static fz_error pdf_loadxref(pdf_xref *xref, char *buf, int bufsize) { fz_error error; fz_obj *size; int i; error = pdf_loadversion(xref); if (error) return fz_rethrow(error, "cannot read version marker"); error = pdf_readstartxref(xref); if (error) return fz_rethrow(error, "cannot read startxref"); error = pdf_readtrailer(xref, buf, bufsize); if (error) return fz_rethrow(error, "cannot read trailer"); size = fz_dictgets(xref->trailer, "Size"); if (!size) return fz_throw("trailer missing Size entry"); pdf_logxref("\tsize %d at 0x%x\n", fz_toint(size), xref->startxref); xref->len = fz_toint(size); xref->cap = xref->len + 1; /* for hack to allow broken pdf generators with off-by-one errors */ xref->table = fz_malloc(xref->cap * sizeof(pdf_xrefentry)); for (i = 0; i < xref->cap; i++) { xref->table[i].ofs = 0; xref->table[i].gen = 0; xref->table[i].stmofs = 0; xref->table[i].obj = nil; xref->table[i].type = 0; } error = pdf_readxrefsections(xref, xref->startxref, buf, bufsize); if (error) return fz_rethrow(error, "cannot read xref"); /* broken pdfs where first object is not free */ if (xref->table[0].type != 'f') return fz_throw("first object in xref is not free"); /* broken pdfs where object offsets are out of range */ for (i = 0; i < xref->len; i++) if (xref->table[i].type == 'n') if (xref->table[i].ofs <= 0 || xref->table[i].ofs >= xref->filesize) return fz_throw("object offset out of range: %d", xref->table[i].ofs); return fz_okay; }
fz_filter * fz_newpredictd(fz_obj *params) { fz_obj *obj; FZ_NEWFILTER(fz_predict, p, predict); p->predictor = 1; p->columns = 1; p->colors = 1; p->bpc = 8; obj = fz_dictgets(params, "Predictor"); if (obj) p->predictor = fz_toint(obj); if (p->predictor != 1 && p->predictor != 2 && p->predictor != 10 && p->predictor != 11 && p->predictor != 12 && p->predictor != 13 && p->predictor != 14 && p->predictor != 15) { fz_warn("invalid predictor: %d", p->predictor); p->predictor = 1; } obj = fz_dictgets(params, "Columns"); if (obj) p->columns = fz_toint(obj); obj = fz_dictgets(params, "Colors"); if (obj) p->colors = fz_toint(obj); obj = fz_dictgets(params, "BitsPerComponent"); if (obj) p->bpc = fz_toint(obj); p->stride = (p->bpc * p->colors * p->columns + 7) / 8; p->bpp = (p->bpc * p->colors + 7) / 8; if (p->predictor >= 10) { p->ref = fz_malloc(p->stride); memset(p->ref, 0, p->stride); } else { p->ref = nil; } return (fz_filter*)p; }
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 * loadiccbased(fz_colorspace **csp, pdf_xref *xref, fz_obj *ref) { fz_error *error; fz_obj *dict; int n; pdf_logrsrc("load ICCBased\n"); error = pdf_loadindirect(&dict, xref, ref); if (error) return error; n = fz_toint(fz_dictgets(dict, "N")); fz_dropobj(dict); switch (n) { case 1: *csp = pdf_devicegray; return nil; case 3: *csp = pdf_devicergb; return nil; case 4: *csp = pdf_devicecmyk; return nil; } return fz_throw("syntaxerror: ICCBased must have 1, 3 or 4 components"); }
fz_error pdf_loadfontdescriptor(pdf_fontdesc *fontdesc, pdf_xref *xref, fz_obj *dict, char *collection) { fz_error error; fz_obj *obj1, *obj2, *obj3, *obj; fz_rect bbox; char *fontname; pdf_logfont("load fontdescriptor {\n"); fontname = fz_toname(fz_dictgets(dict, "FontName")); pdf_logfont("fontname '%s'\n", fontname); fontdesc->flags = fz_toint(fz_dictgets(dict, "Flags")); fontdesc->italicangle = fz_toreal(fz_dictgets(dict, "ItalicAngle")); fontdesc->ascent = fz_toreal(fz_dictgets(dict, "Ascent")); fontdesc->descent = fz_toreal(fz_dictgets(dict, "Descent")); fontdesc->capheight = fz_toreal(fz_dictgets(dict, "CapHeight")); fontdesc->xheight = fz_toreal(fz_dictgets(dict, "XHeight")); fontdesc->missingwidth = fz_toreal(fz_dictgets(dict, "MissingWidth")); bbox = pdf_torect(fz_dictgets(dict, "FontBBox")); pdf_logfont("bbox [%g %g %g %g]\n", bbox.x0, bbox.y0, bbox.x1, bbox.y1); pdf_logfont("flags %d\n", fontdesc->flags); obj1 = fz_dictgets(dict, "FontFile"); obj2 = fz_dictgets(dict, "FontFile2"); obj3 = fz_dictgets(dict, "FontFile3"); obj = obj1 ? obj1 : obj2 ? obj2 : obj3; if (getenv("NOFONT")) obj = nil; if (fz_isindirect(obj)) { error = pdf_loadembeddedfont(fontdesc, xref, obj); if (error) { fz_catch(error, "ignored error when loading embedded font, attempting to load system font"); error = pdf_loadsystemfont(fontdesc, fontname, collection); if (error) return fz_rethrow(error, "cannot load font descriptor"); } } else { error = pdf_loadsystemfont(fontdesc, fontname, collection); if (error) return fz_rethrow(error, "cannot load font descriptor"); } pdf_logfont("}\n"); return fz_okay; }
fz_error fz_newlzwe(fz_filter **fp, fz_obj *params) { FZ_NEWFILTER(fz_lzwe, lzw, lzwe); lzw->earlychange = 0; if (params) { fz_obj *obj; obj = fz_dictgets(params, "EarlyChange"); if (obj) lzw->earlychange = fz_toint(obj) != 0; } lzw->bidx = 0; lzw->bsave = 0; lzw->resume = 0; lzw->code = -1; lzw->hcode = -1; lzw->fcode = -1; lzw->codebits = MINBITS; lzw->nextcode = LZW_FIRST; lzw->oldcode = -1; /* generates LZW_CLEAR */ clearhash(lzw); return fz_okay; }
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; }
fz_error * fz_newpredict(fz_filter **fp, fz_obj *params, int encode) { fz_obj *obj; FZ_NEWFILTER(fz_predict, p, predict); p->encode = encode; p->predictor = 1; p->columns = 1; p->colors = 1; p->bpc = 8; obj = fz_dictgets(params, "Predictor"); if (obj) p->predictor = fz_toint(obj); obj = fz_dictgets(params, "Columns"); if (obj) p->columns = fz_toint(obj); obj = fz_dictgets(params, "Colors"); if (obj) p->colors = fz_toint(obj); obj = fz_dictgets(params, "BitsPerComponent"); if (obj) p->bpc = fz_toint(obj); p->stride = (p->bpc * p->colors * p->columns + 7) / 8; p->bpp = (p->bpc * p->colors + 7) / 8; if (p->predictor >= 10) { p->ref = fz_malloc(p->stride); if (!p->ref) { fz_free(p); return fz_throw("outofmem: predictor reference buffer"); } memset(p->ref, 0, p->stride); } else { p->ref = nil; } return fz_okay; }
static fz_error pdf_readxrefsections(pdf_xref *xref, int ofs, char *buf, int cap) { fz_error error; fz_obj *trailer; fz_obj *prev; fz_obj *xrefstm; error = pdf_readxref(&trailer, xref, ofs, buf, cap); if (error) return fz_rethrow(error, "cannot read xref section"); /* FIXME: do we overwrite free entries properly? */ xrefstm = fz_dictgets(trailer, "XRefStm"); if (xrefstm) { pdf_logxref("load xrefstm\n"); error = pdf_readxrefsections(xref, fz_toint(xrefstm), buf, cap); if (error) { fz_dropobj(trailer); return fz_rethrow(error, "cannot read /XRefStm xref section"); } } prev = fz_dictgets(trailer, "Prev"); if (prev) { pdf_logxref("load prev at 0x%x\n", fz_toint(prev)); error = pdf_readxrefsections(xref, fz_toint(prev), buf, cap); if (error) { fz_dropobj(trailer); return fz_rethrow(error, "cannot read /Prev xref section"); } } fz_dropobj(trailer); 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 pdf_repairobjstm(pdf_xref *xref, int num, int gen) { fz_error error; fz_obj *obj; fz_stream *stm; int tok; int i, n, count; char buf[256]; error = pdf_loadobject(&obj, xref, num, gen); if (error) return fz_rethrow(error, "cannot load object stream object (%d %d R)", num, gen); count = fz_toint(fz_dictgets(obj, "N")); fz_dropobj(obj); error = pdf_openstream(&stm, xref, num, gen); if (error) return fz_rethrow(error, "cannot open object stream object (%d %d R)", num, gen); for (i = 0; i < count; i++) { error = pdf_lex(&tok, stm, buf, sizeof buf, &n); if (error || tok != PDF_TINT) { fz_close(stm); return fz_rethrow(error, "corrupt object stream (%d %d R)", num, gen); } n = atoi(buf); if (n >= xref->len) pdf_resizexref(xref, n + 1); xref->table[n].ofs = num; xref->table[n].gen = i; xref->table[n].stmofs = 0; xref->table[n].obj = nil; xref->table[n].type = 'o'; error = pdf_lex(&tok, stm, buf, sizeof buf, &n); if (error || tok != PDF_TINT) { fz_close(stm); return fz_rethrow(error, "corrupt object stream (%d %d R)", num, gen); } } fz_close(stm); return fz_okay; }
static pdf_outline * pdf_loadoutlineimp(pdf_xref *xref, fz_obj *dict) { pdf_outline *node; fz_obj *obj; if (fz_isnull(dict)) return nil; node = fz_malloc(sizeof(pdf_outline)); node->title = nil; node->link = nil; node->child = nil; node->next = nil; node->count = 0; pdf_logpage("load outline {\n"); obj = fz_dictgets(dict, "Title"); if (obj) { node->title = pdf_toutf8(obj); pdf_logpage("title %s\n", node->title); } obj = fz_dictgets(dict, "Count"); if (obj) { node->count = fz_toint(obj); } if (fz_dictgets(dict, "Dest") || fz_dictgets(dict, "A")) { node->link = pdf_loadlink(xref, dict); } obj = fz_dictgets(dict, "First"); if (obj) { node->child = pdf_loadoutlineimp(xref, obj); } pdf_logpage("}\n"); obj = fz_dictgets(dict, "Next"); if (obj) { node->next = pdf_loadoutlineimp(xref, obj); } return node; }
fz_filter * fz_newdctd(fz_obj *params) { fz_obj *obj; int colortransform; FZ_NEWFILTER(fz_dctd, d, dctd); colortransform = -1; /* "unset" */ if (params) { obj = fz_dictgets(params, "ColorTransform"); if (obj) colortransform = fz_toint(obj); } d->colortransform = colortransform; d->stage = 0; /* setup error callback first thing */ myiniterr(&d->err); d->cinfo.err = (struct jpeg_error_mgr*) &d->err; if (setjmp(d->err.jb)) fz_warn("cannot initialise jpeg: %s", d->err.msg); /* create decompression object. this zeroes cinfo except for err. */ jpeg_create_decompress(&d->cinfo); /* prepare source manager */ d->cinfo.src = (struct jpeg_source_mgr *)&d->src; d->src.super.init_source = myinitsource; d->src.super.fill_input_buffer = myfillinput; d->src.super.skip_input_data = myskipinput; d->src.super.resync_to_restart = jpeg_resync_to_restart; d->src.super.term_source = mytermsource; d->src.super.bytes_in_buffer = 0; d->src.super.next_input_byte = nil; d->src.skip = 0; /* speed up jpeg decoding a bit */ d->cinfo.dct_method = JDCT_FASTEST; d->cinfo.do_fancy_upsampling = FALSE; return (fz_filter *)d; }
/* * Build a filter for reading raw stream data. * This is a null filter to constrain reading to the * stream length, followed by a decryption filter. */ static fz_error * buildrawfilter(fz_filter **filterp, pdf_xref *xref, fz_obj *stmobj, int oid, int gen) { fz_error *error; fz_filter *base; fz_obj *stmlen; int len; stmlen = fz_dictgets(stmobj, "Length"); error = pdf_resolve(&stmlen, xref); if (error) return fz_rethrow(error, "cannot resolve stream /Length"); len = fz_toint(stmlen); fz_dropobj(stmlen); error = fz_newnullfilter(&base, len); if (error) return fz_rethrow(error, "cannot create null filter"); if (xref->crypt) { fz_filter *crypt; fz_filter *pipe; error = pdf_cryptstream(&crypt, xref->crypt, oid, gen); if (error) { fz_dropfilter(base); return fz_rethrow(error, "cannot create decryption filter"); } error = fz_newpipeline(&pipe, base, crypt); fz_dropfilter(base); fz_dropfilter(crypt); if (error) return fz_rethrow(error, "cannot create pipeline filter"); *filterp = pipe; } else { *filterp = base; } return fz_okay; }
fz_stream * fz_opendctd(fz_stream *chain, fz_obj *params) { fz_dctd *state; fz_obj *obj; state = fz_malloc(sizeof(fz_dctd)); memset(state, 0, sizeof(fz_dctd)); state->chain = chain; state->colortransform = -1; /* unset */ state->init = 0; obj = fz_dictgets(params, "ColorTransform"); if (obj) state->colortransform = fz_toint(obj); return fz_newstream(state, readdctd, closedctd); }
fz_stream * fz_openlzwd(fz_stream *chain, fz_obj *params) { fz_lzwd *lzw; fz_obj *obj; int i; lzw = fz_malloc(sizeof(fz_lzwd)); lzw->chain = chain; lzw->eod = 0; lzw->earlychange = 1; obj = fz_dictgets(params, "EarlyChange"); if (obj) lzw->earlychange = !!fz_toint(obj); lzw->bidx = 32; lzw->word = 0; for (i = 0; i < 256; i++) { lzw->table[i].value = i; lzw->table[i].firstchar = i; lzw->table[i].length = 1; lzw->table[i].prev = -1; } for (i = 256; i < NUMCODES; i++) { lzw->table[i].value = 0; lzw->table[i].firstchar = 0; lzw->table[i].length = 0; lzw->table[i].prev = -1; } lzw->codebits = MINBITS; lzw->code = -1; lzw->nextcode = LZW_FIRST; lzw->oldcode = -1; lzw->rp = lzw->bp; lzw->wp = lzw->bp; return fz_newstream(lzw, readlzwd, closelzwd); }
fz_error fz_newlzwd(fz_filter **fp, fz_obj *params) { int i; FZ_NEWFILTER(fz_lzwd, lzw, lzwd); lzw->earlychange = 0; if (params) { fz_obj *obj; obj = fz_dictgets(params, "EarlyChange"); if (obj) lzw->earlychange = fz_toint(obj) != 0; } lzw->bidx = 32; lzw->word = 0; for (i = 0; i < 256; i++) { lzw->table[i].value = i; lzw->table[i].firstchar = i; lzw->table[i].length = 1; lzw->table[i].prev = -1; } for (i = 256; i < NUMCODES; i++) { lzw->table[i].value = 0; lzw->table[i].firstchar = 0; lzw->table[i].length = 0; lzw->table[i].prev = -1; } lzw->codebits = MINBITS; lzw->code = -1; lzw->nextcode = LZW_FIRST; lzw->oldcode = -1; lzw->resume = 0; return fz_okay; }
static fz_error loadiccbased(fz_colorspace **csp, pdf_xref *xref, fz_obj *ref) { fz_obj *dict; int n; pdf_logrsrc("load ICCBased\n"); dict = fz_resolveindirect(ref); n = fz_toint(fz_dictgets(dict, "N")); switch (n) { case 1: *csp = pdf_devicegray; return fz_okay; case 3: *csp = pdf_devicergb; return fz_okay; case 4: *csp = pdf_devicecmyk; return fz_okay; } return fz_throw("syntaxerror: ICCBased must have 1, 3 or 4 components"); }
static void loadcolorkey(int *colorkey, int bpc, int indexed, fz_obj *obj) { int scale = 1; int i; pdf_logimage("keyed mask\n"); if (!indexed) { switch (bpc) { case 1: scale = 255; break; case 2: scale = 85; break; case 4: scale = 17; break; case 8: scale = 1; break; } } for (i = 0; i < fz_arraylen(obj); i++) colorkey[i] = fz_toint(fz_arrayget(obj, i)) * scale; }
fz_error pdf_loadpattern(pdf_pattern **patp, pdf_xref *xref, fz_obj *dict) { fz_error error; pdf_pattern *pat; fz_obj *obj; if ((*patp = pdf_finditem(xref->store, pdf_droppattern, dict))) { pdf_keeppattern(*patp); return fz_okay; } pdf_logrsrc("load pattern (%d %d R) {\n", fz_tonum(dict), fz_togen(dict)); pat = fz_malloc(sizeof(pdf_pattern)); pat->refs = 1; pat->resources = nil; pat->contents = nil; /* Store pattern now, to avoid possible recursion if objects refer back to this one */ pdf_storeitem(xref->store, pdf_keeppattern, pdf_droppattern, dict, pat); pat->ismask = fz_toint(fz_dictgets(dict, "PaintType")) == 2; pat->xstep = fz_toreal(fz_dictgets(dict, "XStep")); pat->ystep = fz_toreal(fz_dictgets(dict, "YStep")); pdf_logrsrc("mask %d\n", pat->ismask); pdf_logrsrc("xstep %g\n", pat->xstep); pdf_logrsrc("ystep %g\n", pat->ystep); obj = fz_dictgets(dict, "BBox"); pat->bbox = pdf_torect(obj); pdf_logrsrc("bbox [%g %g %g %g]\n", pat->bbox.x0, pat->bbox.y0, pat->bbox.x1, pat->bbox.y1); obj = fz_dictgets(dict, "Matrix"); if (obj) pat->matrix = pdf_tomatrix(obj); else pat->matrix = fz_identity; pdf_logrsrc("matrix [%g %g %g %g %g %g]\n", pat->matrix.a, pat->matrix.b, pat->matrix.c, pat->matrix.d, pat->matrix.e, pat->matrix.f); pat->resources = fz_dictgets(dict, "Resources"); if (pat->resources) fz_keepobj(pat->resources); error = pdf_loadstream(&pat->contents, xref, fz_tonum(dict), fz_togen(dict)); if (error) { pdf_removeitem(xref->store, pdf_droppattern, dict); pdf_droppattern(pat); return fz_rethrow(error, "cannot load pattern stream (%d %d R)", fz_tonum(dict), fz_togen(dict)); } pdf_logrsrc("}\n"); *patp = pat; return fz_okay; }
static fz_error * loadshadedict(fz_shade **shadep, pdf_xref *xref, fz_obj *dict, fz_obj *ref, fz_matrix matrix) { fz_error *error; fz_shade *shade; fz_obj *obj; int type; int i; pdf_logshade("load shade dict %d %d {\n", fz_tonum(ref), fz_togen(ref)); shade = fz_malloc(sizeof(fz_shade)); if (!shade) return fz_outofmem; shade->refs = 1; shade->usebackground = 0; shade->usefunction = 0; shade->matrix = matrix; shade->bbox = fz_infiniterect; shade->meshlen = 0; shade->meshcap = 0; shade->mesh = nil; obj = fz_dictgets(dict, "ShadingType"); type = fz_toint(obj); pdf_logshade("type %d\n", type); /* TODO: flatten indexed... */ obj = fz_dictgets(dict, "ColorSpace"); if (obj) { shade->cs = pdf_finditem(xref->store, PDF_KCOLORSPACE, obj); if (shade->cs) fz_keepcolorspace(shade->cs); else { error = pdf_resolve(&obj, xref); if (error) return error; error = pdf_loadcolorspace(&shade->cs, xref, obj); if (error) return error; fz_dropobj(obj); } } pdf_logshade("colorspace %s\n", shade->cs->name); obj = fz_dictgets(dict, "Background"); if (obj) { pdf_logshade("background\n"); shade->usebackground = 1; for (i = 0; i < shade->cs->n; i++) shade->background[i] = fz_toreal(fz_arrayget(obj, i)); } obj = fz_dictgets(dict, "BBox"); if (fz_isarray(obj)) { shade->bbox = pdf_torect(obj); pdf_logshade("bbox [%g %g %g %g]\n", shade->bbox.x0, shade->bbox.y0, shade->bbox.x1, shade->bbox.y1); } switch(type) { case 1: error = pdf_loadtype1shade(shade, xref, dict, ref); if (error) goto cleanup; break; case 2: error = pdf_loadtype2shade(shade, xref, dict, ref); if (error) goto cleanup; break; case 3: error = pdf_loadtype3shade(shade, xref, dict, ref); if (error) goto cleanup; break; case 4: error = pdf_loadtype4shade(shade, xref, dict, ref); if (error) goto cleanup; break; case 5: error = pdf_loadtype5shade(shade, xref, dict, ref); if (error) goto cleanup; break; case 6: error = pdf_loadtype6shade(shade, xref, dict, ref); if (error) goto cleanup; break; case 7: error = pdf_loadtype7shade(shade, xref, dict, ref); if (error) goto cleanup; break; default: fz_warn("syntaxerror: unknown shading type: %d", type); break; }; pdf_logshade("}\n"); *shadep = shade; return nil; cleanup: fz_dropshade(shade); return error; }
void pdfmoz_open(pdfmoz_t *moz, char *filename) { SCROLLINFO si; fz_error error; fz_obj *obj; fz_irect bbox; int rot; int i; strcpy(moz->error, ""); error = fz_newrenderer(&moz->rast, pdf_devicergb, 0, 1024 * 512); if (error) pdfmoz_error(moz, error); /* * Open PDF and load xref table */ moz->filename = filename; moz->xref = pdf_newxref(); error = pdf_loadxref(moz->xref, filename); if (error) { error = pdf_repairxref(moz->xref, filename); if (error) pdfmoz_error(moz, error); } /* * Handle encrypted PDF files */ error = pdf_decryptxref(moz->xref); if (error) pdfmoz_error(moz, error); if (pdf_needspassword(moz->xref)) { pdfmoz_warn(moz, "PDF file is encrypted and needs a password."); } moz->pagecount = pdf_getpagecount(moz->xref); moz->pages = fz_malloc(sizeof(page_t) * moz->pagecount); for (i = 0; i < moz->pagecount; i++) { fz_obj *pageobj; pageobj = pdf_getpageobject(moz->xref, i); moz->pages[i].obj = fz_keepobj(pageobj); moz->pages[i].page = nil; moz->pages[i].image = nil; obj = fz_dictgets(moz->pages[i].obj, "CropBox"); if (!obj) obj = fz_dictgets(moz->pages[i].obj, "MediaBox"); bbox = fz_roundrect(pdf_torect(obj)); moz->pages[i].w = bbox.x1 - bbox.x0; moz->pages[i].h = bbox.y1 - bbox.y0; rot = fz_toint(fz_dictgets(moz->pages[i].obj, "Rotate")); if ((rot / 90) % 2) { int t = moz->pages[i].w; moz->pages[i].w = moz->pages[i].h; moz->pages[i].h = t; } moz->pages[i].px = 1 + PAD; } /* * Load meta information * TODO: move this into mupdf library */ obj = fz_dictgets(moz->xref->trailer, "Root"); moz->xref->root = fz_resolveindirect(obj); if (!moz->xref->root) pdfmoz_error(moz, fz_throw("syntaxerror: missing Root object")); if (moz->xref->root) fz_keepobj(moz->xref->root); obj = fz_dictgets(moz->xref->trailer, "Info"); moz->xref->info = fz_resolveindirect(obj); if (moz->xref->info) fz_keepobj(moz->xref->info); moz->doctitle = filename; if (strrchr(moz->doctitle, '\\')) moz->doctitle = strrchr(moz->doctitle, '\\') + 1; if (strrchr(moz->doctitle, '/')) moz->doctitle = strrchr(moz->doctitle, '/') + 1; if (moz->xref->info) { obj = fz_dictgets(moz->xref->info, "Title"); if (obj) moz->doctitle = pdf_toutf8(obj); } /* * Start at first page */ si.cbSize = sizeof(si); si.fMask = SIF_POS | SIF_RANGE; // XXX | SIF_PAGE; si.nPos = 0; si.nMin = 0; si.nMax = 1; si.nPage = 1; SetScrollInfo(moz->hwnd, SB_VERT, &si, TRUE); moz->scrollpage = 0; moz->scrollyofs = 0; InvalidateRect(moz->hwnd, NULL, FALSE); }
fz_error pdf_loadfontdescriptor(pdf_fontdesc *fontdesc, pdf_xref *xref, fz_obj *dict, char *collection, char *basefont) { fz_error error; fz_obj *obj1, *obj2, *obj3, *obj; fz_rect bbox; char *fontname; char *origname; pdf_logfont("load fontdescriptor {\n"); /* cf. http://code.google.com/p/sumatrapdf/issues/detail?id=1014 */ if (!strchr(basefont, ',') || strchr(basefont, '+')) origname = fz_toname(fz_dictgets(dict, "FontName")); else origname = basefont; fontname = cleanfontname(origname); pdf_logfont("fontname %s -> %s\n", origname, fontname); fontdesc->flags = fz_toint(fz_dictgets(dict, "Flags")); fontdesc->italicangle = fz_toreal(fz_dictgets(dict, "ItalicAngle")); fontdesc->ascent = fz_toreal(fz_dictgets(dict, "Ascent")); fontdesc->descent = fz_toreal(fz_dictgets(dict, "Descent")); fontdesc->capheight = fz_toreal(fz_dictgets(dict, "CapHeight")); fontdesc->xheight = fz_toreal(fz_dictgets(dict, "XHeight")); fontdesc->missingwidth = fz_toreal(fz_dictgets(dict, "MissingWidth")); bbox = pdf_torect(fz_dictgets(dict, "FontBBox")); pdf_logfont("bbox [%g %g %g %g]\n", bbox.x0, bbox.y0, bbox.x1, bbox.y1); pdf_logfont("flags %d\n", fontdesc->flags); obj1 = fz_dictgets(dict, "FontFile"); obj2 = fz_dictgets(dict, "FontFile2"); obj3 = fz_dictgets(dict, "FontFile3"); obj = obj1 ? obj1 : obj2 ? obj2 : obj3; if (getenv("NOFONT")) obj = nil; if (fz_isindirect(obj)) { error = pdf_loadembeddedfont(fontdesc, xref, obj); if (error) { fz_catch(error, "ignored error when loading embedded font, attempting to load system font"); if (origname != fontname) error = pdf_loadbuiltinfont(fontdesc, fontname); else error = pdf_loadsystemfont(fontdesc, fontname, collection); if (error) return fz_rethrow(error, "cannot load font descriptor (%d %d R)", fz_tonum(dict), fz_togen(dict)); } } else { if (origname != fontname && 0 /* SumatraPDF: prefer system fonts to the built-in ones */) error = pdf_loadbuiltinfont(fontdesc, fontname); else error = pdf_loadsystemfont(fontdesc, fontname, collection); if (error) return fz_rethrow(error, "cannot load font descriptor (%d %d R)", fz_tonum(dict), fz_togen(dict)); } fz_strlcpy(fontdesc->font->name, fontname, sizeof fontdesc->font->name); pdf_logfont("}\n"); return fz_okay; }
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 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 fz_error loadindexed(fz_colorspace **csp, pdf_xref *xref, fz_obj *array) { fz_error error; pdf_indexed *cs; fz_obj *baseobj = fz_arrayget(array, 1); fz_obj *highobj = fz_arrayget(array, 2); fz_obj *lookup = fz_arrayget(array, 3); fz_colorspace *base; int n; pdf_logrsrc("load Indexed {\n"); error = pdf_loadcolorspace(&base, xref, baseobj); if (error) return fz_rethrow(error, "cannot load base colorspace"); pdf_logrsrc("base %s\n", base->name); cs = fz_malloc(sizeof(pdf_indexed)); initcs((fz_colorspace*)cs, "Indexed", 1, nil, nil, freeindexed); cs->base = fz_keepcolorspace(base); cs->high = fz_toint(highobj); fz_dropcolorspace(base); n = base->n * (cs->high + 1); cs->lookup = fz_malloc(n); if (fz_isstring(lookup) && fz_tostrlen(lookup) == n) { unsigned char *buf; int i; pdf_logrsrc("string lookup\n"); buf = (unsigned char *) fz_tostrbuf(lookup); for (i = 0; i < n; i++) cs->lookup[i] = buf[i]; } else if (fz_isindirect(lookup)) { fz_buffer *buf; int i; pdf_logrsrc("stream lookup\n"); error = pdf_loadstream(&buf, xref, fz_tonum(lookup), fz_togen(lookup)); if (error) { fz_dropcolorspace((fz_colorspace*)cs); return fz_rethrow(error, "cannot load colorpsace lookup table"); } for (i = 0; i < n && i < (buf->wp - buf->rp); i++) cs->lookup[i] = buf->rp[i]; fz_dropbuffer(buf); } else return fz_throw("cannot parse colorspace lookup table"); pdf_logrsrc("}\n"); *csp = (fz_colorspace*)cs; return fz_okay; }
static fz_error pdf_loadobjstm(pdf_xref *xref, int num, int gen, char *buf, int cap) { fz_error error; fz_stream *stm; fz_obj *objstm; int *numbuf; int *ofsbuf; fz_obj *obj; int first; int count; int i, n; pdf_token_e tok; pdf_logxref("loadobjstm (%d %d R)\n", num, gen); error = pdf_loadobject(&objstm, xref, num, gen); if (error) return fz_rethrow(error, "cannot load object stream object (%d %d R)", num, gen); count = fz_toint(fz_dictgets(objstm, "N")); first = fz_toint(fz_dictgets(objstm, "First")); pdf_logxref("\tcount %d\n", count); numbuf = fz_malloc(count * sizeof(int)); ofsbuf = fz_malloc(count * sizeof(int)); error = pdf_openstream(&stm, xref, num, gen); if (error) { error = fz_rethrow(error, "cannot open object stream (%d %d R)", num, gen); goto cleanupbuf; } for (i = 0; i < count; i++) { error = pdf_lex(&tok, stm, buf, cap, &n); if (error || tok != PDF_TINT) { error = fz_rethrow(error, "corrupt object stream (%d %d R)", num, gen); goto cleanupstm; } numbuf[i] = atoi(buf); error = pdf_lex(&tok, stm, buf, cap, &n); if (error || tok != PDF_TINT) { error = fz_rethrow(error, "corrupt object stream (%d %d R)", num, gen); goto cleanupstm; } ofsbuf[i] = atoi(buf); } fz_seek(stm, first, 0); for (i = 0; i < count; i++) { fz_seek(stm, first + ofsbuf[i], 0); error = pdf_parsestmobj(&obj, xref, stm, buf, cap); if (error) { error = fz_rethrow(error, "cannot parse object %d in stream (%d %d R)", i, num, gen); goto cleanupstm; } if (numbuf[i] < 1 || numbuf[i] >= xref->len) { fz_dropobj(obj); error = fz_throw("object id (%d 0 R) out of range (0..%d)", numbuf[i], xref->len - 1); goto cleanupstm; } if (xref->table[numbuf[i]].type == 'o' && xref->table[numbuf[i]].ofs == num) { if (xref->table[numbuf[i]].obj) fz_dropobj(xref->table[numbuf[i]].obj); xref->table[numbuf[i]].obj = obj; } else { fz_dropobj(obj); } } fz_close(stm); fz_free(ofsbuf); fz_free(numbuf); fz_dropobj(objstm); return fz_okay; cleanupstm: fz_close(stm); cleanupbuf: fz_free(ofsbuf); fz_free(numbuf); fz_dropobj(objstm); return error; /* already rethrown */ }
static fz_error pdf_readnewxref(fz_obj **trailerp, pdf_xref *xref, char *buf, int cap) { fz_error error; fz_stream *stm; fz_obj *trailer; fz_obj *index; fz_obj *obj; int num, gen, stmofs; int size, w0, w1, w2; int t; int i; pdf_logxref("load new xref format\n"); error = pdf_parseindobj(&trailer, xref, xref->file, buf, cap, &num, &gen, &stmofs); if (error) return fz_rethrow(error, "cannot parse compressed xref stream object"); obj = fz_dictgets(trailer, "Size"); if (!obj) { fz_dropobj(trailer); return fz_throw("xref stream missing Size entry (%d %d R)", num, gen); } size = fz_toint(obj); if (size >= xref->cap) { xref->cap = size + 1; /* for hack to allow broken pdf generators with off-by-one errors */ xref->table = fz_realloc(xref->table, xref->cap * sizeof(pdf_xrefentry)); } if (size > xref->len) { for (i = xref->len; i < xref->cap; i++) { xref->table[i].ofs = 0; xref->table[i].gen = 0; xref->table[i].stmofs = 0; xref->table[i].obj = nil; xref->table[i].type = 0; } xref->len = size; } if (num < 0 || num >= xref->len) { if (num == xref->len && num < xref->cap) { /* allow broken pdf files that have off-by-one errors in the xref */ fz_warn("object id (%d %d R) out of range (0..%d)", num, gen, xref->len - 1); xref->len ++; } else { fz_dropobj(trailer); return fz_throw("object id (%d %d R) out of range (0..%d)", num, gen, xref->len - 1); } } pdf_logxref("\tnum=%d gen=%d size=%d\n", num, gen, size); obj = fz_dictgets(trailer, "W"); if (!obj) { fz_dropobj(trailer); return fz_throw("xref stream missing W entry (%d %d R)", num, gen); } w0 = fz_toint(fz_arrayget(obj, 0)); w1 = fz_toint(fz_arrayget(obj, 1)); w2 = fz_toint(fz_arrayget(obj, 2)); index = fz_dictgets(trailer, "Index"); error = pdf_openstreamat(&stm, xref, num, gen, trailer, stmofs); if (error) { fz_dropobj(trailer); return fz_rethrow(error, "cannot open compressed xref stream (%d %d R)", num, gen); } if (!index) { error = pdf_readnewxrefsection(xref, stm, 0, size, w0, w1, w2); if (error) { fz_close(stm); fz_dropobj(trailer); return fz_rethrow(error, "cannot read xref stream (%d %d R)", num, gen); } } else { for (t = 0; t < fz_arraylen(index); t += 2) { int i0 = fz_toint(fz_arrayget(index, t + 0)); int i1 = fz_toint(fz_arrayget(index, t + 1)); error = pdf_readnewxrefsection(xref, stm, i0, i1, w0, w1, w2); if (error) { fz_close(stm); fz_dropobj(trailer); return fz_rethrow(error, "cannot read xref stream section (%d %d R)", num, gen); } } } fz_close(stm); *trailerp = trailer; return fz_okay; }
static fz_error * loadindexed(fz_colorspace **csp, pdf_xref *xref, fz_obj *array) { fz_error *error; pdf_indexed *cs; fz_obj *baseobj = fz_arrayget(array, 1); fz_obj *highobj = fz_arrayget(array, 2); fz_obj *lookup = fz_arrayget(array, 3); fz_colorspace *base; int n; pdf_logrsrc("load Indexed {\n"); error = pdf_resolve(&baseobj, xref); if (error) return error; error = pdf_loadcolorspace(&base, xref, baseobj); fz_dropobj(baseobj); if (error) return error; pdf_logrsrc("base %s\n", base->name); cs = fz_malloc(sizeof(pdf_indexed)); if (!cs) { fz_dropcolorspace(base); return fz_outofmem; } initcs((fz_colorspace*)cs, "Indexed", 1, nil, nil, dropindexed); cs->base = base; cs->high = fz_toint(highobj); n = base->n * (cs->high + 1); cs->lookup = fz_malloc(n); if (!cs->lookup) { fz_dropcolorspace((fz_colorspace*)cs); return fz_outofmem; } if (fz_isstring(lookup) && fz_tostrlen(lookup) == n) { unsigned char *buf; int i; pdf_logrsrc("string lookup\n"); buf = fz_tostrbuf(lookup); for (i = 0; i < n; i++) cs->lookup[i] = buf[i]; } if (fz_isindirect(lookup)) { fz_buffer *buf; int i; pdf_logrsrc("stream lookup\n"); error = pdf_loadstream(&buf, xref, fz_tonum(lookup), fz_togen(lookup)); if (error) { fz_dropcolorspace((fz_colorspace*)cs); return error; } for (i = 0; i < n && i < (buf->wp - buf->rp); i++) cs->lookup[i] = buf->rp[i]; fz_dropbuffer(buf); } pdf_logrsrc("}\n"); *csp = (fz_colorspace*)cs; return nil; }
/* * 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 */ }