static void renumberobj(fz_obj *obj)
{
int i;
if (fz_isdict(obj))
{
for (i = 0; i < fz_dictlen(obj); i++)
{
fz_obj *key = fz_dictgetkey(obj, i);
fz_obj *val = fz_dictgetval(obj, i);
if (fz_isindirect(val))
{
val = fz_newindirect(renumbermap[fz_tonum(val)], 0, xref);
fz_dictput(obj, key, val);
fz_dropobj(val);
}
else
{
renumberobj(val);
}
}
}
else if (fz_isarray(obj))
{
for (i = 0; i < fz_arraylen(obj); i++)
{
fz_obj *val = fz_arrayget(obj, i);
if (fz_isindirect(val))
{
val = fz_newindirect(renumbermap[fz_tonum(val)], 0, xref);
fz_arrayput(obj, i, val);
fz_dropobj(val);
}
else
{
renumberobj(val);
}
}
}
}
fz_error
pdf_parsedict(fz_obj **op, pdf_xref *xref, fz_stream *file, char *buf, int cap)
{
fz_error error = fz_okay;
fz_obj *dict = nil;
fz_obj *key = nil;
fz_obj *val = nil;
pdf_token_e tok;
int len;
int a, b;
dict = fz_newdict(8);
while (1)
{
error = pdf_lex(&tok, file, buf, cap, &len);
if (error)
{
fz_dropobj(dict);
return fz_rethrow(error, "cannot parse dict");
}
skip:
if (tok == PDF_TCDICT)
{
*op = dict;
return fz_okay;
}
/* for BI .. ID .. EI in content streams */
if (tok == PDF_TKEYWORD && !strcmp(buf, "ID"))
{
*op = dict;
return fz_okay;
}
if (tok != PDF_TNAME)
{
fz_dropobj(dict);
return fz_throw("invalid key in dict");;
}
key = fz_newname(buf);
error = pdf_lex(&tok, file, buf, cap, &len);
if (error)
{
fz_dropobj(dict);
return fz_rethrow(error, "cannot parse dict");
}
switch (tok)
{
case PDF_TOARRAY:
error = pdf_parsearray(&val, xref, file, buf, cap);
if (error)
{
fz_dropobj(key);
fz_dropobj(dict);
return fz_rethrow(error, "cannot parse dict");
}
break;
case PDF_TODICT:
error = pdf_parsedict(&val, xref, file, buf, cap);
if (error)
{
fz_dropobj(key);
fz_dropobj(dict);
return fz_rethrow(error, "cannot parse dict");
}
break;
case PDF_TNAME: val = fz_newname(buf); break;
case PDF_TREAL: val = fz_newreal(atof(buf)); break;
case PDF_TSTRING: val = fz_newstring(buf, len); break;
case PDF_TTRUE: val = fz_newbool(1); break;
case PDF_TFALSE: val = fz_newbool(0); break;
case PDF_TNULL: val = fz_newnull(); break;
case PDF_TINT:
/* 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_dropobj(key);
fz_dropobj(dict);
return fz_rethrow(error, "cannot parse dict");
}
if (tok == PDF_TCDICT || tok == PDF_TNAME ||
(tok == PDF_TKEYWORD && !strcmp(buf, "ID")))
{
val = fz_newint(a);
fz_dictput(dict, key, val);
fz_dropobj(val);
fz_dropobj(key);
goto skip;
}
if (tok == PDF_TINT)
{
b = atoi(buf);
error = pdf_lex(&tok, file, buf, cap, &len);
if (error)
{
fz_dropobj(key);
fz_dropobj(dict);
return fz_rethrow(error, "cannot parse dict");
}
if (tok == PDF_TR)
{
val = fz_newindirect(a, b, xref);
break;
}
}
fz_dropobj(key);
fz_dropobj(dict);
return fz_throw("invalid indirect reference in dict");
default:
return fz_throw("unknown token in dict");
}
fz_dictput(dict, key, val);
fz_dropobj(val);
fz_dropobj(key);
}
}
static fz_error *
loadnametreenode(fz_obj *tree, pdf_xref *xref, fz_obj *node)
{
fz_error *error;
fz_obj *names;
fz_obj *kids;
fz_obj *key;
fz_obj *val;
int i, len;
error = pdf_resolve(&node, xref);
if (error)
return error;
names = fz_dictgets(node, "Names");
if (names)
{
error = pdf_resolve(&names, xref);
if (error)
goto cleanup;
len = fz_arraylen(names) / 2;
for (i = 0; i < len; ++i)
{
key = fz_arrayget(names, i * 2 + 0);
val = fz_arrayget(names, i * 2 + 1);
error = fz_dictput(tree, key, val);
if (error)
{
fz_dropobj(names);
goto cleanup;
}
}
fz_dropobj(names);
}
kids = fz_dictgets(node, "Kids");
if (kids)
{
error = pdf_resolve(&kids, xref);
if (error)
goto cleanup;
len = fz_arraylen(kids);
for (i = 0; i < len; ++i)
{
error = loadnametreenode(tree, xref, fz_arrayget(kids, i));
if (error)
{
fz_dropobj(kids);
goto cleanup;
}
}
fz_dropobj(kids);
}
fz_dropobj(node);
return nil;
cleanup:
fz_dropobj(node);
return error;
}
static fz_error parsedict(fz_obj **obj, pdf_xref *xref, char **sp, struct vap *v)
{
fz_error error = fz_okay;
fz_obj *dict = nil;
fz_obj *key = nil;
fz_obj *val = nil;
char *s = *sp;
error = fz_newdict(&dict, 8);
if (error)
return fz_rethrow(error, "cannot create dict");
s += 2; /* skip "<<" */
while (*s)
{
skipwhite(&s);
/* end-of-dict marker >> */
if (*s == '>')
{
s ++;
if (*s == '>')
{
s ++;
break;
}
error = fz_throw("malformed >> marker");
goto cleanup;
}
/* non-name as key, bail */
if (*s != '/')
{
error = fz_throw("key is not a name");
goto cleanup;
}
error = parsename(&key, &s);
if (error)
{
error = fz_rethrow(error, "cannot parse key");
goto cleanup;
}
skipwhite(&s);
error = parseobj(&val, xref, &s, v);
if (error)
{
error = fz_rethrow(error, "cannot parse value");
goto cleanup;
}
error = fz_dictput(dict, key, val);
if (error)
{
error = fz_rethrow(error, "cannot insert dict entry");
goto cleanup;
}
fz_dropobj(val); val = nil;
fz_dropobj(key); key = nil;
}
*obj = dict;
*sp = s;
return fz_okay;
cleanup:
if (val) fz_dropobj(val);
if (key) fz_dropobj(key);
if (dict) fz_dropobj(dict);
*obj = nil;
*sp = s;
return error; /* already rethrown */
}
fz_error *
pdf_parsedict(fz_obj **op, fz_stream *file, char *buf, int cap)
{
fz_error *error = nil;
fz_obj *dict = nil;
fz_obj *key = nil;
fz_obj *val = nil;
int tok, len;
int a, b;
error = fz_newdict(op, 8);
if (error) return error;
dict = *op;
while (1)
{
tok = pdf_lex(file, buf, cap, &len);
skip:
if (tok == PDF_TCDICT)
return nil;
/* for BI .. ID .. EI in content streams */
if (tok == PDF_TKEYWORD && !strcmp(buf, "ID"))
return nil;
if (tok != PDF_TNAME)
goto cleanup;
error = fz_newname(&key, buf);
if (error) goto cleanup;
tok = pdf_lex(file, buf, cap, &len);
switch (tok)
{
case PDF_TOARRAY: error = pdf_parsearray(&val, file, buf, cap); break;
case PDF_TODICT: error = pdf_parsedict(&val, file, buf, cap); break;
case PDF_TNAME: error = fz_newname(&val, buf); break;
case PDF_TREAL: error = fz_newreal(&val, atof(buf)); break;
case PDF_TSTRING: error = fz_newstring(&val, buf, len); break;
case PDF_TTRUE: error = fz_newbool(&val, 1); break;
case PDF_TFALSE: error = fz_newbool(&val, 0); break;
case PDF_TNULL: error = fz_newnull(&val); break;
case PDF_TINT:
a = atoi(buf);
tok = pdf_lex(file, buf, cap, &len);
if (tok == PDF_TCDICT || tok == PDF_TNAME ||
(tok == PDF_TKEYWORD && !strcmp(buf, "ID")))
{
error = fz_newint(&val, a);
if (error) goto cleanup;
error = fz_dictput(dict, key, val);
if (error) goto cleanup;
fz_dropobj(val);
fz_dropobj(key);
key = val = nil;
goto skip;
}
if (tok == PDF_TINT)
{
b = atoi(buf);
tok = pdf_lex(file, buf, cap, &len);
if (tok == PDF_TR)
{
error = fz_newindirect(&val, a, b);
break;
}
}
goto cleanup;
default:
goto cleanup;
}
if (error) goto cleanup;
error = fz_dictput(dict, key, val);
if (error) goto cleanup;
fz_dropobj(val);
fz_dropobj(key);
key = val = nil;
}
cleanup:
if (key) fz_dropobj(key);
if (val) fz_dropobj(val);
if (dict) fz_dropobj(dict);
if (error) return error;
return fz_throw("syntaxerror: corrupt dictionary");
}
