fz_error
pdf_parse_stm_obj(fz_obj **op, pdf_xref *xref, fz_stream *file, char *buf, int cap)
{
fz_error error;
int tok;
int len;
error = pdf_lex(&tok, file, buf, cap, &len);
if (error)
return fz_rethrow(error, "cannot parse token in object stream");
switch (tok)
{
case PDF_TOK_OPEN_ARRAY:
error = pdf_parse_array(op, xref, file, buf, cap);
if (error)
return fz_rethrow(error, "cannot parse object stream");
break;
case PDF_TOK_OPEN_DICT:
error = pdf_parse_dict(op, xref, file, buf, cap);
if (error)
return fz_rethrow(error, "cannot parse object stream");
break;
case PDF_TOK_NAME: *op = fz_new_name(buf); break;
case PDF_TOK_REAL: *op = fz_new_real(fz_atof(buf)); break;
case PDF_TOK_STRING: *op = fz_new_string(buf, len); break;
case PDF_TOK_TRUE: *op = fz_new_bool(1); break;
case PDF_TOK_FALSE: *op = fz_new_bool(0); break;
case PDF_TOK_NULL: *op = fz_new_null(); break;
case PDF_TOK_INT: *op = fz_new_int(atoi(buf)); break;
default: return fz_throw("unknown token in object stream");
}
return fz_okay;
}
fz_error
pdf_cache_object(pdf_xref *xref, int num, int gen)
{
fz_error error;
pdf_xref_entry *x;
int rnum, rgen;
if (num < 0 || num >= xref->len)
return fz_throw("object out of range (%d %d R); xref size %d", num, gen, xref->len);
x = &xref->table[num];
if (x->obj)
return fz_okay;
if (x->type == 'f')
{
x->obj = fz_new_null();
return fz_okay;
}
else if (x->type == 'n')
{
fz_seek(xref->file, x->ofs, 0);
error = pdf_parse_ind_obj(&x->obj, xref, xref->file, xref->scratch, sizeof xref->scratch,
&rnum, &rgen, &x->stm_ofs);
if (error)
return fz_rethrow(error, "cannot parse object (%d %d R)", num, gen);
if (rnum != num)
return fz_throw("found object (%d %d R) instead of (%d %d R)", rnum, rgen, num, gen);
if (xref->crypt)
pdf_crypt_obj(xref->crypt, x->obj, num, gen);
}
else if (x->type == 'o')
{
if (!x->obj)
{
error = pdf_load_obj_stm(xref, x->ofs, 0, xref->scratch, sizeof xref->scratch);
if (error)
return fz_rethrow(error, "cannot load object stream containing object (%d %d R)", num, gen);
if (!x->obj)
return fz_throw("object (%d %d R) was not found in its object stream", num, gen);
}
}
else
{
return fz_throw("assert: corrupt xref struct");
}
return fz_okay;
}
static void addhexfilter(fz_obj *dict)
{
fz_obj *f, *dp, *newf, *newdp;
fz_obj *ahx, *nullobj;
ahx = fz_new_name(ctx, "ASCIIHexDecode");
nullobj = fz_new_null(ctx);
newf = newdp = NULL;
f = fz_dict_gets(dict, "Filter");
dp = fz_dict_gets(dict, "DecodeParms");
if (fz_is_name(f))
{
newf = fz_new_array(ctx, 2);
fz_array_push(newf, ahx);
fz_array_push(newf, f);
f = newf;
if (fz_is_dict(dp))
{
newdp = fz_new_array(ctx, 2);
fz_array_push(newdp, nullobj);
fz_array_push(newdp, dp);
dp = newdp;
}
}
else if (fz_is_array(f))
{
fz_array_insert(f, ahx);
if (fz_is_array(dp))
fz_array_insert(dp, nullobj);
}
else
f = ahx;
fz_dict_puts(dict, "Filter", f);
if (dp)
fz_dict_puts(dict, "DecodeParms", dp);
fz_drop_obj(ahx);
fz_drop_obj(nullobj);
if (newf)
fz_drop_obj(newf);
if (newdp)
fz_drop_obj(newdp);
}
static fz_outline *
pdf_load_outline_imp(pdf_xref *xref, fz_obj *dict)
{
pdf_link *link;
fz_outline *node;
fz_obj *obj;
/* SumatraPDF: prevent potential stack overflow */
fz_outline *prev, *root = NULL;
fz_obj *origDict = dict;
fz_context *ctx = xref->ctx;
if (fz_is_null(ctx, dict))
return NULL;
/* SumatraPDF: prevent cyclic outlines */
do
{
if (fz_dict_gets(ctx, dict, ".seen"))
break;
obj = fz_new_null(ctx);
fz_dict_puts(ctx, dict, ".seen", obj);
fz_drop_obj(ctx, obj);
node = fz_malloc(ctx, sizeof(fz_outline));
node->title = NULL;
node->page = -1;
node->down = NULL;
node->next = NULL;
obj = fz_dict_gets(ctx, dict, "Title");
if (obj)
node->title = pdf_to_utf8(ctx, obj);
/* SumatraPDF: support expansion states */
node->is_open = fz_to_int(ctx, fz_dict_gets(ctx, dict, "Count")) >= 0;
/* SumatraPDF: extended outline actions */
node->data = node->free_data = NULL;
if (fz_dict_gets(ctx, dict, "Dest") || fz_dict_gets(ctx, dict, "A"))
{
link = pdf_load_link(xref, dict);
if (link) /* SumatraPDF: don't crash if it's no link after all */
{
if (link->kind == PDF_LINK_GOTO)
node->page = pdf_find_page_number(xref, fz_array_get(ctx, link->dest, 0));
/* SumatraPDF: extended outline actions */
node->data = link;
node->free_data = pdf_free_link;
}
}
obj = fz_dict_gets(ctx, dict, "First");
if (obj)
node->down = pdf_load_outline_imp(xref, obj);
/* SumatraPDF: prevent potential stack overflow */
if (!root)
prev = root = node;
else
prev = prev->next = node;
dict = fz_dict_gets(ctx, dict, "Next");
} while (dict && !fz_is_null(ctx, dict));
node = root;
/* SumatraPDF: prevent cyclic outlines */
for (dict = origDict; dict && fz_dict_gets(ctx, dict, ".seen"); dict = fz_dict_gets(ctx, dict, "Next"))
fz_dict_dels(ctx, dict, ".seen");
return node;
}
fz_error
pdf_parse_ind_obj(fz_obj **op, pdf_xref *xref,
fz_stream *file, char *buf, int cap,
int *onum, int *ogen, int *ostmofs)
{
fz_error error = fz_okay;
fz_obj *obj = NULL;
int num = 0, gen = 0, stm_ofs;
int tok;
int len;
int a, b;
error = pdf_lex(&tok, file, buf, cap, &len);
if (error)
return fz_rethrow(error, "cannot parse indirect object (%d %d R)", num, gen);
if (tok != PDF_TOK_INT)
return fz_throw("expected object number (%d %d R)", num, gen);
num = atoi(buf);
error = pdf_lex(&tok, file, buf, cap, &len);
if (error)
return fz_rethrow(error, "cannot parse indirect object (%d %d R)", num, gen);
if (tok != PDF_TOK_INT)
return fz_throw("expected generation number (%d %d R)", num, gen);
gen = atoi(buf);
error = pdf_lex(&tok, file, buf, cap, &len);
if (error)
return fz_rethrow(error, "cannot parse indirect object (%d %d R)", num, gen);
if (tok != PDF_TOK_OBJ)
return fz_throw("expected 'obj' keyword (%d %d R)", num, gen);
error = pdf_lex(&tok, file, buf, cap, &len);
if (error)
return fz_rethrow(error, "cannot parse indirect object (%d %d R)", num, gen);
switch (tok)
{
case PDF_TOK_OPEN_ARRAY:
error = pdf_parse_array(&obj, xref, file, buf, cap);
if (error)
return fz_rethrow(error, "cannot parse indirect object (%d %d R)", num, gen);
break;
case PDF_TOK_OPEN_DICT:
error = pdf_parse_dict(&obj, xref, file, buf, cap);
if (error)
return fz_rethrow(error, "cannot parse indirect object (%d %d R)", num, gen);
break;
case PDF_TOK_NAME: obj = fz_new_name(buf); break;
case PDF_TOK_REAL: obj = fz_new_real(fz_atof(buf)); break;
case PDF_TOK_STRING: obj = fz_new_string(buf, len); break;
case PDF_TOK_TRUE: obj = fz_new_bool(1); break;
case PDF_TOK_FALSE: obj = fz_new_bool(0); break;
case PDF_TOK_NULL: obj = fz_new_null(); break;
case PDF_TOK_INT:
a = atoi(buf);
error = pdf_lex(&tok, file, buf, cap, &len);
if (error)
return fz_rethrow(error, "cannot parse indirect object (%d %d R)", num, gen);
if (tok == PDF_TOK_STREAM || tok == PDF_TOK_ENDOBJ)
{
obj = fz_new_int(a);
goto skip;
}
if (tok == PDF_TOK_INT)
{
b = atoi(buf);
error = pdf_lex(&tok, file, buf, cap, &len);
if (error)
return fz_rethrow(error, "cannot parse indirect object (%d %d R)", num, gen);
if (tok == PDF_TOK_R)
{
obj = fz_new_indirect(a, b, xref);
break;
}
}
return fz_throw("expected 'R' keyword (%d %d R)", num, gen);
case PDF_TOK_ENDOBJ:
obj = fz_new_null();
goto skip;
default:
return fz_throw("syntax error in object (%d %d R)", num, gen);
}
error = pdf_lex(&tok, file, buf, cap, &len);
if (error)
{
fz_drop_obj(obj);
return fz_rethrow(error, "cannot parse indirect object (%d %d R)", num, gen);
}
skip:
if (tok == PDF_TOK_STREAM)
{
int c = fz_read_byte(file);
while (c == ' ')
c = fz_read_byte(file);
if (c == '\r')
{
c = fz_peek_byte(file);
if (c != '\n')
fz_warn("line feed missing after stream begin marker (%d %d R)", num, gen);
else
fz_read_byte(file);
}
stm_ofs = fz_tell(file);
}
else if (tok == PDF_TOK_ENDOBJ)
{
stm_ofs = 0;
}
else
{
fz_warn("expected 'endobj' or 'stream' keyword (%d %d R)", num, gen);
stm_ofs = 0;
}
if (onum) *onum = num;
if (ogen) *ogen = gen;
if (ostmofs) *ostmofs = stm_ofs;
*op = obj;
return fz_okay;
}
fz_error
pdf_parse_dict(fz_obj **op, pdf_xref *xref, fz_stream *file, char *buf, int cap)
{
fz_error error = fz_okay;
fz_obj *dict = NULL;
fz_obj *key = NULL;
fz_obj *val = NULL;
int tok;
int len;
int a, b;
dict = fz_new_dict(8);
while (1)
{
error = pdf_lex(&tok, file, buf, cap, &len);
if (error)
{
fz_drop_obj(dict);
return fz_rethrow(error, "cannot parse dict");
}
skip:
if (tok == PDF_TOK_CLOSE_DICT)
{
*op = dict;
return fz_okay;
}
/* for BI .. ID .. EI in content streams */
if (tok == PDF_TOK_KEYWORD && !strcmp(buf, "ID"))
{
*op = dict;
return fz_okay;
}
if (tok != PDF_TOK_NAME)
{
fz_drop_obj(dict);
return fz_throw("invalid key in dict");
}
key = fz_new_name(buf);
error = pdf_lex(&tok, file, buf, cap, &len);
if (error)
{
fz_drop_obj(key);
fz_drop_obj(dict);
return fz_rethrow(error, "cannot parse dict");
}
switch (tok)
{
case PDF_TOK_OPEN_ARRAY:
error = pdf_parse_array(&val, xref, file, buf, cap);
if (error)
{
fz_drop_obj(key);
fz_drop_obj(dict);
return fz_rethrow(error, "cannot parse dict");
}
break;
case PDF_TOK_OPEN_DICT:
error = pdf_parse_dict(&val, xref, file, buf, cap);
if (error)
{
fz_drop_obj(key);
fz_drop_obj(dict);
return fz_rethrow(error, "cannot parse dict");
}
break;
case PDF_TOK_NAME: val = fz_new_name(buf); break;
case PDF_TOK_REAL: val = fz_new_real(fz_atof(buf)); break;
case PDF_TOK_STRING: val = fz_new_string(buf, len); break;
case PDF_TOK_TRUE: val = fz_new_bool(1); break;
case PDF_TOK_FALSE: val = fz_new_bool(0); break;
case PDF_TOK_NULL: val = fz_new_null(); break;
case PDF_TOK_INT:
/* 64-bit to allow for numbers > INT_MAX and overflow */
a = (int) strtoll(buf, 0, 10);
error = pdf_lex(&tok, file, buf, cap, &len);
if (error)
{
fz_drop_obj(key);
fz_drop_obj(dict);
return fz_rethrow(error, "cannot parse dict");
}
if (tok == PDF_TOK_CLOSE_DICT || tok == PDF_TOK_NAME ||
(tok == PDF_TOK_KEYWORD && !strcmp(buf, "ID")))
{
val = fz_new_int(a);
fz_dict_put(dict, key, val);
fz_drop_obj(val);
fz_drop_obj(key);
goto skip;
}
if (tok == PDF_TOK_INT)
{
b = atoi(buf);
error = pdf_lex(&tok, file, buf, cap, &len);
if (error)
{
fz_drop_obj(key);
fz_drop_obj(dict);
return fz_rethrow(error, "cannot parse dict");
}
if (tok == PDF_TOK_R)
{
val = fz_new_indirect(a, b, xref);
break;
}
}
fz_drop_obj(key);
fz_drop_obj(dict);
return fz_throw("invalid indirect reference in dict");
default:
fz_drop_obj(key);
fz_drop_obj(dict);
return fz_throw("unknown token in dict");
}
fz_dict_put(dict, key, val);
fz_drop_obj(val);
fz_drop_obj(key);
}
}
fz_error
pdf_parse_array(fz_obj **op, pdf_xref *xref, fz_stream *file, char *buf, int cap)
{
fz_error error = fz_okay;
fz_obj *ary = NULL;
fz_obj *obj = NULL;
int a = 0, b = 0, n = 0;
int tok;
int len;
ary = fz_new_array(4);
while (1)
{
error = pdf_lex(&tok, file, buf, cap, &len);
if (error)
{
fz_drop_obj(ary);
return fz_rethrow(error, "cannot parse array");
}
if (tok != PDF_TOK_INT && tok != PDF_TOK_R)
{
if (n > 0)
{
obj = fz_new_int(a);
fz_array_push(ary, obj);
fz_drop_obj(obj);
}
if (n > 1)
{
obj = fz_new_int(b);
fz_array_push(ary, obj);
fz_drop_obj(obj);
}
n = 0;
}
if (tok == PDF_TOK_INT && n == 2)
{
obj = fz_new_int(a);
fz_array_push(ary, obj);
fz_drop_obj(obj);
a = b;
n --;
}
switch (tok)
{
case PDF_TOK_CLOSE_ARRAY:
*op = ary;
return fz_okay;
case PDF_TOK_INT:
if (n == 0)
a = atoi(buf);
if (n == 1)
b = atoi(buf);
n ++;
break;
case PDF_TOK_R:
if (n != 2)
{
fz_drop_obj(ary);
return fz_throw("cannot parse indirect reference in array");
}
obj = fz_new_indirect(a, b, xref);
fz_array_push(ary, obj);
fz_drop_obj(obj);
n = 0;
break;
case PDF_TOK_OPEN_ARRAY:
error = pdf_parse_array(&obj, xref, file, buf, cap);
if (error)
{
fz_drop_obj(ary);
return fz_rethrow(error, "cannot parse array");
}
fz_array_push(ary, obj);
fz_drop_obj(obj);
break;
case PDF_TOK_OPEN_DICT:
error = pdf_parse_dict(&obj, xref, file, buf, cap);
if (error)
{
fz_drop_obj(ary);
return fz_rethrow(error, "cannot parse array");
}
fz_array_push(ary, obj);
fz_drop_obj(obj);
break;
case PDF_TOK_NAME:
obj = fz_new_name(buf);
fz_array_push(ary, obj);
fz_drop_obj(obj);
break;
case PDF_TOK_REAL:
obj = fz_new_real(fz_atof(buf));
fz_array_push(ary, obj);
fz_drop_obj(obj);
break;
case PDF_TOK_STRING:
obj = fz_new_string(buf, len);
fz_array_push(ary, obj);
fz_drop_obj(obj);
break;
case PDF_TOK_TRUE:
obj = fz_new_bool(1);
fz_array_push(ary, obj);
fz_drop_obj(obj);
break;
case PDF_TOK_FALSE:
obj = fz_new_bool(0);
fz_array_push(ary, obj);
fz_drop_obj(obj);
break;
case PDF_TOK_NULL:
obj = fz_new_null();
fz_array_push(ary, obj);
fz_drop_obj(obj);
break;
default:
fz_drop_obj(ary);
return fz_throw("cannot parse token in array");
}
}
}