void
pdf_storeitem(pdf_store *store, pdf_itemkind kind, fz_obj *key, void *val)
{
pdf_item *item;
item = fz_malloc(sizeof(pdf_item));
item->kind = kind;
item->key = fz_keepobj(key);
item->val = val;
item->age = 0;
item->next = nil;
if (fz_isindirect(key))
{
struct refkey refkey;
pdf_logrsrc("store item %s (%d %d R) ptr=%p\n", kindstr(kind), fz_tonum(key), fz_togen(key), val);
refkey.kind = kind;
refkey.oid = fz_tonum(key);
refkey.gen = fz_togen(key);
fz_hashinsert(store->hash, &refkey, item);
}
else
{
pdf_logrsrc("store item %s: ... = %p\n", kindstr(kind), val);
item->next = store->root;
store->root = item;
}
keepitem(kind, val);
}
void openxref(char *filename, char *password, int dieonbadpass)
{
fz_error error;
int okay;
basename = strrchr(filename, '/');
if (!basename)
basename = filename;
else
basename++;
xref = pdf_newxref();
error = pdf_loadxref(xref, filename);
if (error)
{
fz_catch(error, "trying to repair");
error = pdf_repairxref(xref, filename);
if (error)
die(error);
}
error = pdf_decryptxref(xref);
if (error)
die(error);
if (pdf_needspassword(xref))
{
okay = pdf_authenticatepassword(xref, password);
if (!okay && !dieonbadpass)
fz_warn("invalid password, attempting to continue.");
else if (!okay && dieonbadpass)
die(fz_throw("invalid password"));
}
xref->root = fz_dictgets(xref->trailer, "Root");
if (xref->root)
fz_keepobj(xref->root);
xref->info = fz_dictgets(xref->trailer, "Info");
if (xref->info)
fz_keepobj(xref->info);
pagecount = pdf_getpagecount(xref);
}
pdf_link *
pdf_newlink(pdf_linkkind kind, fz_rect bbox, fz_obj *dest)
{
pdf_link *link = fz_malloc(sizeof(pdf_link));
link->kind = kind;
link->rect = bbox;
link->dest = fz_keepobj(dest);
link->next = nil;
return link;
}
fz_error *
pdf_storeitem(pdf_store *store, pdf_itemkind kind, fz_obj *key, void *val)
{
fz_error *error;
switch (kind)
{
case PDF_KCOLORSPACE: fz_keepcolorspace(val); break;
case PDF_KFUNCTION: pdf_keepfunction(val); break;
case PDF_KXOBJECT: pdf_keepxobject(val); break;
case PDF_KIMAGE: fz_keepimage(val); break;
case PDF_KPATTERN: pdf_keeppattern(val); break;
case PDF_KSHADE: fz_keepshade(val); break;
case PDF_KCMAP: pdf_keepcmap(val); break;
case PDF_KFONT: fz_keepfont(val); break;
}
if (fz_isindirect(key))
{
struct refkey item;
pdf_logrsrc("store item %d: %d %d R = %p\n", kind, fz_tonum(key), fz_togen(key), val);
item.kind = kind;
item.oid = fz_tonum(key);
item.gen = fz_togen(key);
error = fz_hashinsert(store->hash, &item, val);
if (error)
return error;
}
else
{
pdf_item *item;
item = fz_malloc(sizeof(pdf_item));
if (!item)
return fz_outofmem;
pdf_logrsrc("store item %d: ... = %p\n", kind, val);
item->kind = kind;
item->key = fz_keepobj(key);
item->val = val;
item->next = store->root;
store->root = item;
}
return nil;
}
fz_error
pdf_loadobject(fz_obj **objp, pdf_xref *xref, int num, int gen)
{
fz_error error;
error = pdf_cacheobject(xref, num, gen);
if (error)
return fz_rethrow(error, "cannot load object (%d %d R) into cache", num, gen);
assert(xref->table[num].obj);
*objp = fz_keepobj(xref->table[num].obj);
return fz_okay;
}
fz_error *
pdf_newlink(pdf_link **linkp, fz_rect bbox, fz_obj *dest, pdf_linkkind kind)
{
pdf_link *link;
link = fz_malloc(sizeof(pdf_link));
if (!link)
return fz_outofmem;
link->rect = bbox;
link->dest = fz_keepobj(dest);
link->kind = kind;
link->next = nil;
*linkp = link;
return nil;
}
fz_error
pdf_newlink(pdf_link **linkp, pdf_linkkind kind, fz_rect bbox, fz_obj *dest)
{
pdf_link *link;
link = fz_malloc(sizeof(pdf_link));
if (!link)
return fz_rethrow(-1, "out of memory");
link->kind = kind;
link->rect = bbox;
link->dest = fz_keepobj(dest);
link->next = nil;
*linkp = link;
return fz_okay;
}
/* Replace numbered object -- for use by pdfclean and similar tools */
void
pdf_updateobject(pdf_xref *xref, int num, int gen, fz_obj *newobj)
{
pdf_xrefentry *x;
if (num < 0 || num >= xref->len)
{
fz_warn("object out of range (%d %d R); xref size %d", num, gen, xref->len);
return;
}
x = &xref->table[num];
if (x->obj)
fz_dropobj(x->obj);
x->obj = fz_keepobj(newobj);
x->type = 'n';
x->ofs = 0;
}
fz_error
pdf_repairxref(pdf_xref *xref, char *buf, int bufsize)
{
fz_error error;
fz_obj *dict, *obj;
fz_obj *length;
fz_obj *encrypt = nil;
fz_obj *id = nil;
fz_obj *root = nil;
fz_obj *info = nil;
struct entry *list = nil;
int listlen;
int listcap;
int maxnum = 0;
int num = 0;
int gen = 0;
int tmpofs, numofs = 0, genofs = 0;
int stmlen, stmofs = 0;
int tok;
int next;
int i, n;
pdf_logxref("repairxref %p\n", xref);
fz_seek(xref->file, 0, 0);
listlen = 0;
listcap = 1024;
list = fz_calloc(listcap, sizeof(struct entry));
/* look for '%PDF' version marker within first kilobyte of file */
n = fz_read(xref->file, (unsigned char *)buf, MAX(bufsize, 1024));
if (n < 0)
{
error = fz_rethrow(n, "cannot read from file");
goto cleanup;
}
fz_seek(xref->file, 0, 0);
for (i = 0; i < n - 4; i++)
{
if (memcmp(buf + i, "%PDF", 4) == 0)
{
fz_seek(xref->file, i, 0);
break;
}
}
while (1)
{
tmpofs = fz_tell(xref->file);
if (tmpofs < 0)
{
error = fz_throw("cannot tell in file");
goto cleanup;
}
error = pdf_lex(&tok, xref->file, buf, bufsize, &n);
if (error)
{
fz_catch(error, "ignoring the rest of the file");
break;
}
if (tok == PDF_TINT)
{
numofs = genofs;
num = gen;
genofs = tmpofs;
gen = atoi(buf);
}
if (tok == PDF_TOBJ)
{
error = fz_repairobj(xref->file, buf, bufsize, &stmofs, &stmlen, &encrypt, &id);
if (error)
{
error = fz_rethrow(error, "cannot parse object (%d %d R)", num, gen);
goto cleanup;
}
pdf_logxref("found object: (%d %d R)\n", num, gen);
if (listlen + 1 == listcap)
{
listcap = (listcap * 3) / 2;
list = fz_realloc(list, listcap, sizeof(struct entry));
}
list[listlen].num = num;
list[listlen].gen = gen;
list[listlen].ofs = numofs;
list[listlen].stmofs = stmofs;
list[listlen].stmlen = stmlen;
listlen ++;
if (num > maxnum)
maxnum = num;
}
/* trailer dictionary */
if (tok == PDF_TODICT)
{
error = pdf_parsedict(&dict, xref, xref->file, buf, bufsize);
if (error)
{
error = fz_rethrow(error, "cannot parse object");
goto cleanup;
}
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, "Root");
if (obj)
{
if (root)
fz_dropobj(root);
root = fz_keepobj(obj);
}
obj = fz_dictgets(dict, "Info");
if (obj)
{
if (info)
fz_dropobj(info);
info = fz_keepobj(obj);
}
fz_dropobj(dict);
}
if (tok == PDF_TERROR)
fz_readbyte(xref->file);
if (tok == PDF_TEOF)
break;
}
/* make xref reasonable */
pdf_resizexref(xref, maxnum + 1);
for (i = 0; i < listlen; i++)
{
xref->table[list[i].num].type = 'n';
xref->table[list[i].num].ofs = list[i].ofs;
xref->table[list[i].num].gen = list[i].gen;
xref->table[list[i].num].stmofs = list[i].stmofs;
/* corrected stream length */
if (list[i].stmlen >= 0)
{
pdf_logxref("correct stream length %d %d = %d\n",
list[i].num, list[i].gen, list[i].stmlen);
error = pdf_loadobject(&dict, xref, list[i].num, list[i].gen);
if (error)
{
error = fz_rethrow(error, "cannot load stream object (%d %d R)", list[i].num, list[i].gen);
goto cleanup;
}
length = fz_newint(list[i].stmlen);
fz_dictputs(dict, "Length", length);
fz_dropobj(length);
fz_dropobj(dict);
}
}
xref->table[0].type = 'f';
xref->table[0].ofs = 0;
xref->table[0].gen = 65535;
xref->table[0].stmofs = 0;
xref->table[0].obj = nil;
next = 0;
for (i = xref->len - 1; i >= 0; i--)
{
if (xref->table[i].type == 'f')
{
xref->table[i].ofs = next;
if (xref->table[i].gen < 65535)
xref->table[i].gen ++;
next = i;
}
}
/* create a repaired trailer, Root will be added later */
xref->trailer = fz_newdict(5);
obj = fz_newint(maxnum + 1);
fz_dictputs(xref->trailer, "Size", obj);
fz_dropobj(obj);
if (root)
{
fz_dictputs(xref->trailer, "Root", root);
fz_dropobj(root);
}
if (info)
{
fz_dictputs(xref->trailer, "Info", info);
fz_dropobj(info);
}
if (encrypt)
{
if (fz_isindirect(encrypt))
{
/* create new reference with non-nil xref pointer */
obj = fz_newindirect(fz_tonum(encrypt), fz_togen(encrypt), xref);
fz_dropobj(encrypt);
encrypt = obj;
}
fz_dictputs(xref->trailer, "Encrypt", encrypt);
fz_dropobj(encrypt);
}
if (id)
{
if (fz_isindirect(id))
{
/* create new reference with non-nil xref pointer */
obj = fz_newindirect(fz_tonum(id), fz_togen(id), xref);
fz_dropobj(id);
id = obj;
}
fz_dictputs(xref->trailer, "ID", id);
fz_dropobj(id);
}
fz_free(list);
return fz_okay;
cleanup:
if (encrypt) fz_dropobj(encrypt);
if (id) fz_dropobj(id);
if (root) fz_dropobj(root);
if (info) fz_dropobj(info);
fz_free(list);
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 *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))
fz_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);
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;
fontdesc->font->t3widths[i] = w * 0.001f;
pdf_addhmtx(fontdesc, i, i, w);
}
pdf_endhmtx(fontdesc);
/* Resources -- inherit page resources if the font doesn't have its own */
fontdesc->font->t3resources = fz_dictgets(dict, "Resources");
if (!fontdesc->font->t3resources)
fontdesc->font->t3resources = rdb;
if (fontdesc->font->t3resources)
fz_keepobj(fontdesc->font->t3resources);
if (!fontdesc->font->t3resources)
fz_warn("no resource dictionary for type 3 font!");
fontdesc->font->t3xref = xref;
fontdesc->font->t3run = pdf_runcontents;
/* 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 (pdf_isstream(xref, fz_tonum(obj), fz_togen(obj)))
{
error = pdf_loadstream(&fontdesc->font->t3procs[i], xref, fz_tonum(obj), fz_togen(obj));
if (error)
goto cleanup;
}
}
}
pdf_logfont("}\n");
*fontdescp = fontdesc;
return fz_okay;
cleanup:
fz_dropfont(fontdesc->font);
fz_free(fontdesc);
return fz_rethrow(error, "cannot load type3 font (%d %d R)", fz_tonum(dict), fz_togen(dict));
}
static fz_error *
loadpagetree(pdf_xref *xref, pdf_pagetree *pages,
struct stuff inherit, fz_obj *obj, fz_obj *ref, int *pagenum)
{
fz_error *error;
fz_obj *type;
fz_obj *kids;
fz_obj *kref, *kobj;
fz_obj *inh;
int i;
type = fz_dictgets(obj, "Type");
if (strcmp(fz_toname(type), "Page") == 0)
{
pdf_logpage("page %d, %d %d\n", *pagenum, ref->u.r.oid, ref->u.r.gid);
(*pagenum)++;
if (inherit.resources && !fz_dictgets(obj, "Resources"))
{
pdf_logpage("inherit resources (%d)\n", pages->cursor);
error = fz_dictputs(obj, "Resources", inherit.resources);
if (error) return fz_rethrow(error, "cannot inherit page tree resources");
}
if (inherit.mediabox && !fz_dictgets(obj, "MediaBox"))
{
pdf_logpage("inherit mediabox (%d)\n", pages->cursor);
error = fz_dictputs(obj, "MediaBox", inherit.mediabox);
if (error) return fz_rethrow(error, "cannot inherit page tree mediabox");
}
if (inherit.cropbox && !fz_dictgets(obj, "CropBox"))
{
pdf_logpage("inherit cropbox (%d)\n", pages->cursor);
error = fz_dictputs(obj, "CropBox", inherit.cropbox);
if (error) return fz_rethrow(error, "cannot inherit page tree cropbox");
}
if (inherit.rotate && !fz_dictgets(obj, "Rotate"))
{
pdf_logpage("inherit rotate (%d)\n", pages->cursor);
error = fz_dictputs(obj, "Rotate", inherit.rotate);
if (error) return fz_rethrow(error, "cannot inherit page tree rotate");
}
pages->pref[pages->cursor] = fz_keepobj(ref);
pages->pobj[pages->cursor] = fz_keepobj(obj);
pages->cursor ++;
}
else if (strcmp(fz_toname(type), "Pages") == 0)
{
inh = fz_dictgets(obj, "Resources");
if (inh) inherit.resources = inh;
inh = fz_dictgets(obj, "MediaBox");
if (inh) inherit.mediabox = inh;
inh = fz_dictgets(obj, "CropBox");
if (inh) inherit.cropbox = inh;
inh = fz_dictgets(obj, "Rotate");
if (inh) inherit.rotate = inh;
kids = fz_dictgets(obj, "Kids");
error = pdf_resolve(&kids, xref);
if (error)
return fz_rethrow(error, "cannot resolve /Kids");
pdf_logpage("subtree %d pages, %d %d {\n",
fz_arraylen(kids), ref->u.r.oid, ref->u.r.gid);
for (i = 0; i < fz_arraylen(kids); i++)
{
kref = fz_arrayget(kids, i);
error = pdf_loadindirect(&kobj, xref, kref);
if (error) { fz_dropobj(kids); return fz_rethrow(error, "cannot load kid"); }
if (kobj == obj)
{
/* prevent infinite recursion possible in maliciously crafted PDFs */
fz_dropobj(kids);
return fz_throw("corrupted pdf file");
}
error = loadpagetree(xref, pages, inherit, kobj, kref, pagenum);
fz_dropobj(kobj);
if (error) { fz_dropobj(kids); return fz_rethrow(error, "cannot load subtree"); }
}
fz_dropobj(kids);
pdf_logpage("}\n");
}
else
return fz_throw("pagetree node has unexpected type %s", fz_toname(type));
return fz_okay;
}
void
cleanmain(int argc, char **argv)
{
int doencrypt = 0;
int dogarbage = 0;
int doexpand = 0;
pdf_crypt *encrypt = nil;
char *infile;
char *outfile = "out.pdf";
char *userpw = "";
char *ownerpw = "";
unsigned perms = 0xfffff0c0; /* nothing allowed */
int keylen = 40;
char *password = "";
fz_error *error;
int c;
while ((c = getopt(argc, argv, "d:egn:o:p:u:x")) != -1)
{
switch (c)
{
case 'p':
/* see TABLE 3.15 User access permissions */
perms = 0xfffff0c0;
if (strchr(optarg, 'p')) /* print */
perms |= (1 << 2) | (1 << 11);
if (strchr(optarg, 'm')) /* modify */
perms |= (1 << 3) | (1 << 10);
if (strchr(optarg, 'c')) /* copy */
perms |= (1 << 4) | (1 << 9);
if (strchr(optarg, 'a')) /* annotate / forms */
perms |= (1 << 5) | (1 << 8);
break;
case 'd': password = optarg; break;
case 'e': doencrypt ++; break;
case 'g': dogarbage ++; break;
case 'n': keylen = atoi(optarg); break;
case 'o': ownerpw = optarg; break;
case 'u': userpw = optarg; break;
case 'x': doexpand ++; break;
default: cleanusage(); break;
}
}
if (argc - optind < 1)
cleanusage();
infile = argv[optind++];
if (argc - optind > 0)
outfile = argv[optind++];
opensrc(infile, password, 0);
if (doencrypt)
{
fz_obj *id = fz_dictgets(src->trailer, "ID");
if (!id)
{
error = fz_packobj(&id, "[(ABCDEFGHIJKLMNOP)(ABCDEFGHIJKLMNOP)]");
if (error)
die(error);
}
else
fz_keepobj(id);
error = pdf_newencrypt(&encrypt, userpw, ownerpw, perms, keylen, id);
if (error)
die(error);
fz_dropobj(id);
}
if (doexpand)
cleanexpand();
if (dogarbage)
{
preloadobjstms();
pdf_garbagecollect(src);
}
error = pdf_savexref(src, outfile, encrypt);
if (error)
die(error);
if (encrypt)
pdf_dropcrypt(encrypt);
pdf_closexref(src);
}
static fz_error
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 oid, 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, &oid, &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");
}
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 (oid < 0 || oid >= xref->len)
{
if (oid == xref->len && oid < 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)", oid, gen, xref->len - 1);
xref->len ++;
}
else
{
fz_dropobj(trailer);
return fz_throw("object id (%d %d R) out of range (0..%d)", oid, gen, xref->len - 1);
}
}
xref->table[oid].type = 'n';
xref->table[oid].gen = gen;
xref->table[oid].obj = fz_keepobj(trailer);
xref->table[oid].stmofs = stmofs;
xref->table[oid].ofs = 0;
obj = fz_dictgets(trailer, "W");
if (!obj) {
fz_dropobj(trailer);
return fz_throw("xref stream missing W entry");
}
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_openstream(&stm, xref, oid, gen);
if (error)
{
fz_dropobj(trailer);
return fz_rethrow(error, "cannot open compressed xref stream");
}
if (!index)
{
error = readnewxrefsection(xref, stm, 0, size, w0, w1, w2);
if (error)
{
fz_dropstream(stm);
fz_dropobj(trailer);
return fz_rethrow(error, "cannot read xref stream");
}
}
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 = readnewxrefsection(xref, stm, i0, i1, w0, w1, w2);
if (error)
{
fz_dropstream(stm);
fz_dropobj(trailer);
return fz_rethrow(error, "cannot read xref stream section");
}
}
}
fz_dropstream(stm);
*trailerp = trailer;
return fz_okay;
}
fz_error
pdf_loadpage(pdf_page **pagep, pdf_xref *xref, fz_obj *dict)
{
fz_error error;
pdf_page *page;
fz_obj *obj;
fz_bbox bbox;
pdf_logpage("load page {\n");
// TODO: move this to a more appropriate place
/* Ensure that we have a store for resource objects */
if (!xref->store)
xref->store = pdf_newstore();
page = fz_malloc(sizeof(pdf_page));
page->resources = nil;
page->contents = nil;
page->transparency = 0;
page->list = nil;
page->text = nil;
page->links = nil;
page->annots = nil;
obj = fz_dictgets(dict, "MediaBox");
bbox = fz_roundrect(pdf_torect(obj));
if (fz_isemptyrect(pdf_torect(obj)))
{
fz_warn("cannot find page bounds, guessing page bounds.");
bbox.x0 = 0;
bbox.y0 = 0;
bbox.x1 = 612;
bbox.y1 = 792;
}
obj = fz_dictgets(dict, "CropBox");
if (fz_isarray(obj))
{
fz_bbox cropbox = fz_roundrect(pdf_torect(obj));
bbox = fz_intersectbbox(bbox, cropbox);
}
page->mediabox.x0 = MIN(bbox.x0, bbox.x1);
page->mediabox.y0 = MIN(bbox.y0, bbox.y1);
page->mediabox.x1 = MAX(bbox.x0, bbox.x1);
page->mediabox.y1 = MAX(bbox.y0, bbox.y1);
if (page->mediabox.x1 - page->mediabox.x0 < 1 || page->mediabox.y1 - page->mediabox.y0 < 1)
return fz_throw("invalid page size");
page->rotate = fz_toint(fz_dictgets(dict, "Rotate"));
pdf_logpage("bbox [%d %d %d %d]\n", bbox.x0, bbox.y0, bbox.x1, bbox.y1);
pdf_logpage("rotate %d\n", page->rotate);
obj = fz_dictgets(dict, "Annots");
if (obj)
{
pdf_loadlinks(&page->links, xref, obj);
pdf_loadannots(&page->annots, xref, obj);
}
page->resources = fz_dictgets(dict, "Resources");
if (page->resources)
fz_keepobj(page->resources);
obj = fz_dictgets(dict, "Contents");
error = pdf_loadpagecontents(&page->contents, xref, obj);
if (error)
{
pdf_freepage(page);
return fz_rethrow(error, "cannot load page contents (%d %d R)", fz_tonum(obj), fz_togen(obj));
}
if (page->resources && pdf_resourcesuseblending(page->resources))
page->transparency = 1;
pdf_logpage("} %p\n", page);
*pagep = page;
return fz_okay;
}
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
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;
}
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;
}
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);
}
void pdfapp_open(pdfapp_t *app, char *filename)
{
#if 0
fz_error error;
fz_obj *obj;
char *password = "";
#else
fz_error error;
fz_stream *file;
char *password = "";
fz_obj *obj;
fz_obj *info;
int fd;
fd = open(filename, O_BINARY | O_RDONLY, 0666);
if (fd < 0)
fprintf(stderr, "error, file %s does not exist\n", filename);
#endif
/*
* Open PDF and load xref table
*/
app->filename = filename;
#if 0
app->xref = pdf_newxref();
error = pdf_loadxref(app->xref, filename);
if (error)
{
fz_catch(error, "trying to air");
error = pdf_repairxref(app->xref, filename);
if (error)
pdfapp_error(app, error);
}
error = pdf_decryptxref(app->xref);
if (error)
pdfapp_error(app, error);
#else
file = fz_open_fd(fd);
error = pdf_open_xref_with_stream(&app->xref, file, NULL);
if (error)
pdfapp_error(app, fz_rethrow(error, "cannot open document '%s'", filename));
fz_close(file);
#endif
/*
* Handle encrypted PDF files
*/
if (pdf_needs_password(app->xref))
{
int okay = pdf_authenticate_password(app->xref, password);
while (!okay)
{
//password = winpassword(app, filename);
if (!password)
exit(1);
okay = pdf_authenticate_password(app->xref, password);
if (!okay)
pdfapp_warn(app, "Invalid password.");
}
}
/*
* Load meta information
* TODO: move this into mupdf library
*/
#if 0
obj = fz_dictgets(app->xref->trailer, "Root");
app->xref->root = fz_resolveindirect(obj);
if (!app->xref->root)
pdfapp_error(app, fz_throw("syntaxerror: missing Root object"));
fz_keepobj(app->xref->root);
obj = fz_dictgets(app->xref->trailer, "Info");
app->xref->info = fz_resolveindirect(obj);
if (!app->xref->info)
pdfapp_warn(app, "Could not load PDF meta information.");
if (app->xref->info)
fz_keepobj(app->xref->info);
/*app->outline = pdf_loadoutline(app->xref);*/
app->doctitle = filename;
if (strrchr(app->doctitle, '\\'))
app->doctitle = strrchr(app->doctitle, '\\') + 1;
if (strrchr(app->doctitle, '/'))
app->doctitle = strrchr(app->doctitle, '/') + 1;
if (app->xref->info)
{
obj = fz_dictgets(app->xref->info, "Title");
if (obj)
{
app->doctitle = pdf_toutf8(obj);
}
}
#else
app->outline = pdf_load_outline(app->xref);
app->doctitle = filename;
if (strrchr(app->doctitle, '\\'))
app->doctitle = strrchr(app->doctitle, '\\') + 1;
if (strrchr(app->doctitle, '/'))
app->doctitle = strrchr(app->doctitle, '/') + 1;
info = fz_dict_gets(app->xref->trailer, "Info");
if (info)
{
obj = fz_dict_gets(info, "Title");
if (obj)
app->doctitle = pdf_to_utf8(obj);
}
#endif
/*
* Start at first page
*/
#if 0
app->pagecount = pdf_getpagecount(app->xref);
#else
error = pdf_load_page_tree(app->xref);
if (error)
pdfapp_error(app, fz_rethrow(error, "cannot load page tree"));
app->pagecount = pdf_count_pages(app->xref);
#endif
app->rotate = 0;
}
Epdf_Document* epdf_document_new(const char* filename)
{
Epdf_Document* doc;
fz_error error;
doc = (Epdf_Document*)malloc(sizeof(Epdf_Document));
if(!doc)
return NULL;
doc->xref = NULL;
doc->outline = NULL;
doc->rast = NULL;
doc->pagecount = 0;
doc->zoom = 1.0;
doc->rotate = 0;
doc->locked = false;
error = fz_newrenderer(&doc->rast, pdf_devicergb, 0, 1024 * 512);
if(error)
return pdfdoc_error(doc);
fz_obj* obj;
// Open PDF and load xref table
doc->xref = pdf_newxref();
if(!doc->xref)
return pdfdoc_error(doc);
error = pdf_loadxref(doc->xref, (char*)filename);
if(error)
{
fz_catch(error, "trying to repair");
fprintf(stderr, "There was a problem with file \"%s\".\nIt may be corrupted or generated by faulty software.\nTrying to repair the file.\n", filename);
error = pdf_repairxref(doc->xref, (char*)filename);
if(error)
return pdfdoc_error(doc);
}
error = pdf_decryptxref(doc->xref);
if(error)
pdfdoc_error(doc);
// Handle encrypted PDF files
if(pdf_needspassword(doc->xref))
doc->locked = true;
/*
* Load meta information
* TODO: move this into mupdf library
*/
obj = fz_dictgets(doc->xref->trailer, (char*)"Root");
doc->xref->root = fz_resolveindirect(obj);
if(!doc->xref->root)
{
fz_throw("syntaxerror: missing Root object");
pdfdoc_error(doc);
}
fz_keepobj(doc->xref->root);
obj = fz_dictgets(doc->xref->trailer, "Info");
doc->xref->info = fz_resolveindirect(obj);
if(!doc->xref->info)
fprintf(stderr, "Could not load PDF meta information.\n");
if(doc->xref->info)
fz_keepobj(doc->xref->info);
doc->outline = pdf_loadoutline(doc->xref);
doc->filename = strdup(filename);
if(doc->xref->info)
{
obj = fz_dictgets(doc->xref->info, "Title");
if(obj)
doc->doctitle = pdf_toutf8(obj);
}
/*
* Start at first page
*/
doc->pagecount = pdf_getpagecount(doc->xref);
if(doc->zoom < 0.1)
doc->zoom = 0.1;
if(doc->zoom > 3.0)
doc->zoom = 3.0;
return doc;
}
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 ++;
}
}
