pdf_obj *
pdf_parse_stm_obj(pdf_document *doc, fz_stream *file, pdf_lexbuf *buf)
{
pdf_token tok;
fz_context *ctx = file->ctx;
tok = pdf_lex(file, buf);
switch (tok)
{
case PDF_TOK_OPEN_ARRAY:
return pdf_parse_array(doc, file, buf);
case PDF_TOK_OPEN_DICT:
return pdf_parse_dict(doc, file, buf);
case PDF_TOK_NAME: return pdf_new_name(doc, buf->scratch); break;
case PDF_TOK_REAL: return pdf_new_real(doc, buf->f); break;
case PDF_TOK_STRING: return pdf_new_string(doc, buf->scratch, buf->len); break;
case PDF_TOK_TRUE: return pdf_new_bool(doc, 1); break;
case PDF_TOK_FALSE: return pdf_new_bool(doc, 0); break;
case PDF_TOK_NULL: return pdf_new_null(doc); break;
case PDF_TOK_INT: return pdf_new_int(doc, buf->i); break;
default: fz_throw(ctx, FZ_ERROR_GENERIC, "unknown token in object stream");
}
return NULL; /* Stupid MSVC */
}
pdf_obj *
pdf_parse_stm_obj(pdf_document *xref, fz_stream *file, pdf_lexbuf *buf)
{
int tok;
fz_context *ctx = file->ctx;
tok = pdf_lex(file, buf);
/* RJW: "cannot parse token in object stream") */
switch (tok)
{
case PDF_TOK_OPEN_ARRAY:
return pdf_parse_array(xref, file, buf);
/* RJW: "cannot parse object stream" */
case PDF_TOK_OPEN_DICT:
return pdf_parse_dict(xref, file, buf);
/* RJW: "cannot parse object stream" */
case PDF_TOK_NAME:
return fz_new_name(ctx, buf->scratch);
break;
case PDF_TOK_REAL:
return pdf_new_real(ctx, buf->f);
break;
case PDF_TOK_STRING:
return pdf_new_string(ctx, buf->scratch, buf->len);
break;
case PDF_TOK_TRUE:
return pdf_new_bool(ctx, 1);
break;
case PDF_TOK_FALSE:
return pdf_new_bool(ctx, 0);
break;
case PDF_TOK_NULL:
return pdf_new_null(ctx);
break;
case PDF_TOK_INT:
return pdf_new_int(ctx, buf->i);
break;
default:
fz_throw(ctx, "unknown token in object stream");
}
return NULL; /* Stupid MSVC */
}
pdf_obj *
pdf_parse_ind_obj(pdf_document *xref,
fz_stream *file, pdf_lexbuf *buf,
int *onum, int *ogen, int *ostmofs)
{
pdf_obj *obj = NULL;
int num = 0, gen = 0, stm_ofs;
int tok;
int a, b;
fz_context *ctx = file->ctx;
fz_var(obj);
tok = pdf_lex(file, buf);
/* RJW: cannot parse indirect object (%d %d R)", num, gen */
if (tok != PDF_TOK_INT)
fz_throw(ctx, "expected object number (%d %d R)", num, gen);
num = buf->i;
tok = pdf_lex(file, buf);
/* RJW: "cannot parse indirect object (%d %d R)", num, gen */
if (tok != PDF_TOK_INT)
fz_throw(ctx, "expected generation number (%d %d R)", num, gen);
gen = buf->i;
tok = pdf_lex(file, buf);
/* RJW: "cannot parse indirect object (%d %d R)", num, gen */
if (tok != PDF_TOK_OBJ)
fz_throw(ctx, "expected 'obj' keyword (%d %d R)", num, gen);
tok = pdf_lex(file, buf);
/* RJW: "cannot parse indirect object (%d %d R)", num, gen */
switch (tok)
{
case PDF_TOK_OPEN_ARRAY:
obj = pdf_parse_array(xref, file, buf);
/* RJW: "cannot parse indirect object (%d %d R)", num, gen */
break;
case PDF_TOK_OPEN_DICT:
obj = pdf_parse_dict(xref, file, buf);
/* RJW: "cannot parse indirect object (%d %d R)", num, gen */
break;
case PDF_TOK_NAME:
obj = fz_new_name(ctx, buf->scratch);
break;
case PDF_TOK_REAL:
obj = pdf_new_real(ctx, buf->f);
break;
case PDF_TOK_STRING:
obj = pdf_new_string(ctx, buf->scratch, buf->len);
break;
case PDF_TOK_TRUE:
obj = pdf_new_bool(ctx, 1);
break;
case PDF_TOK_FALSE:
obj = pdf_new_bool(ctx, 0);
break;
case PDF_TOK_NULL:
obj = pdf_new_null(ctx);
break;
case PDF_TOK_INT:
a = buf->i;
tok = pdf_lex(file, buf);
/* "cannot parse indirect object (%d %d R)", num, gen */
if (tok == PDF_TOK_STREAM || tok == PDF_TOK_ENDOBJ)
{
obj = pdf_new_int(ctx, a);
goto skip;
}
if (tok == PDF_TOK_INT)
{
b = buf->i;
tok = pdf_lex(file, buf);
/* RJW: "cannot parse indirect object (%d %d R)", num, gen); */
if (tok == PDF_TOK_R)
{
obj = pdf_new_indirect(ctx, a, b, xref);
break;
}
}
fz_throw(ctx, "expected 'R' keyword (%d %d R)", num, gen);
case PDF_TOK_ENDOBJ:
obj = pdf_new_null(ctx);
goto skip;
default:
fz_throw(ctx, "syntax error in object (%d %d R)", num, gen);
}
fz_try(ctx)
{
tok = pdf_lex(file, buf);
}
fz_catch(ctx)
{
pdf_drop_obj(obj);
fz_throw(ctx, "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(ctx, "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(ctx, "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;
return obj;
}
pdf_obj *
pdf_parse_dict(pdf_document *xref, fz_stream *file, pdf_lexbuf *buf)
{
pdf_obj *dict;
pdf_obj *key = NULL;
pdf_obj *val = NULL;
int tok;
int a, b;
fz_context *ctx = file->ctx;
dict = pdf_new_dict(ctx, 8);
fz_var(key);
fz_var(val);
fz_try(ctx)
{
while (1)
{
tok = pdf_lex(file, buf);
skip:
if (tok == PDF_TOK_CLOSE_DICT)
break;
/* for BI .. ID .. EI in content streams */
if (tok == PDF_TOK_KEYWORD && !strcmp(buf->scratch, "ID"))
break;
if (tok != PDF_TOK_NAME)
fz_throw(ctx, "invalid key in dict");
key = fz_new_name(ctx, buf->scratch);
tok = pdf_lex(file, buf);
switch (tok)
{
case PDF_TOK_OPEN_ARRAY:
/* cf. http://code.google.com/p/sumatrapdf/issues/detail?id=1643 */
fz_try(ctx)
{
val = pdf_parse_array(xref, file, buf);
}
fz_catch(ctx)
{
fz_warn(ctx, "ignoring broken array for '%s'", pdf_to_name(key));
pdf_drop_obj(key);
val = key = NULL;
do
tok = pdf_lex(file, buf);
while (tok != PDF_TOK_CLOSE_DICT && tok != PDF_TOK_CLOSE_ARRAY &&
tok != PDF_TOK_EOF && tok != PDF_TOK_OPEN_ARRAY && tok != PDF_TOK_OPEN_DICT);
if (tok == PDF_TOK_CLOSE_DICT)
goto skip;
if (tok == PDF_TOK_CLOSE_ARRAY)
continue;
fz_throw(ctx, "cannot make sense of broken array after all");
}
break;
case PDF_TOK_OPEN_DICT:
val = pdf_parse_dict(xref, file, buf);
break;
case PDF_TOK_NAME:
val = fz_new_name(ctx, buf->scratch);
break;
case PDF_TOK_REAL:
val = pdf_new_real(ctx, buf->f);
break;
case PDF_TOK_STRING:
val = pdf_new_string(ctx, buf->scratch, buf->len);
break;
case PDF_TOK_TRUE:
val = pdf_new_bool(ctx, 1);
break;
case PDF_TOK_FALSE:
val = pdf_new_bool(ctx, 0);
break;
case PDF_TOK_NULL:
val = pdf_new_null(ctx);
break;
case PDF_TOK_INT:
/* 64-bit to allow for numbers > INT_MAX and overflow */
a = buf->i;
tok = pdf_lex(file, buf);
if (tok == PDF_TOK_CLOSE_DICT || tok == PDF_TOK_NAME ||
(tok == PDF_TOK_KEYWORD && !strcmp(buf->scratch, "ID")))
{
val = pdf_new_int(ctx, a);
fz_dict_put(dict, key, val);
pdf_drop_obj(val);
val = NULL;
pdf_drop_obj(key);
key = NULL;
goto skip;
}
if (tok == PDF_TOK_INT)
{
b = buf->i;
tok = pdf_lex(file, buf);
if (tok == PDF_TOK_R)
{
val = pdf_new_indirect(ctx, a, b, xref);
break;
}
}
fz_throw(ctx, "invalid indirect reference in dict");
default:
fz_throw(ctx, "unknown token in dict");
}
fz_dict_put(dict, key, val);
pdf_drop_obj(val);
val = NULL;
pdf_drop_obj(key);
key = NULL;
}
}
fz_catch(ctx)
{
pdf_drop_obj(dict);
pdf_drop_obj(key);
pdf_drop_obj(val);
fz_throw(ctx, "cannot parse dict");
}
return dict;
}
pdf_obj *
pdf_parse_array(pdf_document *xref, fz_stream *file, pdf_lexbuf *buf)
{
pdf_obj *ary = NULL;
pdf_obj *obj = NULL;
int a = 0, b = 0, n = 0;
int tok;
fz_context *ctx = file->ctx;
pdf_obj *op;
fz_var(obj);
ary = pdf_new_array(ctx, 4);
fz_try(ctx)
{
while (1)
{
tok = pdf_lex(file, buf);
if (tok != PDF_TOK_INT && tok != PDF_TOK_R)
{
if (n > 0)
{
obj = pdf_new_int(ctx, a);
pdf_array_push(ary, obj);
pdf_drop_obj(obj);
obj = NULL;
}
if (n > 1)
{
obj = pdf_new_int(ctx, b);
pdf_array_push(ary, obj);
pdf_drop_obj(obj);
obj = NULL;
}
n = 0;
}
if (tok == PDF_TOK_INT && n == 2)
{
obj = pdf_new_int(ctx, a);
pdf_array_push(ary, obj);
pdf_drop_obj(obj);
obj = NULL;
a = b;
n --;
}
switch (tok)
{
case PDF_TOK_CLOSE_ARRAY:
op = ary;
goto end;
case PDF_TOK_INT:
if (n == 0)
a = buf->i;
if (n == 1)
b = buf->i;
n ++;
break;
case PDF_TOK_R:
if (n != 2)
fz_throw(ctx, "cannot parse indirect reference in array");
obj = pdf_new_indirect(ctx, a, b, xref);
pdf_array_push(ary, obj);
pdf_drop_obj(obj);
obj = NULL;
n = 0;
break;
case PDF_TOK_OPEN_ARRAY:
obj = pdf_parse_array(xref, file, buf);
pdf_array_push(ary, obj);
pdf_drop_obj(obj);
obj = NULL;
break;
case PDF_TOK_OPEN_DICT:
obj = pdf_parse_dict(xref, file, buf);
pdf_array_push(ary, obj);
pdf_drop_obj(obj);
obj = NULL;
break;
case PDF_TOK_NAME:
obj = fz_new_name(ctx, buf->scratch);
pdf_array_push(ary, obj);
pdf_drop_obj(obj);
obj = NULL;
break;
case PDF_TOK_REAL:
obj = pdf_new_real(ctx, buf->f);
pdf_array_push(ary, obj);
pdf_drop_obj(obj);
obj = NULL;
break;
case PDF_TOK_STRING:
obj = pdf_new_string(ctx, buf->scratch, buf->len);
pdf_array_push(ary, obj);
pdf_drop_obj(obj);
obj = NULL;
break;
case PDF_TOK_TRUE:
obj = pdf_new_bool(ctx, 1);
pdf_array_push(ary, obj);
pdf_drop_obj(obj);
obj = NULL;
break;
case PDF_TOK_FALSE:
obj = pdf_new_bool(ctx, 0);
pdf_array_push(ary, obj);
pdf_drop_obj(obj);
obj = NULL;
break;
case PDF_TOK_NULL:
obj = pdf_new_null(ctx);
pdf_array_push(ary, obj);
pdf_drop_obj(obj);
obj = NULL;
break;
default:
fz_throw(ctx, "cannot parse token in array");
}
}
end:
{}
}
fz_catch(ctx)
{
pdf_drop_obj(obj);
pdf_drop_obj(ary);
fz_throw(ctx, "cannot parse array");
}
return op;
}
/*
When resetting or submitting a form, the fields to act upon are defined
by an array of either field references or field names, plus a flag determining
whether to act upon the fields in the array, or all fields other than those in
the array. specified_fields interprets this information and produces the array
of fields to be acted upon.
*/
static pdf_obj *specified_fields(pdf_document *doc, pdf_obj *fields, int exclude)
{
fz_context *ctx = doc->ctx;
pdf_obj *form = pdf_dict_getp(pdf_trailer(doc), "Root/AcroForm/Fields");
int i, n;
pdf_obj *result = pdf_new_array(doc, 0);
pdf_obj *nil = NULL;
fz_var(nil);
fz_try(ctx)
{
/* The 'fields' array not being present signals that all fields
* should be acted upon, so handle it using the exclude case - excluding none */
if (exclude || !fields)
{
/* mark the fields we don't want to act upon */
nil = pdf_new_null(doc);
n = pdf_array_len(fields);
for (i = 0; i < n; i++)
{
pdf_obj *field = pdf_array_get(fields, i);
if (pdf_is_string(field))
field = pdf_lookup_field(form, pdf_to_str_buf(field));
if (field)
pdf_dict_puts(field, "Exclude", nil);
}
/* Act upon all unmarked fields */
n = pdf_array_len(form);
for (i = 0; i < n; i++)
add_field_hierarchy_to_array(result, pdf_array_get(form, i));
/* Unmark the marked fields */
n = pdf_array_len(fields);
for (i = 0; i < n; i++)
{
pdf_obj *field = pdf_array_get(fields, i);
if (pdf_is_string(field))
field = pdf_lookup_field(form, pdf_to_str_buf(field));
if (field)
pdf_dict_dels(field, "Exclude");
}
}
else
{
n = pdf_array_len(fields);
for (i = 0; i < n; i++)
{
pdf_obj *field = pdf_array_get(fields, i);
if (pdf_is_string(field))
field = pdf_lookup_field(form, pdf_to_str_buf(field));
if (field)
add_field_hierarchy_to_array(result, field);
}
}
}
fz_always(ctx)
{
pdf_drop_obj(nil);
}
fz_catch(ctx)
{
pdf_drop_obj(result);
fz_rethrow(ctx);
}
return result;
}
pdf_obj *
pdf_parse_dict(pdf_document *doc, fz_stream *file, pdf_lexbuf *buf)
{
pdf_obj *dict;
pdf_obj *key = NULL;
pdf_obj *val = NULL;
pdf_token tok;
int a, b;
fz_context *ctx = file->ctx;
dict = pdf_new_dict(doc, 8);
fz_var(key);
fz_var(val);
fz_try(ctx)
{
while (1)
{
tok = pdf_lex(file, buf);
skip:
if (tok == PDF_TOK_CLOSE_DICT)
break;
/* for BI .. ID .. EI in content streams */
if (tok == PDF_TOK_KEYWORD && !strcmp(buf->scratch, "ID"))
break;
if (tok != PDF_TOK_NAME)
fz_throw(ctx, FZ_ERROR_GENERIC, "invalid key in dict");
key = pdf_new_name(doc, buf->scratch);
tok = pdf_lex(file, buf);
switch (tok)
{
case PDF_TOK_OPEN_ARRAY:
val = pdf_parse_array(doc, file, buf);
break;
case PDF_TOK_OPEN_DICT:
val = pdf_parse_dict(doc, file, buf);
break;
case PDF_TOK_NAME: val = pdf_new_name(doc, buf->scratch); break;
case PDF_TOK_REAL: val = pdf_new_real(doc, buf->f); break;
case PDF_TOK_STRING: val = pdf_new_string(doc, buf->scratch, buf->len); break;
case PDF_TOK_TRUE: val = pdf_new_bool(doc, 1); break;
case PDF_TOK_FALSE: val = pdf_new_bool(doc, 0); break;
case PDF_TOK_NULL: val = pdf_new_null(doc); break;
case PDF_TOK_INT:
/* 64-bit to allow for numbers > INT_MAX and overflow */
a = buf->i;
tok = pdf_lex(file, buf);
if (tok == PDF_TOK_CLOSE_DICT || tok == PDF_TOK_NAME ||
(tok == PDF_TOK_KEYWORD && !strcmp(buf->scratch, "ID")))
{
val = pdf_new_int(doc, a);
pdf_dict_put(dict, key, val);
pdf_drop_obj(val);
val = NULL;
pdf_drop_obj(key);
key = NULL;
goto skip;
}
if (tok == PDF_TOK_INT)
{
b = buf->i;
tok = pdf_lex(file, buf);
if (tok == PDF_TOK_R)
{
val = pdf_new_indirect(doc, a, b);
break;
}
}
fz_throw(ctx, FZ_ERROR_GENERIC, "invalid indirect reference in dict");
default:
fz_throw(ctx, FZ_ERROR_GENERIC, "unknown token in dict");
}
pdf_dict_put(dict, key, val);
pdf_drop_obj(val);
val = NULL;
pdf_drop_obj(key);
key = NULL;
}
}
fz_catch(ctx)
{
pdf_drop_obj(dict);
pdf_drop_obj(key);
pdf_drop_obj(val);
fz_rethrow_message(ctx, "cannot parse dict");
}
return dict;
}
pdf_obj *
pdf_parse_ind_obj(fz_context *ctx, pdf_document *doc,
fz_stream *file, pdf_lexbuf *buf,
int *onum, int *ogen, fz_off_t *ostmofs, int *try_repair)
{
pdf_obj *obj = NULL;
int num = 0, gen = 0;
fz_off_t stm_ofs;
pdf_token tok;
fz_off_t a, b;
fz_var(obj);
tok = pdf_lex(ctx, file, buf);
if (tok != PDF_TOK_INT)
{
if (try_repair)
*try_repair = 1;
fz_throw(ctx, FZ_ERROR_GENERIC, "expected object number");
}
num = buf->i;
tok = pdf_lex(ctx, file, buf);
if (tok != PDF_TOK_INT)
{
if (try_repair)
*try_repair = 1;
fz_throw(ctx, FZ_ERROR_GENERIC, "expected generation number (%d ? obj)", num);
}
gen = buf->i;
tok = pdf_lex(ctx, file, buf);
if (tok != PDF_TOK_OBJ)
{
if (try_repair)
*try_repair = 1;
fz_throw(ctx, FZ_ERROR_GENERIC, "expected 'obj' keyword (%d %d ?)", num, gen);
}
tok = pdf_lex(ctx, file, buf);
switch (tok)
{
case PDF_TOK_OPEN_ARRAY:
obj = pdf_parse_array(ctx, doc, file, buf);
break;
case PDF_TOK_OPEN_DICT:
obj = pdf_parse_dict(ctx, doc, file, buf);
break;
case PDF_TOK_NAME: obj = pdf_new_name(ctx, doc, buf->scratch); break;
case PDF_TOK_REAL: obj = pdf_new_real(ctx, doc, buf->f); break;
case PDF_TOK_STRING: obj = pdf_new_string(ctx, doc, buf->scratch, buf->len); break;
case PDF_TOK_TRUE: obj = pdf_new_bool(ctx, doc, 1); break;
case PDF_TOK_FALSE: obj = pdf_new_bool(ctx, doc, 0); break;
case PDF_TOK_NULL: obj = pdf_new_null(ctx, doc); break;
case PDF_TOK_INT:
a = buf->i;
tok = pdf_lex(ctx, file, buf);
if (tok == PDF_TOK_STREAM || tok == PDF_TOK_ENDOBJ)
{
obj = pdf_new_int_offset(ctx, doc, a);
goto skip;
}
if (tok == PDF_TOK_INT)
{
b = buf->i;
tok = pdf_lex(ctx, file, buf);
if (tok == PDF_TOK_R)
{
obj = pdf_new_indirect(ctx, doc, a, b);
break;
}
}
fz_throw(ctx, FZ_ERROR_GENERIC, "expected 'R' keyword (%d %d R)", num, gen);
case PDF_TOK_ENDOBJ:
obj = pdf_new_null(ctx, doc);
goto skip;
default:
fz_throw(ctx, FZ_ERROR_GENERIC, "syntax error in object (%d %d R)", num, gen);
}
fz_try(ctx)
{
tok = pdf_lex(ctx, file, buf);
}
fz_catch(ctx)
{
pdf_drop_obj(ctx, obj);
fz_rethrow(ctx);
}
skip:
if (tok == PDF_TOK_STREAM)
{
int c = fz_read_byte(ctx, file);
while (c == ' ')
c = fz_read_byte(ctx, file);
if (c == '\r')
{
c = fz_peek_byte(ctx, file);
if (c != '\n')
fz_warn(ctx, "line feed missing after stream begin marker (%d %d R)", num, gen);
else
fz_read_byte(ctx, file);
}
stm_ofs = fz_tell(ctx, file);
}
else if (tok == PDF_TOK_ENDOBJ)
{
stm_ofs = 0;
}
else
{
fz_warn(ctx, "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;
return obj;
}
