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;
}
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 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));
}
/*
* 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 */
}
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
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_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_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");
}
void
pdf_loadpagetreenode(pdf_xref *xref, fz_obj *node, struct info info)
{
fz_obj *dict, *kids, *count;
fz_obj *obj, *tmp;
int i, n;
/* prevent infinite recursion */
if (fz_dictgets(node, ".seen"))
return;
kids = fz_dictgets(node, "Kids");
count = fz_dictgets(node, "Count");
if (fz_isarray(kids) && fz_isint(count))
{
obj = fz_dictgets(node, "Resources");
if (obj)
info.resources = obj;
obj = fz_dictgets(node, "MediaBox");
if (obj)
info.mediabox = obj;
obj = fz_dictgets(node, "CropBox");
if (obj)
info.cropbox = obj;
obj = fz_dictgets(node, "Rotate");
if (obj)
info.rotate = obj;
tmp = fz_newnull();
fz_dictputs(node, ".seen", tmp);
fz_dropobj(tmp);
n = fz_arraylen(kids);
for (i = 0; i < n; i++)
{
obj = fz_arrayget(kids, i);
pdf_loadpagetreenode(xref, obj, info);
}
fz_dictdels(node, ".seen");
}
else
{
dict = fz_resolveindirect(node);
if (info.resources && !fz_dictgets(dict, "Resources"))
fz_dictputs(dict, "Resources", info.resources);
if (info.mediabox && !fz_dictgets(dict, "MediaBox"))
fz_dictputs(dict, "MediaBox", info.mediabox);
if (info.cropbox && !fz_dictgets(dict, "CropBox"))
fz_dictputs(dict, "CropBox", info.cropbox);
if (info.rotate && !fz_dictgets(dict, "Rotate"))
fz_dictputs(dict, "Rotate", info.rotate);
if (xref->pagelen == xref->pagecap)
{
fz_warn("found more pages than expected");
xref->pagecap ++;
xref->pagerefs = fz_realloc(xref->pagerefs, sizeof(fz_obj*) * xref->pagecap);
xref->pageobjs = fz_realloc(xref->pageobjs, sizeof(fz_obj*) * xref->pagecap);
}
xref->pagerefs[xref->pagelen] = fz_keepobj(node);
xref->pageobjs[xref->pagelen] = fz_keepobj(dict);
xref->pagelen ++;
}
}
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;
}
