fz_dropobj(resources);
cleanup:
fz_dropfont((fz_font*)font);
return fz_rethrow(error, "cannot load type3 font");
}
static fz_error
loadsimplefont(pdf_fontdesc **fontdescp, pdf_xref *xref, fz_obj *dict)
{
fz_error error;
fz_obj *descriptor;
fz_obj *encoding;
fz_obj *widths;
unsigned short *etable = nil;
pdf_fontdesc *fontdesc;
fz_bbox bbox;
FT_Face face;
FT_CharMap cmap;
int kind;
int symbolic;
char *basefont;
char *fontname;
char *estrings[256];
char ebuffer[256][32];
int i, k, n;
int fterr;
basefont = fz_toname(fz_dictgets(dict, "BaseFont"));
fontname = cleanfontname(basefont);
/* Load font file */
fontdesc = pdf_newfontdesc();
pdf_logfont("load simple font (%d %d R) ptr=%p {\n", fz_tonum(dict), fz_togen(dict), fontdesc);
pdf_logfont("basefont %s -> %s\n", basefont, fontname);
descriptor = fz_dictgets(dict, "FontDescriptor");
if (descriptor)
error = pdf_loadfontdescriptor(fontdesc, xref, descriptor, nil, basefont);
else
error = pdf_loadbuiltinfont(fontdesc, fontname);
if (error)
goto cleanup;
face = fontdesc->font->ftface;
kind = ftkind(face);
pdf_logfont("ft name '%s' '%s'\n", face->family_name, face->style_name);
bbox.x0 = (face->bbox.xMin * 1000) / face->units_per_EM;
bbox.y0 = (face->bbox.yMin * 1000) / face->units_per_EM;
bbox.x1 = (face->bbox.xMax * 1000) / face->units_per_EM;
bbox.y1 = (face->bbox.yMax * 1000) / face->units_per_EM;
pdf_logfont("ft bbox [%d %d %d %d]\n", bbox.x0, bbox.y0, bbox.x1, bbox.y1);
if (bbox.x0 == bbox.x1)
fz_setfontbbox(fontdesc->font, -1000, -1000, 2000, 2000);
else
fz_setfontbbox(fontdesc->font, bbox.x0, bbox.y0, bbox.x1, bbox.y1);
/* Encoding */
symbolic = fontdesc->flags & 4;
if (face->num_charmaps > 0)
cmap = face->charmaps[0];
else
cmap = nil;
for (i = 0; i < face->num_charmaps; i++)
{
FT_CharMap test = face->charmaps[i];
if (kind == TYPE1)
{
if (test->platform_id == 7)
cmap = test;
}
if (kind == TRUETYPE)
{
if (test->platform_id == 1 && test->encoding_id == 0)
cmap = test;
if (test->platform_id == 3 && test->encoding_id == 1)
cmap = test;
}
}
if (cmap)
{
fterr = FT_Set_Charmap(face, cmap);
if (fterr)
fz_warn("freetype could not set cmap: %s", ft_errorstring(fterr));
}
else
fz_warn("freetype could not find any cmaps");
etable = fz_malloc(sizeof(unsigned short) * 256);
for (i = 0; i < 256; i++)
{
estrings[i] = nil;
etable[i] = 0;
}
encoding = fz_dictgets(dict, "Encoding");
if (encoding)
{
if (fz_isname(encoding))
pdf_loadencoding(estrings, fz_toname(encoding));
if (fz_isdict(encoding))
{
fz_obj *base, *diff, *item;
base = fz_dictgets(encoding, "BaseEncoding");
if (fz_isname(base))
pdf_loadencoding(estrings, fz_toname(base));
else if (!fontdesc->isembedded && !symbolic)
pdf_loadencoding(estrings, "StandardEncoding");
/* cf. http://bugs.ghostscript.com/show_bug.cgi?id=690615 and http://code.google.com/p/sumatrapdf/issues/detail?id=687 */
/* try to extract an encoding from the font or synthesize a likely one */
/* note: FT_Get_Name_Index fails for symbolic CFF fonts, so let them be encoded by index */
else if (!fontdesc->encoding && !ftloadt1encoding(face, estrings) && !(symbolic && !strcmp(FT_Get_X11_Font_Format(face), "CFF")))
pdf_loadencoding(estrings, "StandardEncoding");
diff = fz_dictgets(encoding, "Differences");
if (fz_isarray(diff))
{
n = fz_arraylen(diff);
k = 0;
for (i = 0; i < n; i++)
{
item = fz_arrayget(diff, i);
if (fz_isint(item))
k = fz_toint(item);
if (fz_isname(item))
estrings[k++] = fz_toname(item);
if (k < 0) k = 0;
if (k > 255) k = 255;
}
}
}
}
/* start with the builtin encoding */
for (i = 0; i < 256; i++)
etable[i] = ftcharindex(face, i);
/* encode by glyph name where we can */
if (kind == TYPE1)
{
pdf_logfont("encode type1/cff by strings\n");
for (i = 0; i < 256; i++)
{
if (estrings[i])
{
etable[i] = FT_Get_Name_Index(face, estrings[i]);
if (etable[i] == 0)
{
int aglcode = pdf_lookupagl(estrings[i]);
char **aglnames = pdf_lookupaglnames(aglcode);
while (*aglnames)
{
etable[i] = FT_Get_Name_Index(face, *aglnames);
if (etable[i])
break;
aglnames++;
}
}
}
}
}
/* encode by glyph name where we can */
if (kind == TRUETYPE)
{
/* Unicode cmap */
if (!symbolic && face->charmap && face->charmap->platform_id == 3)
{
pdf_logfont("encode truetype via unicode\n");
for (i = 0; i < 256; i++)
{
if (estrings[i])
{
int aglcode = pdf_lookupagl(estrings[i]);
if (!aglcode)
etable[i] = FT_Get_Name_Index(face, estrings[i]);
else
etable[i] = ftcharindex(face, aglcode);
}
}
}
/* MacRoman cmap */
else if (!symbolic && face->charmap && face->charmap->platform_id == 1)
{
pdf_logfont("encode truetype via macroman\n");
for (i = 0; i < 256; i++)
{
if (estrings[i])
{
k = mrecode(estrings[i]);
if (k <= 0)
etable[i] = FT_Get_Name_Index(face, estrings[i]);
else
etable[i] = ftcharindex(face, k);
}
}
}
/* Symbolic cmap */
else
{
pdf_logfont("encode truetype symbolic\n");
for (i = 0; i < 256; i++)
{
if (estrings[i])
{
etable[i] = FT_Get_Name_Index(face, estrings[i]);
if (etable[i] == 0)
etable[i] = ftcharindex(face, i);
}
}
}
}
/* try to reverse the glyph names from the builtin encoding */
for (i = 0; i < 256; i++)
{
if (etable[i] && !estrings[i])
{
if (FT_HAS_GLYPH_NAMES(face))
{
fterr = FT_Get_Glyph_Name(face, etable[i], ebuffer[i], 32);
if (fterr)
fz_warn("freetype get glyph name (gid %d): %s", etable[i], ft_errorstring(fterr));
if (ebuffer[i][0])
estrings[i] = ebuffer[i];
}
else
{
estrings[i] = (char*) pdf_winansi[i]; /* discard const */
}
}
}
/* Prevent encoding Differences from being overwritten by reloading them */
/* cf. http://code.google.com/p/sumatrapdf/issues/detail?id=115 */
if (fz_isdict(encoding))
{
fz_obj *diff, *item;
diff = fz_dictgets(encoding, "Differences");
if (fz_isarray(diff))
{
n = fz_arraylen(diff);
k = 0;
for (i = 0; i < n; i++)
{
item = fz_arrayget(diff, i);
if (fz_isint(item))
k = fz_toint(item);
if (fz_isname(item))
estrings[k++] = fz_toname(item);
if (k < 0) k = 0;
if (k > 255) k = 255;
}
}
}
fontdesc->encoding = pdf_newidentitycmap(0, 1);
fontdesc->ncidtogid = 256;
fontdesc->cidtogid = etable;
error = pdf_loadtounicode(fontdesc, xref, estrings, nil, fz_dictgets(dict, "ToUnicode"));
if (error)
goto cleanup;
/* Widths */
pdf_setdefaulthmtx(fontdesc, fontdesc->missingwidth);
widths = fz_dictgets(dict, "Widths");
if (widths)
{
int first, last;
first = fz_toint(fz_dictgets(dict, "FirstChar"));
last = fz_toint(fz_dictgets(dict, "LastChar"));
if (first < 0 || last > 255 || first > last)
first = last = 0;
for (i = 0; i < last - first + 1; i++)
{
int wid = fz_toint(fz_arrayget(widths, i));
pdf_addhmtx(fontdesc, i + first, i + first, wid);
}
}
else
{
fterr = FT_Set_Char_Size(face, 1000, 1000, 72, 72);
if (fterr)
fz_warn("freetype set character size: %s", ft_errorstring(fterr));
for (i = 0; i < 256; i++)
{
pdf_addhmtx(fontdesc, i, i, ftwidth(fontdesc, i));
}
}
pdf_endhmtx(fontdesc);
pdf_logfont("}\n");
*fontdescp = fontdesc;
return fz_okay;
cleanup:
if (etable != fontdesc->cidtogid)
fz_free(etable);
pdf_dropfont(fontdesc);
return fz_rethrow(error, "cannot load simple font (%d %d R)", fz_tonum(dict), fz_togen(dict));
}
fz_error
pdf_loadtype3font(pdf_fontdesc **fontdescp, pdf_xref *xref, fz_obj *rdb, fz_obj *dict)
{
fz_error error;
char buf[256];
char *estrings[256];
pdf_fontdesc *fontdesc;
fz_obj *encoding;
fz_obj *widths;
fz_obj *resources;
fz_obj *charprocs;
fz_obj *obj;
int first, last;
int i, k, n;
fz_rect bbox;
fz_matrix matrix;
obj = fz_dictgets(dict, "Name");
if (fz_isname(obj))
strlcpy(buf, fz_toname(obj), sizeof buf);
else
sprintf(buf, "Unnamed-T3");
fontdesc = pdf_newfontdesc();
pdf_logfont("load type3 font (%d %d R) ptr=%p {\n", fz_tonum(dict), fz_togen(dict), fontdesc);
pdf_logfont("name %s\n", buf);
obj = fz_dictgets(dict, "FontMatrix");
matrix = pdf_tomatrix(obj);
pdf_logfont("matrix [%g %g %g %g %g %g]\n",
matrix.a, matrix.b,
matrix.c, matrix.d,
matrix.e, matrix.f);
obj = fz_dictgets(dict, "FontBBox");
bbox = pdf_torect(obj);
pdf_logfont("bbox [%g %g %g %g]\n",
bbox.x0, bbox.y0,
bbox.x1, bbox.y1);
bbox = fz_transformaabb(matrix, bbox);
bbox.x0 = fz_floor(bbox.x0 * 1000);
bbox.y0 = fz_floor(bbox.y0 * 1000);
bbox.x1 = fz_ceil(bbox.x1 * 1000);
bbox.y1 = fz_ceil(bbox.y1 * 1000);
fontdesc->font = fz_newtype3font(buf, matrix);
fz_setfontbbox(fontdesc->font, bbox.x0, bbox.y0, bbox.x1, bbox.y1);
/*
* Encoding
*/
for (i = 0; i < 256; i++)
estrings[i] = nil;
encoding = fz_dictgets(dict, "Encoding");
if (!encoding)
{
error = fz_throw("syntaxerror: Type3 font missing Encoding");
goto cleanup;
}
if (fz_isname(encoding))
pdf_loadencoding(estrings, fz_toname(encoding));
if (fz_isdict(encoding))
{
fz_obj *base, *diff, *item;
base = fz_dictgets(encoding, "BaseEncoding");
if (fz_isname(base))
pdf_loadencoding(estrings, fz_toname(base));
diff = fz_dictgets(encoding, "Differences");
if (fz_isarray(diff))
{
n = fz_arraylen(diff);
k = 0;
for (i = 0; i < n; i++)
{
item = fz_arrayget(diff, i);
if (fz_isint(item))
k = fz_toint(item);
if (fz_isname(item))
estrings[k++] = fz_toname(item);
if (k < 0) k = 0;
if (k > 255) k = 255;
}
}
}
fontdesc->encoding = pdf_newidentitycmap(0, 1);
error = pdf_loadtounicode(fontdesc, xref, estrings, nil, fz_dictgets(dict, "ToUnicode"));
if (error)
goto cleanup;
/*
* Widths
*/
pdf_setdefaulthmtx(fontdesc, 0);
first = fz_toint(fz_dictgets(dict, "FirstChar"));
last = fz_toint(fz_dictgets(dict, "LastChar"));
widths = fz_dictgets(dict, "Widths");
if (!widths)
{
error = fz_throw("syntaxerror: Type3 font missing Widths");
goto cleanup;
}
for (i = first; i <= last; i++)
{
float w = fz_toreal(fz_arrayget(widths, i - first));
w = fontdesc->font->t3matrix.a * w * 1000.0;
fontdesc->font->t3widths[i] = w * 0.001;
pdf_addhmtx(fontdesc, i, i, w);
}
pdf_endhmtx(fontdesc);
/*
* Resources
*/
resources = fz_dictgets(dict, "Resources");
/* Inherit page's resource dict if type3 font does not have one */
if (!resources && rdb)
resources = rdb;
else if (!resources && !rdb)
fz_warn("no resource dictionary for type 3 font!");
/*
* CharProcs
*/
charprocs = fz_dictgets(dict, "CharProcs");
if (!charprocs)
{
error = fz_throw("syntaxerror: Type3 font missing CharProcs");
goto cleanup;
}
for (i = 0; i < 256; i++)
{
if (estrings[i])
{
obj = fz_dictgets(charprocs, estrings[i]);
if (obj)
{
pdf_logfont("load charproc %s {\n", estrings[i]);
error = loadcharproc(&fontdesc->font->t3procs[i], xref, resources, obj);
if (error)
goto cleanup;
pdf_logfont("}\n");
}
}
}
pdf_logfont("}\n");
*fontdescp = fontdesc;
return fz_okay;
cleanup:
fz_dropfont(fontdesc->font);
fz_free(fontdesc);
return fz_rethrow(error, "cannot load type3 font");
}
fz_error
pdf_openxrefwithstream(pdf_xref **xrefp, fz_stream *file, char *password)
{
pdf_xref *xref;
fz_error error;
fz_obj *encrypt;
fz_obj *id;
xref = fz_malloc(sizeof(pdf_xref));
memset(xref, 0, sizeof(pdf_xref));
pdf_logxref("openxref %p\n", xref);
xref->file = fz_keepstream(file);
error = pdf_loadxref(xref, xref->scratch, sizeof xref->scratch);
if (error)
{
fz_catch(error, "trying to repair");
if (xref->table)
{
fz_free(xref->table);
xref->table = NULL;
xref->len = 0;
xref->cap = 0;
}
error = pdf_repairxref(xref, xref->scratch, sizeof xref->scratch);
if (error)
{
pdf_freexref(xref);
return fz_rethrow(error, "cannot repair document");
}
}
encrypt = fz_dictgets(xref->trailer, "Encrypt");
id = fz_dictgets(xref->trailer, "ID");
if (fz_isdict(encrypt))
{
error = pdf_newcrypt(&xref->crypt, encrypt, id);
if (error)
{
pdf_freexref(xref);
return fz_rethrow(error, "cannot decrypt document");
}
}
if (pdf_needspassword(xref))
{
/* Only care if we have a password */
if (password)
{
int okay = pdf_authenticatepassword(xref, password);
if (!okay)
{
pdf_freexref(xref);
return fz_throw("invalid password");
}
}
}
*xrefp = xref;
return fz_okay;
}
static fz_error
loadsimplefont(pdf_fontdesc **fontdescp, pdf_xref *xref, fz_obj *dict)
{
fz_error error;
fz_obj *descriptor = nil;
fz_obj *encoding = nil;
fz_obj *widths = nil;
unsigned short *etable = nil;
pdf_fontdesc *fontdesc;
fz_irect bbox;
FT_Face face;
FT_CharMap cmap;
int kind;
int symbolic;
char *basefont;
char *fontname;
char *estrings[256];
char ebuffer[256][32];
int i, k, n;
int fterr;
basefont = fz_toname(fz_dictgets(dict, "BaseFont"));
fontname = cleanfontname(basefont);
/*
* Load font file
*/
fontdesc = pdf_newfontdesc();
if (!fontdesc)
return fz_rethrow(-1, "out of memory");
pdf_logfont("load simple font (%d %d R) ptr=%p {\n", fz_tonum(dict), fz_togen(dict), fontdesc);
pdf_logfont("basefont0 %s\n", basefont);
pdf_logfont("basefont1 %s\n", fontname);
descriptor = fz_dictgets(dict, "FontDescriptor");
if (descriptor && basefont == fontname)
error = pdf_loadfontdescriptor(fontdesc, xref, descriptor, nil);
else
error = pdf_loadbuiltinfont(fontdesc, fontname);
if (error)
goto cleanup;
face = fontdesc->font->ftface;
kind = ftkind(face);
pdf_logfont("ft name '%s' '%s'\n", face->family_name, face->style_name);
bbox.x0 = (face->bbox.xMin * 1000) / face->units_per_EM;
bbox.y0 = (face->bbox.yMin * 1000) / face->units_per_EM;
bbox.x1 = (face->bbox.xMax * 1000) / face->units_per_EM;
bbox.y1 = (face->bbox.yMax * 1000) / face->units_per_EM;
pdf_logfont("ft bbox [%d %d %d %d]\n", bbox.x0, bbox.y0, bbox.x1, bbox.y1);
if (bbox.x0 == bbox.x1)
fz_setfontbbox(fontdesc->font, -1000, -1000, 2000, 2000);
else
fz_setfontbbox(fontdesc->font, bbox.x0, bbox.y0, bbox.x1, bbox.y1);
/*
* Encoding
*/
symbolic = fontdesc->flags & 4;
if (face->num_charmaps > 0)
cmap = face->charmaps[0];
else
cmap = nil;
for (i = 0; i < face->num_charmaps; i++)
{
FT_CharMap test = face->charmaps[i];
if (kind == TYPE1)
{
if (test->platform_id == 7)
cmap = test;
}
if (kind == TRUETYPE)
{
if (test->platform_id == 1 && test->encoding_id == 0)
cmap = test;
if (test->platform_id == 3 && test->encoding_id == 1)
cmap = test;
}
}
if (cmap)
{
fterr = FT_Set_Charmap(face, cmap);
if (fterr)
fz_warn("freetype could not set cmap: %s", ft_errorstring(fterr));
}
else
fz_warn("freetype could not find any cmaps");
etable = fz_malloc(sizeof(unsigned short) * 256);
if (!etable)
goto cleanup;
for (i = 0; i < 256; i++)
{
estrings[i] = nil;
etable[i] = 0;
}
encoding = fz_dictgets(dict, "Encoding");
if (encoding && !(kind == TRUETYPE && symbolic))
{
if (fz_isname(encoding))
pdf_loadencoding(estrings, fz_toname(encoding));
if (fz_isdict(encoding))
{
fz_obj *base, *diff, *item;
base = fz_dictgets(encoding, "BaseEncoding");
if (fz_isname(base))
pdf_loadencoding(estrings, fz_toname(base));
else if (!fontdesc->isembedded)
pdf_loadencoding(estrings, "StandardEncoding");
diff = fz_dictgets(encoding, "Differences");
if (fz_isarray(diff))
{
n = fz_arraylen(diff);
k = 0;
for (i = 0; i < n; i++)
{
item = fz_arrayget(diff, i);
if (fz_isint(item))
k = fz_toint(item);
if (fz_isname(item))
estrings[k++] = fz_toname(item);
if (k < 0) k = 0;
if (k > 255) k = 255;
}
}
}
if (kind == TYPE1)
{
pdf_logfont("encode type1/cff by strings\n");
for (i = 0; i < 256; i++)
if (estrings[i])
etable[i] = FT_Get_Name_Index(face, estrings[i]);
else
etable[i] = ftcharindex(face, i);
}
if (kind == TRUETYPE)
{
/* Unicode cmap */
if (face->charmap && face->charmap->platform_id == 3)
{
pdf_logfont("encode truetype via unicode\n");
for (i = 0; i < 256; i++)
if (estrings[i])
{
int aglbuf[256];
int aglnum;
aglnum = pdf_lookupagl(estrings[i], aglbuf, nelem(aglbuf));
if (aglnum != 1)
etable[i] = FT_Get_Name_Index(face, estrings[i]);
else
etable[i] = ftcharindex(face, aglbuf[0]);
}
else
etable[i] = ftcharindex(face, i);
}
/* MacRoman cmap */
else if (face->charmap && face->charmap->platform_id == 1)
{
pdf_logfont("encode truetype via macroman\n");
for (i = 0; i < 256; i++)
if (estrings[i])
{
k = mrecode(estrings[i]);
if (k <= 0)
etable[i] = FT_Get_Name_Index(face, estrings[i]);
else
etable[i] = ftcharindex(face, k);
}
else
etable[i] = ftcharindex(face, i);
}
/* Symbolic cmap */
else
{
pdf_logfont("encode truetype symbolic\n");
for (i = 0; i < 256; i++)
{
etable[i] = ftcharindex(face, i);
fterr = FT_Get_Glyph_Name(face, etable[i], ebuffer[i], 32);
if (fterr)
{
error = fz_throw("freetype get glyph name (gid %d): %s", etable[i], ft_errorstring(fterr));
goto cleanup;
}
if (ebuffer[i][0])
estrings[i] = ebuffer[i];
}
}
}
}
else
{
pdf_logfont("encode builtin\n");
for (i = 0; i < 256; i++)
{
etable[i] = ftcharindex(face, i);
if (etable[i] == 0)
continue;
if (FT_HAS_GLYPH_NAMES(face))
{
fterr = FT_Get_Glyph_Name(face, etable[i], ebuffer[i], 32);
if (fterr)
{
error = fz_throw("freetype get glyph name (gid %d): %s", etable[i], ft_errorstring(fterr));
goto cleanup;
}
if (ebuffer[i][0])
estrings[i] = ebuffer[i];
}
}
}
error = pdf_newidentitycmap(&fontdesc->encoding, 0, 1);
if (error)
goto cleanup;
fontdesc->ncidtogid = 256;
fontdesc->cidtogid = etable;
error = pdf_loadtounicode(fontdesc, xref, estrings, nil, fz_dictgets(dict, "ToUnicode"));
if (error)
goto cleanup;
/*
* Widths
*/
pdf_setdefaulthmtx(fontdesc, fontdesc->missingwidth);
widths = fz_dictgets(dict, "Widths");
if (widths)
{
int first, last;
first = fz_toint(fz_dictgets(dict, "FirstChar"));
last = fz_toint(fz_dictgets(dict, "LastChar"));
if (first < 0 || last > 255 || first > last)
first = last = 0;
for (i = 0; i < last - first + 1; i++)
{
int wid = fz_toint(fz_arrayget(widths, i));
error = pdf_addhmtx(fontdesc, i + first, i + first, wid);
if (error)
goto cleanup;
}
}
else
{
fterr = FT_Set_Char_Size(face, 1000, 1000, 72, 72);
if (fterr)
fz_warn("freetype set character size: %s", ft_errorstring(fterr));
for (i = 0; i < 256; i++)
{
error = pdf_addhmtx(fontdesc, i, i, ftwidth(fontdesc, i));
if (error)
goto cleanup;
}
}
error = pdf_endhmtx(fontdesc);
if (error)
goto cleanup;
pdf_logfont("}\n");
*fontdescp = fontdesc;
return fz_okay;
cleanup:
fz_free(etable);
fz_dropfont(fontdesc->font);
fz_free(fontdesc);
return fz_rethrow(error, "cannot load simple font");
}
static fz_error
gatherpatterns(int page, fz_obj *pageobj, fz_obj *dict)
{
int i;
for (i = 0; i < fz_dictlen(dict); i++)
{
fz_obj *ref;
fz_obj *patterndict;
fz_obj *type;
fz_obj *paint = nil;
fz_obj *tiling = nil;
int k;
patterndict = ref = fz_dictgetval(dict, i);
if (!fz_isdict(patterndict))
return fz_throw("not a pattern dict (%d %d R)", fz_tonum(ref), fz_togen(ref));
type = fz_dictgets(patterndict, "PatternType");
if (!fz_isint(type) || fz_toint(type) < 1 || fz_toint(type) > 2)
{
fz_warn("not a pattern type (%d %d R)", fz_tonum(ref), fz_togen(ref));
type = nil;
}
if (fz_toint(type) == 1)
{
paint = fz_dictgets(patterndict, "PaintType");
if (!fz_isint(paint) || fz_toint(paint) < 1 || fz_toint(paint) > 2)
{
fz_warn("not a pattern paint type (%d %d R)", fz_tonum(ref), fz_togen(ref));
paint = nil;
}
tiling = fz_dictgets(patterndict, "TilingType");
if (!fz_isint(tiling) || fz_toint(tiling) < 1 || fz_toint(tiling) > 3)
{
fz_warn("not a pattern tiling type (%d %d R)", fz_tonum(ref), fz_togen(ref));
tiling = nil;
}
}
for (k = 0; k < patterns; k++)
if (fz_tonum(pattern[k]->ref) == fz_tonum(ref) &&
fz_togen(pattern[k]->ref) == fz_togen(ref))
break;
if (k < patterns)
continue;
patterns++;
pattern = fz_realloc(pattern, patterns * sizeof (struct info *));
if (!pattern)
return fz_throw("out of memory");
pattern[patterns - 1] = fz_malloc(sizeof (struct info));
if (!pattern[patterns - 1])
return fz_throw("out of memory");
pattern[patterns - 1]->page = page;
pattern[patterns - 1]->pageobj = pageobj;
pattern[patterns - 1]->ref = ref;
pattern[patterns - 1]->u.pattern.pattern = type;
pattern[patterns - 1]->u.pattern.paint = paint;
pattern[patterns - 1]->u.pattern.tiling = tiling;
}
return fz_okay;
}
static fz_error
pdf_loadimageimp(fz_pixmap **imgp, pdf_xref *xref, fz_obj *rdb, fz_obj *dict, fz_stream *cstm, int forcemask)
{
fz_stream *stm;
fz_pixmap *tile;
fz_obj *obj, *res;
fz_error error;
int w, h, bpc, n;
int imagemask;
int interpolate;
int indexed;
fz_colorspace *colorspace;
fz_pixmap *mask; /* explicit mask/softmask image */
int usecolorkey;
int colorkey[FZ_MAXCOLORS * 2];
float decode[FZ_MAXCOLORS * 2];
int scale;
int stride;
unsigned char *samples;
int i, len;
/* special case for JPEG2000 images */
if (pdf_isjpximage(dict))
{
tile = nil;
error = pdf_loadjpximage(&tile, xref, dict);
if (error)
return fz_rethrow(error, "cannot load jpx image");
if (forcemask)
{
if (tile->n != 2)
{
fz_droppixmap(tile);
return fz_throw("softmask must be grayscale");
}
mask = fz_alphafromgray(tile, 1);
fz_droppixmap(tile);
*imgp = mask;
return fz_okay;
}
*imgp = tile;
return fz_okay;
}
w = fz_toint(fz_dictgetsa(dict, "Width", "W"));
h = fz_toint(fz_dictgetsa(dict, "Height", "H"));
bpc = fz_toint(fz_dictgetsa(dict, "BitsPerComponent", "BPC"));
imagemask = fz_tobool(fz_dictgetsa(dict, "ImageMask", "IM"));
interpolate = fz_tobool(fz_dictgetsa(dict, "Interpolate", "I"));
indexed = 0;
usecolorkey = 0;
colorspace = nil;
mask = nil;
if (imagemask)
bpc = 1;
if (w == 0)
return fz_throw("image width is zero");
if (h == 0)
return fz_throw("image height is zero");
if (bpc == 0)
return fz_throw("image depth is zero");
if (w > (1 << 16))
return fz_throw("image is too wide");
if (h > (1 << 16))
return fz_throw("image is too high");
obj = fz_dictgetsa(dict, "ColorSpace", "CS");
if (obj && !imagemask && !forcemask)
{
/* colorspace resource lookup is only done for inline images */
if (fz_isname(obj))
{
res = fz_dictget(fz_dictgets(rdb, "ColorSpace"), obj);
if (res)
obj = res;
}
error = pdf_loadcolorspace(&colorspace, xref, obj);
if (error)
return fz_rethrow(error, "cannot load image colorspace");
if (!strcmp(colorspace->name, "Indexed"))
indexed = 1;
n = colorspace->n;
}
else
{
n = 1;
}
obj = fz_dictgetsa(dict, "Decode", "D");
if (obj)
{
for (i = 0; i < n * 2; i++)
decode[i] = fz_toreal(fz_arrayget(obj, i));
}
else
{
float maxval = indexed ? (1 << bpc) - 1 : 1;
for (i = 0; i < n * 2; i++)
decode[i] = i & 1 ? maxval : 0;
}
obj = fz_dictgetsa(dict, "SMask", "Mask");
if (fz_isdict(obj))
{
/* Not allowed for inline images */
if (!cstm)
{
error = pdf_loadimageimp(&mask, xref, rdb, obj, nil, 1);
if (error)
{
if (colorspace)
fz_dropcolorspace(colorspace);
return fz_rethrow(error, "cannot load image mask/softmask");
}
}
}
else if (fz_isarray(obj))
{
usecolorkey = 1;
for (i = 0; i < n * 2; i++)
colorkey[i] = fz_toint(fz_arrayget(obj, i));
}
stride = (w * n * bpc + 7) / 8;
samples = fz_calloc(h, stride);
if (cstm)
{
stm = pdf_openinlinestream(cstm, xref, dict, stride * h);
}
else
{
error = pdf_openstream(&stm, xref, fz_tonum(dict), fz_togen(dict));
if (error)
{
if (colorspace)
fz_dropcolorspace(colorspace);
if (mask)
fz_droppixmap(mask);
return fz_rethrow(error, "cannot open image data stream (%d 0 R)", fz_tonum(dict));
}
}
len = fz_read(stm, samples, h * stride);
if (len < 0)
{
fz_close(stm);
if (colorspace)
fz_dropcolorspace(colorspace);
if (mask)
fz_droppixmap(mask);
return fz_rethrow(len, "cannot read image data");
}
/* Make sure we read the EOF marker (for inline images only) */
if (cstm)
{
unsigned char tbuf[512];
int tlen = fz_read(stm, tbuf, sizeof tbuf);
if (tlen < 0)
fz_catch(tlen, "ignoring error at end of image");
if (tlen > 0)
fz_warn("ignoring garbage at end of image");
}
fz_close(stm);
/* Pad truncated images */
if (len < stride * h)
{
fz_warn("padding truncated image (%d 0 R)", fz_tonum(dict));
memset(samples + len, 0, stride * h - len);
}
/* Invert 1-bit image masks */
if (imagemask)
{
/* 0=opaque and 1=transparent so we need to invert */
unsigned char *p = samples;
len = h * stride;
for (i = 0; i < len; i++)
p[i] = ~p[i];
}
pdf_logimage("size %dx%d n=%d bpc=%d imagemask=%d indexed=%d\n", w, h, n, bpc, imagemask, indexed);
/* Unpack samples into pixmap */
tile = fz_newpixmap(colorspace, 0, 0, w, h);
scale = 1;
if (!indexed)
{
switch (bpc)
{
case 1: scale = 255; break;
case 2: scale = 85; break;
case 4: scale = 17; break;
}
}
fz_unpacktile(tile, samples, n, bpc, stride, scale);
if (usecolorkey)
pdf_maskcolorkey(tile, n, colorkey);
if (indexed)
{
fz_pixmap *conv;
fz_decodeindexedtile(tile, decode, (1 << bpc) - 1);
conv = pdf_expandindexedpixmap(tile);
fz_droppixmap(tile);
tile = conv;
}
else
{
fz_decodetile(tile, decode);
}
if (colorspace)
fz_dropcolorspace(colorspace);
tile->mask = mask;
tile->interpolate = interpolate;
fz_free(samples);
*imgp = tile;
return fz_okay;
}
static fz_error
pdf_loadjpximage(fz_pixmap **imgp, pdf_xref *xref, fz_obj *dict)
{
fz_error error;
fz_buffer *buf;
fz_pixmap *img;
fz_obj *obj;
pdf_logimage("jpeg2000\n");
error = pdf_loadstream(&buf, xref, fz_tonum(dict), fz_togen(dict));
if (error)
return fz_rethrow(error, "cannot load jpx image data");
error = fz_loadjpximage(&img, buf->data, buf->len);
if (error)
{
fz_dropbuffer(buf);
return fz_rethrow(error, "cannot load jpx image");
}
fz_dropbuffer(buf);
obj = fz_dictgetsa(dict, "SMask", "Mask");
if (fz_isdict(obj))
{
error = pdf_loadimageimp(&img->mask, xref, nil, obj, nil, 1);
if (error)
{
fz_droppixmap(img);
return fz_rethrow(error, "cannot load image mask/softmask");
}
}
obj = fz_dictgets(dict, "ColorSpace");
if (obj)
{
fz_colorspace *original = img->colorspace;
img->colorspace = nil;
error = pdf_loadcolorspace(&img->colorspace, xref, obj);
if (error)
{
fz_dropcolorspace(original);
return fz_rethrow(error, "cannot load image colorspace");
}
if (original->n != img->colorspace->n)
{
fz_warn("jpeg-2000 colorspace (%s) does not match promised colorspace (%s)", original->name, img->colorspace->name);
fz_dropcolorspace(img->colorspace);
img->colorspace = original;
}
else
fz_dropcolorspace(original);
if (!strcmp(img->colorspace->name, "Indexed"))
{
fz_pixmap *conv;
conv = pdf_expandindexedpixmap(img);
fz_droppixmap(img);
img = conv;
}
}
*imgp = img;
return fz_okay;
}
static fz_error
gatherfonts(int page, fz_obj *pageobj, fz_obj *dict)
{
int i;
for (i = 0; i < fz_dictlen(dict); i++)
{
fz_obj *ref;
fz_obj *fontdict;
fz_obj *subtype;
fz_obj *basefont;
fz_obj *name;
int k;
fontdict = ref = fz_dictgetval(dict, i);
if (!fz_isdict(fontdict))
{
fz_warn("not a font dict (%d %d R)", fz_tonum(ref), fz_togen(ref));
continue;
}
subtype = fz_dictgets(fontdict, "Subtype");
if (!fz_isname(subtype))
fz_warn("not a font dict subtype (%d %d R)", fz_tonum(ref), fz_togen(ref));
basefont = fz_dictgets(fontdict, "BaseFont");
if (basefont)
{
if (!fz_isname(basefont))
return fz_throw("not a font dict basefont (%d %d R)", fz_tonum(ref), fz_togen(ref));
}
else
{
name = fz_dictgets(fontdict, "Name");
if (name && !fz_isname(name))
return fz_throw("not a font dict name (%d %d R)", fz_tonum(ref), fz_togen(ref));
}
for (k = 0; k < fonts; k++)
if (fz_tonum(font[k]->ref) == fz_tonum(ref) &&
fz_togen(font[k]->ref) == fz_togen(ref))
break;
if (k < fonts)
continue;
fonts++;
font = fz_realloc(font, fonts * sizeof (struct info *));
if (!font)
return fz_throw("out of memory");
font[fonts - 1] = fz_malloc(sizeof (struct info));
if (!font[fonts - 1])
return fz_throw("out of memory");
font[fonts - 1]->page = page;
font[fonts - 1]->pageobj = pageobj;
font[fonts - 1]->ref = ref;
font[fonts - 1]->u.font.subtype = subtype;
font[fonts - 1]->u.font.name = basefont ? basefont : name;
}
return fz_okay;
}
/*
* Create a filter given a name and param dictionary.
*/
static fz_error *
buildonefilter(fz_filter **fp, fz_obj *f, fz_obj *p)
{
fz_filter *decompress;
fz_filter *predict;
fz_error *error;
char *s;
s = fz_toname(f);
if (!strcmp(s, "ASCIIHexDecode") || !strcmp(s, "AHx"))
error = fz_newahxd(fp, p);
else if (!strcmp(s, "ASCII85Decode") || !strcmp(s, "A85"))
error = fz_newa85d(fp, p);
else if (!strcmp(s, "CCITTFaxDecode") || !strcmp(s, "CCF"))
error = fz_newfaxd(fp, p);
else if (!strcmp(s, "DCTDecode") || !strcmp(s, "DCT"))
error = fz_newdctd(fp, p);
else if (!strcmp(s, "RunLengthDecode") || !strcmp(s, "RL"))
error = fz_newrld(fp, p);
else if (!strcmp(s, "FlateDecode") || !strcmp(s, "Fl"))
{
if (fz_isdict(p))
{
fz_obj *obj = fz_dictgets(p, "Predictor");
if (obj)
{
error = fz_newflated(&decompress, p);
if (error)
return fz_rethrow(error, "cannot create deflate filter");
error = fz_newpredictd(&predict, p);
if (error)
{
fz_dropfilter(decompress);
return fz_rethrow(error, "cannot create predictor filter");
}
error = fz_newpipeline(fp, decompress, predict);
fz_dropfilter(decompress);
fz_dropfilter(predict);
if (error)
return fz_rethrow(error, "cannot create pipeline filter");
return fz_okay;
}
}
error = fz_newflated(fp, p);
}
else if (!strcmp(s, "LZWDecode") || !strcmp(s, "LZW"))
{
if (fz_isdict(p))
{
fz_obj *obj = fz_dictgets(p, "Predictor");
if (obj)
{
error = fz_newlzwd(&decompress, p);
if (error)
return fz_rethrow(error, "cannot create lzwd filter");
error = fz_newpredictd(&predict, p);
if (error)
{
fz_dropfilter(decompress);
return fz_rethrow(error, "cannot create predictor filter");
}
error = fz_newpipeline(fp, decompress, predict);
fz_dropfilter(decompress);
fz_dropfilter(predict);
if (error)
return fz_rethrow(error, "cannot create pipeline filter");
return fz_okay;
}
}
error = fz_newlzwd(fp, p);
}
#ifdef HAVE_JBIG2DEC
else if (!strcmp(s, "JBIG2Decode"))
{
/* TODO: extract and feed JBIG2Global */
error = fz_newjbig2d(fp, p);
}
#endif
#ifdef HAVE_JASPER
else if (!strcmp(s, "JPXDecode"))
error = fz_newjpxd(fp, p);
#endif
else
{
return fz_throw("unknown filter name (%s)", s);
}
if (error)
return fz_rethrow(error, "cannot create filter");
return fz_okay;
}
static fz_error
gatherforms(int page, fz_obj *pageobj, fz_obj *dict)
{
int i;
for (i = 0; i < fz_dictlen(dict); i++)
{
fz_obj *ref;
fz_obj *xobjdict;
fz_obj *type;
fz_obj *subtype;
fz_obj *group;
fz_obj *reference;
int k;
xobjdict = ref = fz_dictgetval(dict, i);
if (!fz_isdict(xobjdict))
return fz_throw("not a xobject dict (%d %d R)", fz_tonum(ref), fz_togen(ref));
type = fz_dictgets(xobjdict, "Subtype");
if (!fz_isname(type))
return fz_throw("not a xobject type (%d %d R)", fz_tonum(ref), fz_togen(ref));
if (strcmp(fz_toname(type), "Form"))
continue;
subtype = fz_dictgets(xobjdict, "Subtype2");
if (subtype && !fz_isname(subtype))
return fz_throw("not a xobject subtype (%d %d R)", fz_tonum(ref), fz_togen(ref));
if (strcmp(fz_toname(subtype), "PS"))
continue;
group = fz_dictgets(xobjdict, "Group");
if (group && !fz_isdict(group))
return fz_throw("not a form xobject group dict (%d %d R)", fz_tonum(ref), fz_togen(ref));
reference = fz_dictgets(xobjdict, "Ref");
if (reference && !fz_isdict(reference))
return fz_throw("not a form xobject reference dict (%d %d R)", fz_tonum(ref), fz_togen(ref));
for (k = 0; k < forms; k++)
if (fz_tonum(form[k]->ref) == fz_tonum(ref) &&
fz_togen(form[k]->ref) == fz_togen(ref))
break;
if (k < forms)
continue;
forms++;
form = fz_realloc(form, forms * sizeof (struct info *));
if (!form)
return fz_throw("out of memory");
form[forms - 1] = fz_malloc(sizeof (struct info));
if (!form[forms - 1])
return fz_throw("out of memory");
form[forms - 1]->page = page;
form[forms - 1]->pageobj = pageobj;
form[forms - 1]->ref = ref;
form[forms - 1]->u.form.group = group;
form[forms - 1]->u.form.reference = reference;
}
return fz_okay;
}
