static fz_error
loadtype0(pdf_fontdesc **fontdescp, pdf_xref *xref, fz_obj *dict)
{
fz_error error;
fz_obj *dfonts;
fz_obj *dfont;
fz_obj *subtype;
fz_obj *encoding;
fz_obj *tounicode;
dfonts = fz_dictgets(dict, "DescendantFonts");
if (!dfonts)
return fz_throw("cid font is missing descendant fonts");
dfont = fz_arrayget(dfonts, 0);
subtype = fz_dictgets(dfont, "Subtype");
encoding = fz_dictgets(dict, "Encoding");
tounicode = fz_dictgets(dict, "ToUnicode");
if (fz_isname(subtype) && !strcmp(fz_toname(subtype), "CIDFontType0"))
error = loadcidfont(fontdescp, xref, dfont, encoding, tounicode);
else if (fz_isname(subtype) && !strcmp(fz_toname(subtype), "CIDFontType2"))
error = loadcidfont(fontdescp, xref, dfont, encoding, tounicode);
else
error = fz_throw("syntaxerror: unknown cid font type");
if (error)
return fz_rethrow(error, "cannot load descendant font (%d %d R)", fz_tonum(dfont), fz_togen(dfont));
return fz_okay;
}
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;
}
void
pdf_loadannots(pdf_comment **cp, pdf_link **lp, pdf_xref *xref, fz_obj *annots)
{
pdf_comment *comment;
pdf_link *link;
fz_obj *subtype;
fz_obj *obj;
int i;
comment = nil;
link = nil;
pdf_logpage("load annotations {\n");
for (i = 0; i < fz_arraylen(annots); i++)
{
obj = fz_arrayget(annots, i);
subtype = fz_dictgets(obj, "Subtype");
if (fz_isname(subtype) && !strcmp(fz_toname(subtype), "Link"))
{
pdf_link *temp = pdf_loadlink(xref, obj);
if (temp)
{
temp->next = link;
link = temp;
}
}
}
pdf_logpage("}\n");
*cp = comment;
*lp = link;
}
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;
}
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 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 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);
}
fz_error
pdf_loadannots(pdf_comment **cp, pdf_link **lp, pdf_xref *xref, fz_obj *annots)
{
fz_error error;
pdf_comment *comment;
pdf_link *link;
fz_obj *subtype;
fz_obj *obj;
int i;
comment = nil;
link = nil;
pdf_logpage("load annotations {\n");
for (i = 0; i < fz_arraylen(annots); i++)
{
obj = fz_arrayget(annots, i);
subtype = fz_dictgets(obj, "Subtype");
if (fz_isname(subtype) && !strcmp(fz_toname(subtype), "Link"))
{
pdf_link *temp = nil;
error = pdf_loadlink(&temp, xref, obj);
if (error)
{
if (link)
pdf_droplink(link);
return fz_rethrow(error, "cannot load annotation link");
}
if (temp)
{
temp->next = link;
link = temp;
}
}
else
{
error = loadcomment(&comment, xref, obj);
if (error)
{
if (link)
pdf_droplink(link);
return fz_rethrow(error, "cannot load annotation comment");
}
}
}
pdf_logpage("}\n");
*cp = comment;
*lp = link;
return fz_okay;
}
static int
pdf_extgstateusesblending(fz_obj *dict)
{
fz_obj *obj;
obj = fz_dictgets(dict, "BM");
if (fz_isname(obj) && strcmp(fz_toname(obj), "Normal"))
return 1;
return 0;
}
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);
}
/*
* Construct a filter to decode a stream, without
* constraining to stream length, and without decryption.
*/
fz_error *
pdf_buildinlinefilter(fz_filter **filterp, fz_obj *stmobj)
{
fz_error *error;
fz_obj *filters;
fz_obj *params;
filters = fz_dictgetsa(stmobj, "Filter", "F");
params = fz_dictgetsa(stmobj, "DecodeParms", "DP");
if (filters)
{
if (fz_isname(filters))
error = buildonefilter(filterp, filters, params);
else
error = buildfilterchain(filterp, nil, filters, params);
}
else
error = fz_newnullfilter(filterp, -1);
if (error)
return fz_rethrow(error, "cannot create inline filter chain");
return fz_okay;
}
pdf_link *
pdf_loadlink(pdf_xref *xref, fz_obj *dict)
{
fz_obj *dest;
fz_obj *action;
fz_obj *obj;
fz_rect bbox;
pdf_linkkind kind;
pdf_logpage("load link {\n");
dest = nil;
obj = fz_dictgets(dict, "Rect");
if (obj)
{
bbox = pdf_torect(obj);
pdf_logpage("rect [%g %g %g %g]\n",
bbox.x0, bbox.y0,
bbox.x1, bbox.y1);
}
else
bbox = fz_emptyrect;
obj = fz_dictgets(dict, "Dest");
if (obj)
{
kind = PDF_LGOTO;
dest = resolvedest(xref, obj);
pdf_logpage("dest (%d %d R)\n", fz_tonum(dest), fz_togen(dest));
}
action = fz_dictgets(dict, "A");
if (action)
{
obj = fz_dictgets(action, "S");
if (fz_isname(obj) && !strcmp(fz_toname(obj), "GoTo"))
{
kind = PDF_LGOTO;
dest = resolvedest(xref, fz_dictgets(action, "D"));
pdf_logpage("action goto (%d %d R)\n", fz_tonum(dest), fz_togen(dest));
}
else if (fz_isname(obj) && !strcmp(fz_toname(obj), "URI"))
{
kind = PDF_LURI;
dest = fz_dictgets(action, "URI");
pdf_logpage("action uri %s\n", fz_tostrbuf(dest));
}
/* cf. http://code.google.com/p/sumatrapdf/issues/detail?id=726 */
else if (fz_isname(obj) && !strcmp(fz_toname(obj), "Launch"))
{
kind = PDF_LLAUNCH;
dest = fz_dictgets(action, "F");
pdf_logpage("action launch (%d %d R)\n", fz_tonum(dest), fz_togen(dest));
}
else
{
pdf_logpage("unhandled link action, ignoring link\n");
dest = nil;
}
}
pdf_logpage("}\n");
if (dest)
{
return pdf_newlink(kind, bbox, dest);
}
return nil;
}
fz_error
pdf_loadxobject(pdf_xobject **formp, pdf_xref *xref, fz_obj *dict)
{
fz_error error;
pdf_xobject *form;
fz_obj *obj;
if ((*formp = pdf_finditem(xref->store, PDF_KXOBJECT, dict)))
{
pdf_keepxobject(*formp);
return fz_okay;
}
form = fz_malloc(sizeof(pdf_xobject));
form->refs = 1;
form->resources = nil;
form->contents = nil;
/* Store item immediately, to avoid possible recursion if objects refer back to this one */
pdf_storeitem(xref->store, PDF_KXOBJECT, dict, form);
pdf_logrsrc("load xobject (%d %d R) ptr=%p {\n", fz_tonum(dict), fz_togen(dict), form);
obj = fz_dictgets(dict, "BBox");
form->bbox = pdf_torect(obj);
pdf_logrsrc("bbox [%g %g %g %g]\n",
form->bbox.x0, form->bbox.y0,
form->bbox.x1, form->bbox.y1);
obj = fz_dictgets(dict, "Matrix");
if (obj)
form->matrix = pdf_tomatrix(obj);
else
form->matrix = fz_identity();
pdf_logrsrc("matrix [%g %g %g %g %g %g]\n",
form->matrix.a, form->matrix.b,
form->matrix.c, form->matrix.d,
form->matrix.e, form->matrix.f);
form->isolated = 0;
form->knockout = 0;
form->transparency = 0;
obj = fz_dictgets(dict, "Group");
if (obj)
{
fz_obj *attrs = obj;
form->isolated = fz_tobool(fz_dictgets(attrs, "I"));
form->knockout = fz_tobool(fz_dictgets(attrs, "K"));
obj = fz_dictgets(attrs, "S");
if (fz_isname(obj) && !strcmp(fz_toname(obj), "Transparency"))
form->transparency = 1;
}
pdf_logrsrc("isolated %d\n", form->isolated);
pdf_logrsrc("knockout %d\n", form->knockout);
pdf_logrsrc("transparency %d\n", form->transparency);
form->resources = fz_dictgets(dict, "Resources");
if (form->resources)
fz_keepobj(form->resources);
error = pdf_loadstream(&form->contents, xref, fz_tonum(dict), fz_togen(dict));
if (error)
{
pdf_removeitem(xref->store, PDF_KXOBJECT, dict);
pdf_dropxobject(form);
return fz_rethrow(error, "cannot load xobject content stream (%d %d R)", fz_tonum(dict), fz_togen(dict));
}
pdf_logrsrc("stream %d bytes\n", form->contents->wp - form->contents->rp);
pdf_logrsrc("}\n");
*formp = form;
return fz_okay;
}
/*
* Load CMap stream in PDF file
*/
fz_error
pdf_loadembeddedcmap(pdf_cmap **cmapp, pdf_xref *xref, fz_obj *stmref)
{
fz_error error = fz_okay;
fz_obj *stmobj;
fz_stream *file = nil;
pdf_cmap *cmap = nil;
pdf_cmap *usecmap;
fz_obj *wmode;
fz_obj *obj;
if ((*cmapp = pdf_finditem(xref->store, PDF_KCMAP, stmref)))
{
pdf_keepcmap(*cmapp);
return fz_okay;
}
pdf_logfont("load embedded cmap (%d %d R) {\n", fz_tonum(stmref), fz_togen(stmref));
stmobj = fz_resolveindirect(stmref);
error = pdf_openstream(&file, xref, fz_tonum(stmref), fz_togen(stmref));
if (error)
{
error = fz_rethrow(error, "cannot open cmap stream");
goto cleanup;
}
error = pdf_parsecmap(&cmap, file);
if (error)
{
error = fz_rethrow(error, "cannot parse cmap stream");
goto cleanup;
}
fz_dropstream(file);
wmode = fz_dictgets(stmobj, "WMode");
if (fz_isint(wmode))
{
pdf_logfont("wmode %d\n", wmode);
pdf_setwmode(cmap, fz_toint(wmode));
}
obj = fz_dictgets(stmobj, "UseCMap");
if (fz_isname(obj))
{
pdf_logfont("usecmap /%s\n", fz_toname(obj));
error = pdf_loadsystemcmap(&usecmap, fz_toname(obj));
if (error)
{
error = fz_rethrow(error, "cannot load system usecmap '%s'", fz_toname(obj));
goto cleanup;
}
pdf_setusecmap(cmap, usecmap);
pdf_dropcmap(usecmap);
}
else if (fz_isindirect(obj))
{
pdf_logfont("usecmap (%d %d R)\n", fz_tonum(obj), fz_togen(obj));
error = pdf_loadembeddedcmap(&usecmap, xref, obj);
if (error)
{
error = fz_rethrow(error, "cannot load embedded usecmap");
goto cleanup;
}
pdf_setusecmap(cmap, usecmap);
pdf_dropcmap(usecmap);
}
pdf_logfont("}\n");
error = pdf_storeitem(xref->store, PDF_KCMAP, stmref, cmap);
if (error)
{
error = fz_rethrow(error, "cannot store cmap resource");
goto cleanup;
}
*cmapp = cmap;
return fz_okay;
cleanup:
if (file)
fz_dropstream(file);
if (cmap)
pdf_dropcmap(cmap);
return error; /* already rethrown */
}
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");
}
static fz_error
loadcidfont(pdf_fontdesc **fontdescp, pdf_xref *xref, fz_obj *dict, fz_obj *encoding, fz_obj *tounicode)
{
fz_error error;
fz_obj *widths;
fz_obj *descriptor;
pdf_fontdesc *fontdesc;
FT_Face face;
fz_bbox bbox;
int kind;
char collection[256];
char *basefont;
int i, k, fterr;
fz_obj *obj;
int dw;
/* Get font name and CID collection */
basefont = fz_toname(fz_dictgets(dict, "BaseFont"));
{
fz_obj *cidinfo;
char tmpstr[64];
int tmplen;
cidinfo = fz_dictgets(dict, "CIDSystemInfo");
if (!cidinfo)
return fz_throw("cid font is missing info");
obj = fz_dictgets(cidinfo, "Registry");
tmplen = MIN(sizeof tmpstr - 1, fz_tostrlen(obj));
memcpy(tmpstr, fz_tostrbuf(obj), tmplen);
tmpstr[tmplen] = '\0';
fz_strlcpy(collection, tmpstr, sizeof collection);
fz_strlcat(collection, "-", sizeof collection);
obj = fz_dictgets(cidinfo, "Ordering");
tmplen = MIN(sizeof tmpstr - 1, fz_tostrlen(obj));
memcpy(tmpstr, fz_tostrbuf(obj), tmplen);
tmpstr[tmplen] = '\0';
fz_strlcat(collection, tmpstr, sizeof collection);
}
/* Load font file */
fontdesc = pdf_newfontdesc();
pdf_logfont("load cid font (%d %d R) ptr=%p {\n", fz_tonum(dict), fz_togen(dict), fontdesc);
pdf_logfont("basefont %s\n", basefont);
pdf_logfont("collection %s\n", collection);
descriptor = fz_dictgets(dict, "FontDescriptor");
if (descriptor)
error = pdf_loadfontdescriptor(fontdesc, xref, descriptor, collection, basefont);
else
error = fz_throw("syntaxerror: missing font descriptor");
if (error)
goto cleanup;
face = fontdesc->font->ftface;
kind = ftkind(face);
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 */
error = fz_okay;
if (fz_isname(encoding))
{
pdf_logfont("encoding /%s\n", fz_toname(encoding));
if (!strcmp(fz_toname(encoding), "Identity-H"))
fontdesc->encoding = pdf_newidentitycmap(0, 2);
else if (!strcmp(fz_toname(encoding), "Identity-V"))
fontdesc->encoding = pdf_newidentitycmap(1, 2);
else
error = pdf_loadsystemcmap(&fontdesc->encoding, fz_toname(encoding));
}
else if (fz_isindirect(encoding))
{
pdf_logfont("encoding %d %d R\n", fz_tonum(encoding), fz_togen(encoding));
error = pdf_loadembeddedcmap(&fontdesc->encoding, xref, encoding);
}
else
{
error = fz_throw("syntaxerror: font missing encoding");
}
if (error)
goto cleanup;
pdf_setfontwmode(fontdesc, pdf_getwmode(fontdesc->encoding));
pdf_logfont("wmode %d\n", pdf_getwmode(fontdesc->encoding));
if (kind == TRUETYPE)
{
fz_obj *cidtogidmap;
cidtogidmap = fz_dictgets(dict, "CIDToGIDMap");
if (fz_isindirect(cidtogidmap))
{
fz_buffer *buf;
pdf_logfont("cidtogidmap stream\n");
error = pdf_loadstream(&buf, xref, fz_tonum(cidtogidmap), fz_togen(cidtogidmap));
if (error)
goto cleanup;
fontdesc->ncidtogid = (buf->len) / 2;
fontdesc->cidtogid = fz_malloc(fontdesc->ncidtogid * sizeof(unsigned short));
for (i = 0; i < fontdesc->ncidtogid; i++)
fontdesc->cidtogid[i] = (buf->data[i * 2] << 8) + buf->data[i * 2 + 1];
fz_dropbuffer(buf);
}
/* if truetype font is external, cidtogidmap should not be identity */
/* so we map from cid to unicode and then map that through the (3 1) */
/* unicode cmap to get a glyph id */
else if (fontdesc->font->ftsubstitute)
{
pdf_logfont("emulate ttf cidfont\n");
fterr = FT_Select_Charmap(face, ft_encoding_unicode);
if (fterr)
{
error = fz_throw("fonterror: no unicode cmap when emulating CID font: %s", ft_errorstring(fterr));
goto cleanup;
}
if (!strcmp(collection, "Adobe-CNS1"))
error = pdf_loadsystemcmap(&fontdesc->tottfcmap, "Adobe-CNS1-UCS2");
else if (!strcmp(collection, "Adobe-GB1"))
error = pdf_loadsystemcmap(&fontdesc->tottfcmap, "Adobe-GB1-UCS2");
else if (!strcmp(collection, "Adobe-Japan1"))
error = pdf_loadsystemcmap(&fontdesc->tottfcmap, "Adobe-Japan1-UCS2");
else if (!strcmp(collection, "Adobe-Japan2"))
error = pdf_loadsystemcmap(&fontdesc->tottfcmap, "Adobe-Japan2-UCS2");
else if (!strcmp(collection, "Adobe-Korea1"))
error = pdf_loadsystemcmap(&fontdesc->tottfcmap, "Adobe-Korea1-UCS2");
else
error = fz_okay;
if (error)
{
error = fz_rethrow(error, "cannot load system cmap %s", collection);
goto cleanup;
}
}
}
error = pdf_loadtounicode(fontdesc, xref, nil, collection, tounicode);
if (error)
goto cleanup;
/* Horizontal */
dw = 1000;
obj = fz_dictgets(dict, "DW");
if (obj)
dw = fz_toint(obj);
pdf_setdefaulthmtx(fontdesc, dw);
widths = fz_dictgets(dict, "W");
if (widths)
{
int c0, c1, w;
for (i = 0; i < fz_arraylen(widths); )
{
c0 = fz_toint(fz_arrayget(widths, i));
obj = fz_arrayget(widths, i + 1);
if (fz_isarray(obj))
{
for (k = 0; k < fz_arraylen(obj); k++)
{
w = fz_toint(fz_arrayget(obj, k));
pdf_addhmtx(fontdesc, c0 + k, c0 + k, w);
}
i += 2;
}
else
{
c1 = fz_toint(obj);
w = fz_toint(fz_arrayget(widths, i + 2));
pdf_addhmtx(fontdesc, c0, c1, w);
i += 3;
}
}
}
pdf_endhmtx(fontdesc);
/* Vertical */
if (pdf_getwmode(fontdesc->encoding) == 1)
{
int dw2y = 880;
int dw2w = -1000;
obj = fz_dictgets(dict, "DW2");
if (obj)
{
dw2y = fz_toint(fz_arrayget(obj, 0));
dw2w = fz_toint(fz_arrayget(obj, 1));
}
pdf_setdefaultvmtx(fontdesc, dw2y, dw2w);
widths = fz_dictgets(dict, "W2");
if (widths)
{
int c0, c1, w, x, y;
for (i = 0; i < fz_arraylen(widths); )
{
c0 = fz_toint(fz_arrayget(widths, i));
obj = fz_arrayget(widths, i + 1);
if (fz_isarray(obj))
{
for (k = 0; k < fz_arraylen(obj); k += 3)
{
w = fz_toint(fz_arrayget(obj, k + 0));
x = fz_toint(fz_arrayget(obj, k + 1));
y = fz_toint(fz_arrayget(obj, k + 2));
pdf_addvmtx(fontdesc, c0 + k, c0 + k, x, y, w);
}
i += 2;
}
else
{
c1 = fz_toint(obj);
w = fz_toint(fz_arrayget(widths, i + 2));
x = fz_toint(fz_arrayget(widths, i + 3));
y = fz_toint(fz_arrayget(widths, i + 4));
pdf_addvmtx(fontdesc, c0, c1, x, y, w);
i += 5;
}
}
}
pdf_endvmtx(fontdesc);
}
pdf_logfont("}\n");
*fontdescp = fontdesc;
return fz_okay;
cleanup:
pdf_dropfont(fontdesc);
return fz_rethrow(error, "cannot load cid font (%d %d R)", fz_tonum(dict), fz_togen(dict));
}
static fz_error
fz_repairobj(fz_stream *file, char *buf, int cap, int *stmofsp, int *stmlenp, fz_obj **encrypt, fz_obj **id)
{
fz_error error;
int tok;
int stmlen;
int len;
int n;
*stmofsp = 0;
*stmlenp = -1;
stmlen = 0;
error = pdf_lex(&tok, file, buf, cap, &len);
if (error)
return fz_rethrow(error, "cannot parse object");
if (tok == PDF_TODICT)
{
fz_obj *dict, *obj;
/* Send nil xref so we don't try to resolve references */
error = pdf_parsedict(&dict, nil, file, buf, cap);
if (error)
return fz_rethrow(error, "cannot parse object");
obj = fz_dictgets(dict, "Type");
if (fz_isname(obj) && !strcmp(fz_toname(obj), "XRef"))
{
obj = fz_dictgets(dict, "Encrypt");
if (obj)
{
if (*encrypt)
fz_dropobj(*encrypt);
*encrypt = fz_keepobj(obj);
}
obj = fz_dictgets(dict, "ID");
if (obj)
{
if (*id)
fz_dropobj(*id);
*id = fz_keepobj(obj);
}
}
obj = fz_dictgets(dict, "Length");
if (fz_isint(obj))
stmlen = fz_toint(obj);
fz_dropobj(dict);
}
while ( tok != PDF_TSTREAM &&
tok != PDF_TENDOBJ &&
tok != PDF_TERROR &&
tok != PDF_TEOF )
{
error = pdf_lex(&tok, file, buf, cap, &len);
if (error)
return fz_rethrow(error, "cannot scan for endobj or stream token");
}
if (tok == PDF_TSTREAM)
{
int c = fz_readbyte(file);
if (c == '\r') {
c = fz_peekbyte(file);
if (c == '\n')
fz_readbyte(file);
}
*stmofsp = fz_tell(file);
if (*stmofsp < 0)
return fz_throw("cannot seek in file");
if (stmlen > 0)
{
fz_seek(file, *stmofsp + stmlen, 0);
error = pdf_lex(&tok, file, buf, cap, &len);
if (error)
fz_catch(error, "cannot find endstream token, falling back to scanning");
if (tok == PDF_TENDSTREAM)
goto atobjend;
fz_seek(file, *stmofsp, 0);
}
n = fz_read(file, (unsigned char *) buf, 9);
if (n < 0)
return fz_rethrow(n, "cannot read from file");
while (memcmp(buf, "endstream", 9) != 0)
{
c = fz_readbyte(file);
if (c == EOF)
break;
memmove(buf, buf + 1, 8);
buf[8] = c;
}
*stmlenp = fz_tell(file) - *stmofsp - 9;
atobjend:
error = pdf_lex(&tok, file, buf, cap, &len);
if (error)
return fz_rethrow(error, "cannot scan for endobj token");
if (tok != PDF_TENDOBJ)
fz_warn("object missing 'endobj' token");
}
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;
}
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;
}
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;
}
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_loadcolorspaceimp(fz_colorspace **csp, pdf_xref *xref, fz_obj *obj)
{
if (fz_isname(obj))
{
if (!strcmp(fz_toname(obj), "DeviceGray"))
*csp = pdf_devicegray;
else if (!strcmp(fz_toname(obj), "DeviceRGB"))
*csp = pdf_devicergb;
else if (!strcmp(fz_toname(obj), "DeviceCMYK"))
*csp = pdf_devicecmyk;
else if (!strcmp(fz_toname(obj), "G"))
*csp = pdf_devicegray;
else if (!strcmp(fz_toname(obj), "RGB"))
*csp = pdf_devicergb;
else if (!strcmp(fz_toname(obj), "CMYK"))
*csp = pdf_devicecmyk;
else if (!strcmp(fz_toname(obj), "Pattern"))
*csp = pdf_devicepattern;
else
return fz_throw("unknown colorspace: %s", fz_toname(obj));
return fz_okay;
}
else if (fz_isarray(obj))
{
fz_obj *name = fz_arrayget(obj, 0);
if (fz_isname(name))
{
if (!strcmp(fz_toname(name), "CalCMYK"))
*csp = pdf_devicecmyk;
#ifdef USECAL
else if (!strcmp(fz_toname(name), "CalGray"))
*csp = loadcalgray(xref, fz_arrayget(obj, 1));
else if (!strcmp(fz_toname(name), "CalRGB"))
*csp = loadcalrgb(xref, fz_arrayget(obj, 1));
else if (!strcmp(fz_toname(name), "Lab"))
*csp = loadlab(xref, fz_arrayget(obj, 1));
#else
else if (!strcmp(fz_toname(name), "CalGray"))
*csp = pdf_devicegray;
else if (!strcmp(fz_toname(name), "CalRGB"))
*csp = pdf_devicergb;
else if (!strcmp(fz_toname(name), "Lab"))
*csp = pdf_devicelab;
#endif
else if (!strcmp(fz_toname(name), "ICCBased"))
return loadiccbased(csp, xref, fz_arrayget(obj, 1));
else if (!strcmp(fz_toname(name), "Indexed"))
return loadindexed(csp, xref, obj);
else if (!strcmp(fz_toname(name), "I"))
return loadindexed(csp, xref, obj);
else if (!strcmp(fz_toname(name), "Separation"))
return loadseparation(csp, xref, obj);
else if (!strcmp(fz_toname(name), "DeviceN"))
return loadseparation(csp, xref, obj);
/* load base colorspace instead */
else if (!strcmp(fz_toname(name), "Pattern"))
{
fz_error error;
obj = fz_arrayget(obj, 1);
if (!obj)
{
*csp = pdf_devicepattern;
return fz_okay;
}
error = pdf_loadcolorspace(csp, xref, obj);
if (error)
return fz_rethrow(error, "cannot load pattern");
}
else if (!strcmp(fz_toname(name), "DeviceGray"))
*csp = pdf_devicegray;
else if (!strcmp(fz_toname(name), "DeviceRGB"))
*csp = pdf_devicergb;
else if (!strcmp(fz_toname(name), "DeviceCMYK"))
*csp = pdf_devicecmyk;
else if (!strcmp(fz_toname(name), "G"))
*csp = pdf_devicegray;
else if (!strcmp(fz_toname(name), "RGB"))
*csp = pdf_devicergb;
else if (!strcmp(fz_toname(name), "CMYK"))
*csp = pdf_devicecmyk;
else
return fz_throw("syntaxerror: unknown colorspace %s", fz_toname(name));
return fz_okay;
}
}
return fz_throw("syntaxerror: could not parse color space");
}
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));
}
static void savefont(fz_obj *dict, int num, int gen)
{
fz_error error;
char name[1024];
char *subtype;
fz_buffer *buf;
fz_obj *stream = nil;
fz_obj *obj;
char *ext = "";
FILE *f;
unsigned char *p;
char *fontname = "font";
obj = fz_dictgets(dict, "FontName");
if (obj)
fontname = fz_toname(obj);
obj = fz_dictgets(dict, "FontFile");
if (obj)
{
stream = obj;
ext = "pfa";
}
obj = fz_dictgets(dict, "FontFile2");
if (obj)
{
stream = obj;
ext = "ttf";
}
obj = fz_dictgets(dict, "FontFile3");
if (obj)
{
stream = obj;
obj = fz_dictgets(obj, "Subtype");
if (obj && !fz_isname(obj))
die(fz_throw("Invalid font descriptor subtype"));
subtype = fz_toname(obj);
if (!strcmp(subtype, "Type1C"))
ext = "cff";
else if (!strcmp(subtype, "CIDFontType0C"))
ext = "cid";
else
die(fz_throw("Unhandled font type '%s'", subtype));
}
if (!stream)
{
fz_warn("Unhandled font type");
return;
}
error = fz_newbuffer(&buf, 0);
if (error)
die(error);
error = pdf_loadstream(&buf, xref, fz_tonum(stream), fz_togen(stream));
if (error)
die(error);
sprintf(name, "%s-%04d.%s", fontname, num, ext);
f = fopen(name, "wb");
if (f == NULL)
die(fz_throw("Error creating image file"));
for (p = buf->rp; p < buf->wp; p ++)
fprintf(f, "%c", *p);
if (fclose(f) < 0)
die(fz_throw("Error closing image file"));
fz_dropbuffer(buf);
}
fz_error *
pdf_loadinlineimage(pdf_image **imgp, pdf_xref *xref,
fz_obj *rdb, fz_obj *dict, fz_stream *file)
{
fz_error *error;
pdf_image *img;
fz_filter *filter;
fz_obj *f;
fz_obj *cs;
fz_obj *d;
int ismask;
int i;
img = fz_malloc(sizeof(pdf_image));
if (!img)
return fz_outofmem;
pdf_logimage("load inline image %p {\n", img);
img->super.loadtile = pdf_loadtile;
img->super.drop = pdf_dropimage;
img->super.n = 0;
img->super.a = 0;
img->super.refs = 0;
img->indexed = nil;
img->usecolorkey = 0;
img->mask = nil;
img->super.w = fz_toint(fz_dictgetsa(dict, "Width", "W"));
img->super.h = fz_toint(fz_dictgetsa(dict, "Height", "H"));
img->bpc = fz_toint(fz_dictgetsa(dict, "BitsPerComponent", "BPC"));
ismask = fz_tobool(fz_dictgetsa(dict, "ImageMask", "IM"));
d = fz_dictgetsa(dict, "Decode", "D");
cs = fz_dictgetsa(dict, "ColorSpace", "CS");
pdf_logimage("size %dx%d %d\n", img->super.w, img->super.h, img->bpc);
if (ismask)
{
pdf_logimage("is mask\n");
img->super.cs = nil;
img->super.n = 0;
img->super.a = 1;
img->bpc = 1;
}
if (cs)
{
img->super.cs = nil;
if (fz_isname(cs))
{
fz_obj *csd = fz_dictgets(rdb, "ColorSpace");
if (csd)
{
fz_obj *cso = fz_dictget(csd, cs);
img->super.cs = pdf_finditem(xref->store, PDF_KCOLORSPACE, cso);
if (img->super.cs)
fz_keepcolorspace(img->super.cs);
}
}
if (!img->super.cs)
{
/* XXX danger! danger! does this resolve? */
error = pdf_loadcolorspace(&img->super.cs, xref, cs);
if (error)
return error;
}
if (!strcmp(img->super.cs->name, "Indexed"))
{
pdf_logimage("indexed\n");
img->indexed = (pdf_indexed*)img->super.cs;
img->super.cs = img->indexed->base;
}
pdf_logimage("colorspace %s\n", img->super.cs->name);
img->super.n = img->super.cs->n;
img->super.a = 0;
}
if (fz_isarray(d))
{
pdf_logimage("decode array\n");
if (img->indexed)
for (i = 0; i < 2; i++)
img->decode[i] = fz_toreal(fz_arrayget(d, i));
else
for (i = 0; i < (img->super.n + img->super.a) * 2; i++)
img->decode[i] = fz_toreal(fz_arrayget(d, i));
}
else
{
if (img->indexed)
for (i = 0; i < 2; i++)
img->decode[i] = i & 1 ? (1 << img->bpc) - 1 : 0;
else
for (i = 0; i < (img->super.n + img->super.a) * 2; i++)
img->decode[i] = i & 1;
}
if (img->indexed)
img->stride = (img->super.w * img->bpc + 7) / 8;
else
img->stride = (img->super.w * (img->super.n + img->super.a) * img->bpc + 7) / 8;
/* load image data */
f = fz_dictgetsa(dict, "Filter", "F");
if (f)
{
fz_stream *tempfile;
error = pdf_buildinlinefilter(&filter, dict);
if (error)
return error;
error = fz_openrfilter(&tempfile, filter, file);
if (error)
return error;
i = fz_readall(&img->samples, tempfile);
if (i < 0)
return fz_ioerror(tempfile);
fz_dropfilter(filter);
fz_dropstream(tempfile);
}
else
{
error = fz_newbuffer(&img->samples, img->super.h * img->stride);
if (error)
return error;
i = fz_read(file, img->samples->bp, img->super.h * img->stride);
if (i < 0)
return fz_ioerror(file);
img->samples->wp += img->super.h * img->stride;
}
/* 0 means opaque and 1 means transparent, so we invert to get alpha */
if (ismask)
{
unsigned char *p;
for (p = img->samples->bp; p < img->samples->ep; p++)
*p = ~*p;
}
pdf_logimage("}\n");
*imgp = img;
return nil;
}
static int isfontdesc(fz_obj *obj)
{
fz_obj *type = fz_dictgets(obj, "Type");
return fz_isname(type) && !strcmp(fz_toname(type), "FontDescriptor");
}
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");
}
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");
}
/*
* Construct a filter to decode a stream, constraining
* to stream length and decrypting.
*/
static fz_error *
pdf_buildfilter(fz_filter **filterp, pdf_xref *xref, fz_obj *stmobj, int oid, int gen)
{
fz_error *error;
fz_filter *base, *pipe, *tmp;
fz_obj *filters;
fz_obj *params;
error = buildrawfilter(&base, xref, stmobj, oid, gen);
if (error)
return fz_rethrow(error, "cannot create raw filter chain");
filters = fz_dictgetsa(stmobj, "Filter", "F");
params = fz_dictgetsa(stmobj, "DecodeParms", "DP");
if (filters)
{
error = pdf_resolve(&filters, xref);
if (error)
{
error = fz_rethrow(error, "cannot resolve stream /Filter");
goto cleanup0;
}
if (params)
{
error = pdf_resolve(¶ms, xref);
if (error)
{
error = fz_rethrow(error, "cannot resolve stream /DecodeParms");
goto cleanup1;
}
}
if (fz_isname(filters))
{
error = buildonefilter(&tmp, filters, params);
if (error)
{
error = fz_rethrow(error, "cannot create filter");
goto cleanup2;
}
error = fz_newpipeline(&pipe, base, tmp);
fz_dropfilter(base);
fz_dropfilter(tmp);
if (error)
{
error = fz_rethrow(error, "cannot create filter pipeline");
goto cleanup2;
}
}
else
{
error = buildfilterchain(&pipe, base, filters, params);
if (error)
{
error = fz_rethrow(error, "cannot create filter chain");
goto cleanup2;
}
}
if (params)
fz_dropobj(params);
fz_dropobj(filters);
*filterp = pipe;
}
else
{
*filterp = base;
}
return fz_okay;
cleanup2:
if (params)
fz_dropobj(params);
cleanup1:
fz_dropobj(filters);
cleanup0:
fz_dropfilter(base);
return error; /* already rethrown */
}
static int isimage(fz_obj *obj)
{
fz_obj *type = fz_dictgets(obj, "Subtype");
return fz_isname(type) && !strcmp(fz_toname(type), "Image");
}
