godot/modules/fbx/fbx_parser/FBXParser.cpp
Rémi Verschelde 5b16020846
Replace remaining uses of NULL with nullptr
Follow-up to #38736 (these uses were likely added after this PR was merged).
2021-04-29 11:53:27 +02:00

1330 lines
38 KiB
C++

/*************************************************************************/
/* FBXParser.cpp */
/*************************************************************************/
/* This file is part of: */
/* GODOT ENGINE */
/* https://godotengine.org */
/*************************************************************************/
/* Copyright (c) 2007-2021 Juan Linietsky, Ariel Manzur. */
/* Copyright (c) 2014-2021 Godot Engine contributors (cf. AUTHORS.md). */
/* */
/* Permission is hereby granted, free of charge, to any person obtaining */
/* a copy of this software and associated documentation files (the */
/* "Software"), to deal in the Software without restriction, including */
/* without limitation the rights to use, copy, modify, merge, publish, */
/* distribute, sublicense, and/or sell copies of the Software, and to */
/* permit persons to whom the Software is furnished to do so, subject to */
/* the following conditions: */
/* */
/* The above copyright notice and this permission notice shall be */
/* included in all copies or substantial portions of the Software. */
/* */
/* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, */
/* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF */
/* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.*/
/* IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY */
/* CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, */
/* TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE */
/* SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. */
/*************************************************************************/
/*
Open Asset Import Library (assimp)
----------------------------------------------------------------------
Copyright (c) 2006-2019, assimp team
All rights reserved.
Redistribution and use of this software in source and binary forms,
with or without modification, are permitted provided that the
following conditions are met:
* Redistributions of source code must retain the above
copyright notice, this list of conditions and the
following disclaimer.
* Redistributions in binary form must reproduce the above
copyright notice, this list of conditions and the
following disclaimer in the documentation and/or other
materials provided with the distribution.
* Neither the name of the assimp team, nor the names of its
contributors may be used to endorse or promote products
derived from this software without specific prior
written permission of the assimp team.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
----------------------------------------------------------------------
*/
/** @file FBXParser.cpp
* @brief Implementation of the FBX parser and the rudimentary DOM that we use
*/
#include <stdlib.h> /* strtol */
#include <zlib.h>
#include "ByteSwapper.h"
#include "FBXParseTools.h"
#include "FBXParser.h"
#include "FBXTokenizer.h"
#include "core/math/math_defs.h"
#include "core/math/transform.h"
#include "core/math/vector3.h"
#include "core/string/print_string.h"
using namespace FBXDocParser;
namespace {
// Initially, we did reinterpret_cast, breaking strict aliasing rules.
// This actually caused trouble on Android, so let's be safe this time.
// https://github.com/assimp/assimp/issues/24
template <typename T>
T SafeParse(const char *data, const char *end) {
// Actual size validation happens during Tokenization so
// this is valid as an assertion.
(void)(end);
//ai_assert(static_cast<size_t>(end - data) >= sizeof(T));
T result = static_cast<T>(0);
::memcpy(&result, data, sizeof(T));
return result;
}
} // namespace
namespace FBXDocParser {
// ------------------------------------------------------------------------------------------------
Element::Element(const TokenPtr key_token, Parser &parser) :
key_token(key_token) {
TokenPtr n = nullptr;
do {
n = parser.AdvanceToNextToken();
if (n == nullptr) {
continue;
}
if (!n) {
print_error("unexpected end of file, expected closing bracket" + String(parser.LastToken()->StringContents().c_str()));
}
if (n && n->Type() == TokenType_DATA) {
tokens.push_back(n);
TokenPtr prev = n;
n = parser.AdvanceToNextToken();
if (n == nullptr) {
break;
}
if (!n) {
print_error("unexpected end of file, expected bracket, comma or key" + String(parser.LastToken()->StringContents().c_str()));
parser.corrupt = true;
return;
}
const TokenType ty = n->Type();
// some exporters are missing a comma on the next line
if (ty == TokenType_DATA && prev->Type() == TokenType_DATA && (n->Line() == prev->Line() + 1)) {
tokens.push_back(n);
continue;
}
if (ty != TokenType_OPEN_BRACKET && ty != TokenType_CLOSE_BRACKET && ty != TokenType_COMMA && ty != TokenType_KEY) {
print_error("unexpected token; expected bracket, comma or key" + String(n->StringContents().c_str()));
parser.corrupt = true;
return;
}
}
if (n && n->Type() == TokenType_OPEN_BRACKET) {
compound = new_Scope(parser);
parser.scopes.push_back(compound);
if (parser.corrupt) {
return;
}
// current token should be a TOK_CLOSE_BRACKET
n = parser.CurrentToken();
if (n && n->Type() != TokenType_CLOSE_BRACKET) {
print_error("expected closing bracket" + String(n->StringContents().c_str()));
parser.corrupt = true;
return;
}
parser.AdvanceToNextToken();
return;
}
} while (n && n->Type() != TokenType_KEY && n->Type() != TokenType_CLOSE_BRACKET);
}
// ------------------------------------------------------------------------------------------------
Element::~Element() {
}
// ------------------------------------------------------------------------------------------------
Scope::Scope(Parser &parser, bool topLevel) {
if (!topLevel) {
TokenPtr t = parser.CurrentToken();
if (t->Type() != TokenType_OPEN_BRACKET) {
print_error("expected open bracket" + String(t->StringContents().c_str()));
parser.corrupt = true;
return;
}
}
TokenPtr n = parser.AdvanceToNextToken();
if (n == nullptr) {
print_error("unexpected end of file");
parser.corrupt = true;
return;
}
// note: empty scopes are allowed
while (n && n->Type() != TokenType_CLOSE_BRACKET) {
if (n->Type() != TokenType_KEY) {
print_error("unexpected token, expected TOK_KEY" + String(n->StringContents().c_str()));
parser.corrupt = true;
return;
}
const std::string str = n->StringContents();
if (parser.corrupt) {
return;
}
// std::multimap<std::string, ElementPtr> (key and value)
elements.insert(ElementMap::value_type(str, new_Element(n, parser)));
// Element() should stop at the next Key token (or right after a Close token)
n = parser.CurrentToken();
if (n == nullptr) {
if (topLevel) {
return;
}
//print_error("unexpected end of file" + String(parser.LastToken()->StringContents().c_str()));
}
}
}
// ------------------------------------------------------------------------------------------------
Scope::~Scope() {
for (ElementMap::value_type &v : elements) {
delete v.second;
v.second = nullptr;
}
elements.clear();
}
// ------------------------------------------------------------------------------------------------
Parser::Parser(const TokenList &tokens, bool is_binary) :
corrupt(false), tokens(tokens), cursor(tokens.begin()), is_binary(is_binary) {
root = new_Scope(*this, true);
scopes.push_back(root);
}
// ------------------------------------------------------------------------------------------------
Parser::~Parser() {
for (ScopePtr scope : scopes) {
delete scope;
scope = nullptr;
}
}
// ------------------------------------------------------------------------------------------------
TokenPtr Parser::AdvanceToNextToken() {
last = current;
if (cursor == tokens.end()) {
current = nullptr;
} else {
current = *cursor++;
}
return current;
}
// ------------------------------------------------------------------------------------------------
TokenPtr Parser::CurrentToken() const {
return current;
}
// ------------------------------------------------------------------------------------------------
TokenPtr Parser::LastToken() const {
return last;
}
// ------------------------------------------------------------------------------------------------
uint64_t ParseTokenAsID(const TokenPtr t, const char *&err_out) {
ERR_FAIL_COND_V_MSG(t == nullptr, 0L, "Invalid token passed to ParseTokenAsID");
err_out = nullptr;
if (t->Type() != TokenType_DATA) {
err_out = "expected TOK_DATA token";
return 0L;
}
if (t->IsBinary()) {
const char *data = t->begin();
if (data[0] != 'L') {
err_out = "failed to parse ID, unexpected data type, expected L(ong) (binary)";
return 0L;
}
uint64_t id = SafeParse<uint64_t>(data + 1, t->end());
return id;
}
// XXX: should use size_t here
unsigned int length = static_cast<unsigned int>(t->end() - t->begin());
//ai_assert(length > 0);
const char *out = nullptr;
bool errored = false;
const uint64_t id = strtoul10_64(t->begin(), errored, &out, &length);
if (errored || out > t->end()) {
err_out = "failed to parse ID (text)";
return 0L;
}
return id;
}
// ------------------------------------------------------------------------------------------------
// wrapper around ParseTokenAsID() with print_error handling
uint64_t ParseTokenAsID(const TokenPtr t) {
const char *err = nullptr;
const uint64_t i = ParseTokenAsID(t, err);
if (err) {
print_error(String(err) + " " + String(t->StringContents().c_str()));
}
return i;
}
// ------------------------------------------------------------------------------------------------
size_t ParseTokenAsDim(const TokenPtr t, const char *&err_out) {
// same as ID parsing, except there is a trailing asterisk
err_out = nullptr;
if (t->Type() != TokenType_DATA) {
err_out = "expected TOK_DATA token";
return 0;
}
if (t->IsBinary()) {
const char *data = t->begin();
if (data[0] != 'L') {
err_out = "failed to parse ID, unexpected data type, expected L(ong) (binary)";
return 0;
}
uint64_t id = SafeParse<uint64_t>(data + 1, t->end());
AI_SWAP8(id);
return static_cast<size_t>(id);
}
if (*t->begin() != '*') {
err_out = "expected asterisk before array dimension";
return 0;
}
// XXX: should use size_t here
unsigned int length = static_cast<unsigned int>(t->end() - t->begin());
if (length == 0) {
err_out = "expected valid integer number after asterisk";
return 0;
}
const char *out = nullptr;
bool errored = false;
const size_t id = static_cast<size_t>(strtoul10_64(t->begin() + 1, errored, &out, &length));
if (errored || out > t->end()) {
print_error("failed to parse id");
err_out = "failed to parse ID";
return 0;
}
return id;
}
// ------------------------------------------------------------------------------------------------
float ParseTokenAsFloat(const TokenPtr t, const char *&err_out) {
err_out = nullptr;
if (t->Type() != TokenType_DATA) {
err_out = "expected TOK_DATA token";
return 0.0f;
}
if (t->IsBinary()) {
const char *data = t->begin();
if (data[0] != 'F' && data[0] != 'D') {
err_out = "failed to parse F(loat) or D(ouble), unexpected data type (binary)";
return 0.0f;
}
if (data[0] == 'F') {
return SafeParse<float>(data + 1, t->end());
} else {
return static_cast<float>(SafeParse<double>(data + 1, t->end()));
}
}
// need to copy the input string to a temporary buffer
// first - next in the fbx token stream comes ',',
// which fast_atof could interpret as decimal point.
#define MAX_FLOAT_LENGTH 31
char temp[MAX_FLOAT_LENGTH + 1];
const size_t length = static_cast<size_t>(t->end() - t->begin());
std::copy(t->begin(), t->end(), temp);
temp[std::min(static_cast<size_t>(MAX_FLOAT_LENGTH), length)] = '\0';
return atof(temp);
}
// ------------------------------------------------------------------------------------------------
int ParseTokenAsInt(const TokenPtr t, const char *&err_out) {
err_out = nullptr;
if (t->Type() != TokenType_DATA) {
err_out = "expected TOK_DATA token";
return 0;
}
// binary files are simple to parse
if (t->IsBinary()) {
const char *data = t->begin();
if (data[0] != 'I') {
err_out = "failed to parse I(nt), unexpected data type (binary)";
return 0;
}
int32_t ival = SafeParse<int32_t>(data + 1, t->end());
AI_SWAP4(ival);
return static_cast<int>(ival);
}
// ASCII files are unsafe.
const size_t length = static_cast<size_t>(t->end() - t->begin());
if (length == 0) {
err_out = "expected valid integer number after asterisk";
ERR_FAIL_V_MSG(0, "expected valid integer number after asterisk");
}
// must not be null for strtol to work
char *out = (char *)t->end();
// string begin, end ptr ref, base 10
const int value = strtol(t->begin(), &out, 10);
if (out == nullptr || out != t->end()) {
err_out = "failed to parse ID";
ERR_FAIL_V_MSG(0, "failed to parse ID");
}
return value;
}
// ------------------------------------------------------------------------------------------------
int64_t ParseTokenAsInt64(const TokenPtr t, const char *&err_out) {
err_out = nullptr;
if (t->Type() != TokenType_DATA) {
err_out = "expected TOK_DATA token";
return 0L;
}
if (t->IsBinary()) {
const char *data = t->begin();
if (data[0] != 'L') {
err_out = "failed to parse Int64, unexpected data type";
return 0L;
}
int64_t id = SafeParse<int64_t>(data + 1, t->end());
AI_SWAP8(id);
return id;
}
// XXX: should use size_t here
unsigned int length = static_cast<unsigned int>(t->end() - t->begin());
//ai_assert(length > 0);
char *out = nullptr;
const int64_t id = strtol(t->begin(), &out, length);
if (out > t->end()) {
err_out = "failed to parse Int64 (text)";
return 0L;
}
return id;
}
// ------------------------------------------------------------------------------------------------
std::string ParseTokenAsString(const TokenPtr t, const char *&err_out) {
err_out = nullptr;
if (t->Type() != TokenType_DATA) {
err_out = "expected TOK_DATA token";
return "";
}
if (t->IsBinary()) {
const char *data = t->begin();
if (data[0] != 'S') {
err_out = "failed to parse String, unexpected data type (binary)";
return "";
}
// read string length
int32_t len = SafeParse<int32_t>(data + 1, t->end());
AI_SWAP4(len);
//ai_assert(t.end() - data == 5 + len);
return std::string(data + 5, len);
}
const size_t length = static_cast<size_t>(t->end() - t->begin());
if (length < 2) {
err_out = "token is too short to hold a string";
return "";
}
const char *s = t->begin(), *e = t->end() - 1;
if (*s != '\"' || *e != '\"') {
err_out = "expected double quoted string";
return "";
}
return std::string(s + 1, length - 2);
}
namespace {
// ------------------------------------------------------------------------------------------------
// read the type code and element count of a binary data array and stop there
void ReadBinaryDataArrayHead(const char *&data, const char *end, char &type, uint32_t &count,
const ElementPtr el) {
TokenPtr token = el->KeyToken();
if (static_cast<size_t>(end - data) < 5) {
print_error("binary data array is too short, need five (5) bytes for type signature and element count: " + String(token->StringContents().c_str()));
}
// data type
type = *data;
// read number of elements
uint32_t len = SafeParse<uint32_t>(data + 1, end);
AI_SWAP4(len);
count = len;
data += 5;
}
// ------------------------------------------------------------------------------------------------
// read binary data array, assume cursor points to the 'compression mode' field (i.e. behind the header)
void ReadBinaryDataArray(char type, uint32_t count, const char *&data, const char *end,
std::vector<char> &buff,
const ElementPtr /*el*/) {
uint32_t encmode = SafeParse<uint32_t>(data, end);
AI_SWAP4(encmode);
data += 4;
// next comes the compressed length
uint32_t comp_len = SafeParse<uint32_t>(data, end);
AI_SWAP4(comp_len);
data += 4;
//ai_assert(data + comp_len == end);
// determine the length of the uncompressed data by looking at the type signature
uint32_t stride = 0;
switch (type) {
case 'f':
case 'i':
stride = 4;
break;
case 'd':
case 'l':
stride = 8;
break;
}
const uint32_t full_length = stride * count;
buff.resize(full_length);
if (encmode == 0) {
//ai_assert(full_length == comp_len);
// plain data, no compression
std::copy(data, end, buff.begin());
} else if (encmode == 1) {
// zlib/deflate, next comes ZIP head (0x78 0x01)
// see http://www.ietf.org/rfc/rfc1950.txt
z_stream zstream;
zstream.opaque = Z_NULL;
zstream.zalloc = Z_NULL;
zstream.zfree = Z_NULL;
zstream.data_type = Z_BINARY;
// http://hewgill.com/journal/entries/349-how-to-decompress-gzip-stream-with-zlib
if (Z_OK != inflateInit(&zstream)) {
print_error("failure initializing zlib");
}
zstream.next_in = reinterpret_cast<Bytef *>(const_cast<char *>(data));
zstream.avail_in = comp_len;
zstream.avail_out = static_cast<uInt>(buff.size());
zstream.next_out = reinterpret_cast<Bytef *>(&*buff.begin());
const int ret = inflate(&zstream, Z_FINISH);
if (ret != Z_STREAM_END && ret != Z_OK) {
print_error("failure decompressing compressed data section");
}
// terminate zlib
inflateEnd(&zstream);
}
#ifdef ASSIMP_BUILD_DEBUG
else {
// runtime check for this happens at tokenization stage
//ai_assert(false);
}
#endif
data += comp_len;
//ai_assert(data == end);
}
} // namespace
// ------------------------------------------------------------------------------------------------
// read an array of float3 tuples
void ParseVectorDataArray(std::vector<Vector3> &out, const ElementPtr el) {
out.resize(0);
const TokenList &tok = el->Tokens();
TokenPtr token = el->KeyToken();
if (tok.empty()) {
print_error("unexpected empty element" + String(token->StringContents().c_str()));
}
if (tok[0]->IsBinary()) {
const char *data = tok[0]->begin(), *end = tok[0]->end();
char type;
uint32_t count;
ReadBinaryDataArrayHead(data, end, type, count, el);
if (count % 3 != 0) {
print_error("number of floats is not a multiple of three (3) (binary)" + String(token->StringContents().c_str()));
}
if (!count) {
return;
}
if (type != 'd' && type != 'f') {
print_error("expected float or double array (binary)" + String(token->StringContents().c_str()));
}
std::vector<char> buff;
ReadBinaryDataArray(type, count, data, end, buff, el);
//ai_assert(data == end);
//ai_assert(buff.size() == count * (type == 'd' ? 8 : 4));
const uint32_t count3 = count / 3;
out.reserve(count3);
if (type == 'd') {
const double *d = reinterpret_cast<const double *>(&buff[0]);
for (unsigned int i = 0; i < count3; ++i, d += 3) {
out.push_back(Vector3(static_cast<real_t>(d[0]),
static_cast<real_t>(d[1]),
static_cast<real_t>(d[2])));
}
// for debugging
/*for ( size_t i = 0; i < out.size(); i++ ) {
aiVector3D vec3( out[ i ] );
std::stringstream stream;
stream << " vec3.x = " << vec3.x << " vec3.y = " << vec3.y << " vec3.z = " << vec3.z << std::endl;
DefaultLogger::get()->info( stream.str() );
}*/
} else if (type == 'f') {
const float *f = reinterpret_cast<const float *>(&buff[0]);
for (unsigned int i = 0; i < count3; ++i, f += 3) {
out.push_back(Vector3(f[0], f[1], f[2]));
}
}
return;
}
const size_t dim = ParseTokenAsDim(tok[0]);
// may throw bad_alloc if the input is rubbish, but this need
// not to be prevented - importing would fail but we wouldn't
// crash since assimp handles this case properly.
out.reserve(dim);
const ScopePtr scope = GetRequiredScope(el);
const ElementPtr a = GetRequiredElement(scope, "a", el);
if (a->Tokens().size() % 3 != 0) {
print_error("number of floats is not a multiple of three (3)" + String(token->StringContents().c_str()));
} else {
for (TokenList::const_iterator it = a->Tokens().begin(), end = a->Tokens().end(); it != end;) {
Vector3 v;
v.x = ParseTokenAsFloat(*it++);
v.y = ParseTokenAsFloat(*it++);
v.z = ParseTokenAsFloat(*it++);
out.push_back(v);
}
}
}
// ------------------------------------------------------------------------------------------------
// read an array of color4 tuples
void ParseVectorDataArray(std::vector<Color> &out, const ElementPtr el) {
out.resize(0);
const TokenList &tok = el->Tokens();
TokenPtr token = el->KeyToken();
if (tok.empty()) {
print_error("unexpected empty element" + String(token->StringContents().c_str()));
}
if (tok[0]->IsBinary()) {
const char *data = tok[0]->begin(), *end = tok[0]->end();
char type;
uint32_t count;
ReadBinaryDataArrayHead(data, end, type, count, el);
if (count % 4 != 0) {
print_error("number of floats is not a multiple of four (4) (binary)" + String(token->StringContents().c_str()));
}
if (!count) {
return;
}
if (type != 'd' && type != 'f') {
print_error("expected float or double array (binary)" + String(token->StringContents().c_str()));
}
std::vector<char> buff;
ReadBinaryDataArray(type, count, data, end, buff, el);
//ai_assert(data == end);
//ai_assert(buff.size() == count * (type == 'd' ? 8 : 4));
const uint32_t count4 = count / 4;
out.reserve(count4);
if (type == 'd') {
const double *d = reinterpret_cast<const double *>(&buff[0]);
for (unsigned int i = 0; i < count4; ++i, d += 4) {
out.push_back(Color(static_cast<float>(d[0]),
static_cast<float>(d[1]),
static_cast<float>(d[2]),
static_cast<float>(d[3])));
}
} else if (type == 'f') {
const float *f = reinterpret_cast<const float *>(&buff[0]);
for (unsigned int i = 0; i < count4; ++i, f += 4) {
out.push_back(Color(f[0], f[1], f[2], f[3]));
}
}
return;
}
const size_t dim = ParseTokenAsDim(tok[0]);
// see notes in ParseVectorDataArray() above
out.reserve(dim);
const ScopePtr scope = GetRequiredScope(el);
const ElementPtr a = GetRequiredElement(scope, "a", el);
if (a->Tokens().size() % 4 != 0) {
print_error("number of floats is not a multiple of four (4)" + String(token->StringContents().c_str()));
}
for (TokenList::const_iterator it = a->Tokens().begin(), end = a->Tokens().end(); it != end;) {
Color v;
v.r = ParseTokenAsFloat(*it++);
v.g = ParseTokenAsFloat(*it++);
v.b = ParseTokenAsFloat(*it++);
v.a = ParseTokenAsFloat(*it++);
out.push_back(v);
}
}
// ------------------------------------------------------------------------------------------------
// read an array of float2 tuples
void ParseVectorDataArray(std::vector<Vector2> &out, const ElementPtr el) {
out.resize(0);
const TokenList &tok = el->Tokens();
TokenPtr token = el->KeyToken();
if (tok.empty()) {
print_error("unexpected empty element" + String(token->StringContents().c_str()));
}
if (tok[0]->IsBinary()) {
const char *data = tok[0]->begin(), *end = tok[0]->end();
char type;
uint32_t count;
ReadBinaryDataArrayHead(data, end, type, count, el);
if (count % 2 != 0) {
print_error("number of floats is not a multiple of two (2) (binary)" + String(token->StringContents().c_str()));
}
if (!count) {
return;
}
if (type != 'd' && type != 'f') {
print_error("expected float or double array (binary)" + String(token->StringContents().c_str()));
}
std::vector<char> buff;
ReadBinaryDataArray(type, count, data, end, buff, el);
//ai_assert(data == end);
//ai_assert(buff.size() == count * (type == 'd' ? 8 : 4));
const uint32_t count2 = count / 2;
out.reserve(count2);
if (type == 'd') {
const double *d = reinterpret_cast<const double *>(&buff[0]);
for (unsigned int i = 0; i < count2; ++i, d += 2) {
out.push_back(Vector2(static_cast<float>(d[0]),
static_cast<float>(d[1])));
}
} else if (type == 'f') {
const float *f = reinterpret_cast<const float *>(&buff[0]);
for (unsigned int i = 0; i < count2; ++i, f += 2) {
out.push_back(Vector2(f[0], f[1]));
}
}
return;
}
const size_t dim = ParseTokenAsDim(tok[0]);
// see notes in ParseVectorDataArray() above
out.reserve(dim);
const ScopePtr scope = GetRequiredScope(el);
const ElementPtr a = GetRequiredElement(scope, "a", el);
if (a->Tokens().size() % 2 != 0) {
print_error("number of floats is not a multiple of two (2)" + String(token->StringContents().c_str()));
} else {
for (TokenList::const_iterator it = a->Tokens().begin(), end = a->Tokens().end(); it != end;) {
Vector2 v;
v.x = ParseTokenAsFloat(*it++);
v.y = ParseTokenAsFloat(*it++);
out.push_back(v);
}
}
}
// ------------------------------------------------------------------------------------------------
// read an array of ints
void ParseVectorDataArray(std::vector<int> &out, const ElementPtr el) {
out.resize(0);
const TokenList &tok = el->Tokens();
TokenPtr token = el->KeyToken();
if (tok.empty()) {
print_error("unexpected empty element" + String(token->StringContents().c_str()));
}
if (tok[0]->IsBinary()) {
const char *data = tok[0]->begin(), *end = tok[0]->end();
char type;
uint32_t count;
ReadBinaryDataArrayHead(data, end, type, count, el);
if (!count) {
return;
}
if (type != 'i') {
print_error("expected int array (binary)" + String(token->StringContents().c_str()));
}
std::vector<char> buff;
ReadBinaryDataArray(type, count, data, end, buff, el);
//ai_assert(data == end);
//ai_assert(buff.size() == count * 4);
out.reserve(count);
const int32_t *ip = reinterpret_cast<const int32_t *>(&buff[0]);
for (unsigned int i = 0; i < count; ++i, ++ip) {
int32_t val = *ip;
AI_SWAP4(val);
out.push_back(val);
}
return;
}
const size_t dim = ParseTokenAsDim(tok[0]);
// see notes in ParseVectorDataArray()
out.reserve(dim);
const ScopePtr scope = GetRequiredScope(el);
const ElementPtr a = GetRequiredElement(scope, "a", el);
for (TokenList::const_iterator it = a->Tokens().begin(), end = a->Tokens().end(); it != end;) {
const int ival = ParseTokenAsInt(*it++);
out.push_back(ival);
}
}
// ------------------------------------------------------------------------------------------------
// read an array of floats
void ParseVectorDataArray(std::vector<float> &out, const ElementPtr el) {
out.resize(0);
const TokenList &tok = el->Tokens();
TokenPtr token = el->KeyToken();
if (tok.empty()) {
print_error("unexpected empty element: " + String(token->StringContents().c_str()));
}
if (tok[0]->IsBinary()) {
const char *data = tok[0]->begin(), *end = tok[0]->end();
char type;
uint32_t count;
ReadBinaryDataArrayHead(data, end, type, count, el);
if (!count) {
return;
}
if (type != 'd' && type != 'f') {
print_error("expected float or double array (binary) " + String(token->StringContents().c_str()));
}
std::vector<char> buff;
ReadBinaryDataArray(type, count, data, end, buff, el);
//ai_assert(data == end);
//ai_assert(buff.size() == count * (type == 'd' ? 8 : 4));
if (type == 'd') {
const double *d = reinterpret_cast<const double *>(&buff[0]);
for (unsigned int i = 0; i < count; ++i, ++d) {
out.push_back(static_cast<float>(*d));
}
} else if (type == 'f') {
const float *f = reinterpret_cast<const float *>(&buff[0]);
for (unsigned int i = 0; i < count; ++i, ++f) {
out.push_back(*f);
}
}
return;
}
const size_t dim = ParseTokenAsDim(tok[0]);
// see notes in ParseVectorDataArray()
out.reserve(dim);
const ScopePtr scope = GetRequiredScope(el);
const ElementPtr a = GetRequiredElement(scope, "a", el);
for (TokenList::const_iterator it = a->Tokens().begin(), end = a->Tokens().end(); it != end;) {
const float ival = ParseTokenAsFloat(*it++);
out.push_back(ival);
}
}
// ------------------------------------------------------------------------------------------------
// read an array of uints
void ParseVectorDataArray(std::vector<unsigned int> &out, const ElementPtr el) {
out.resize(0);
const TokenList &tok = el->Tokens();
const TokenPtr token = el->KeyToken();
ERR_FAIL_COND_MSG(!token, "invalid ParseVectorDataArrat token invalid");
if (tok.empty()) {
print_error("unexpected empty element: " + String(token->StringContents().c_str()));
}
if (tok[0]->IsBinary()) {
const char *data = tok[0]->begin(), *end = tok[0]->end();
char type;
uint32_t count;
ReadBinaryDataArrayHead(data, end, type, count, el);
if (!count) {
return;
}
if (type != 'i') {
print_error("expected (u)int array (binary)" + String(token->StringContents().c_str()));
}
std::vector<char> buff;
ReadBinaryDataArray(type, count, data, end, buff, el);
//ai_assert(data == end);
//ai_assert(buff.size() == count * 4);
out.reserve(count);
const int32_t *ip = reinterpret_cast<const int32_t *>(&buff[0]);
for (unsigned int i = 0; i < count; ++i, ++ip) {
int32_t val = *ip;
if (val < 0) {
print_error("encountered negative integer index (binary)");
}
out.push_back(val);
}
return;
}
const size_t dim = ParseTokenAsDim(tok[0]);
// see notes in ParseVectorDataArray()
out.reserve(dim);
const ScopePtr scope = GetRequiredScope(el);
const ElementPtr a = GetRequiredElement(scope, "a", el);
for (TokenList::const_iterator it = a->Tokens().begin(), end = a->Tokens().end(); it != end;) {
const int ival = ParseTokenAsInt(*it++);
if (ival < 0) {
print_error("encountered negative integer index");
}
out.push_back(static_cast<unsigned int>(ival));
}
}
// ------------------------------------------------------------------------------------------------
// read an array of uint64_ts
void ParseVectorDataArray(std::vector<uint64_t> &out, const ElementPtr el) {
out.resize(0);
const TokenList &tok = el->Tokens();
TokenPtr token = el->KeyToken();
ERR_FAIL_COND(!token);
if (tok.empty()) {
print_error("unexpected empty element " + String(token->StringContents().c_str()));
}
if (tok[0]->IsBinary()) {
const char *data = tok[0]->begin(), *end = tok[0]->end();
char type;
uint32_t count;
ReadBinaryDataArrayHead(data, end, type, count, el);
if (!count) {
return;
}
if (type != 'l') {
print_error("expected long array (binary): " + String(token->StringContents().c_str()));
}
std::vector<char> buff;
ReadBinaryDataArray(type, count, data, end, buff, el);
//ai_assert(data == end);
//ai_assert(buff.size() == count * 8);
out.reserve(count);
const uint64_t *ip = reinterpret_cast<const uint64_t *>(&buff[0]);
for (unsigned int i = 0; i < count; ++i, ++ip) {
uint64_t val = *ip;
AI_SWAP8(val);
out.push_back(val);
}
return;
}
const size_t dim = ParseTokenAsDim(tok[0]);
// see notes in ParseVectorDataArray()
out.reserve(dim);
const ScopePtr scope = GetRequiredScope(el);
const ElementPtr a = GetRequiredElement(scope, "a", el);
for (TokenList::const_iterator it = a->Tokens().begin(), end = a->Tokens().end(); it != end;) {
const uint64_t ival = ParseTokenAsID(*it++);
out.push_back(ival);
}
}
// ------------------------------------------------------------------------------------------------
// read an array of int64_ts
void ParseVectorDataArray(std::vector<int64_t> &out, const ElementPtr el) {
out.resize(0);
const TokenList &tok = el->Tokens();
TokenPtr token = el->KeyToken();
ERR_FAIL_COND(!token);
if (tok.empty()) {
print_error("unexpected empty element: " + String(token->StringContents().c_str()));
}
if (tok[0]->IsBinary()) {
const char *data = tok[0]->begin(), *end = tok[0]->end();
char type;
uint32_t count;
ReadBinaryDataArrayHead(data, end, type, count, el);
if (!count) {
return;
}
if (type != 'l') {
print_error("expected long array (binary) " + String(token->StringContents().c_str()));
}
std::vector<char> buff;
ReadBinaryDataArray(type, count, data, end, buff, el);
//ai_assert(data == end);
//ai_assert(buff.size() == count * 8);
out.reserve(count);
const int64_t *ip = reinterpret_cast<const int64_t *>(&buff[0]);
for (unsigned int i = 0; i < count; ++i, ++ip) {
int64_t val = *ip;
AI_SWAP8(val);
out.push_back(val);
}
return;
}
const size_t dim = ParseTokenAsDim(tok[0]);
// see notes in ParseVectorDataArray()
out.reserve(dim);
const ScopePtr scope = GetRequiredScope(el);
const ElementPtr a = GetRequiredElement(scope, "a", el);
for (TokenList::const_iterator it = a->Tokens().begin(), end = a->Tokens().end(); it != end;) {
const int64_t val = ParseTokenAsInt64(*it++);
out.push_back(val);
}
}
// ------------------------------------------------------------------------------------------------
Transform ReadMatrix(const ElementPtr element) {
std::vector<float> values;
ParseVectorDataArray(values, element);
if (values.size() != 16) {
print_error("expected 16 matrix elements");
}
// clean values to prevent any IBM damage on inverse() / affine_inverse()
for (float &value : values) {
if (::Math::is_equal_approx(0, value)) {
value = 0;
}
}
Transform xform;
Basis basis;
basis.set(
Vector3(values[0], values[1], values[2]),
Vector3(values[4], values[5], values[6]),
Vector3(values[8], values[9], values[10]));
xform.basis = basis;
xform.origin = Vector3(values[12], values[13], values[14]);
// determine if we need to think about this with dynamic rotation order?
// for example:
// xform.basis = z_axis * y_axis * x_axis;
//xform.basis.transpose();
print_verbose("xform verbose basis: " + (xform.basis.get_euler() * (180 / Math_PI)) + " xform origin:" + xform.origin);
return xform;
}
// ------------------------------------------------------------------------------------------------
// wrapper around ParseTokenAsString() with print_error handling
std::string ParseTokenAsString(const TokenPtr t) {
ERR_FAIL_COND_V(!t, "");
const char *err;
const std::string &i = ParseTokenAsString(t, err);
if (err) {
print_error(String(err) + ", " + String(t->StringContents().c_str()));
}
return i;
}
// ------------------------------------------------------------------------------------------------
// extract a required element from a scope, abort if the element cannot be found
ElementPtr GetRequiredElement(const ScopePtr sc, const std::string &index, const ElementPtr element /*= nullptr*/) {
const ElementPtr el = sc->GetElement(index);
TokenPtr token = el->KeyToken();
ERR_FAIL_COND_V(!token, nullptr);
if (!el) {
print_error("did not find required element \"" + String(index.c_str()) + "\" " + String(token->StringContents().c_str()));
}
return el;
}
bool HasElement(const ScopePtr sc, const std::string &index) {
const ElementPtr el = sc->GetElement(index);
if (nullptr == el) {
return false;
}
return true;
}
// ------------------------------------------------------------------------------------------------
// extract a required element from a scope, abort if the element cannot be found
ElementPtr GetOptionalElement(const ScopePtr sc, const std::string &index, const ElementPtr element /*= nullptr*/) {
const ElementPtr el = sc->GetElement(index);
return el;
}
// ------------------------------------------------------------------------------------------------
// extract required compound scope
ScopePtr GetRequiredScope(const ElementPtr el) {
if (el) {
ScopePtr s = el->Compound();
TokenPtr token = el->KeyToken();
ERR_FAIL_COND_V(!token, nullptr);
if (s) {
return s;
}
ERR_FAIL_V_MSG(nullptr, "expected compound scope " + String(token->StringContents().c_str()));
}
ERR_FAIL_V_MSG(nullptr, "Invalid element supplied to parser");
}
// ------------------------------------------------------------------------------------------------
// extract optional compound scope
ScopePtr GetOptionalScope(const ElementPtr el) {
if (el) {
ScopePtr s = el->Compound();
TokenPtr token = el->KeyToken();
if (token && s) {
return s;
}
}
return nullptr;
}
// ------------------------------------------------------------------------------------------------
// get token at a particular index
TokenPtr GetRequiredToken(const ElementPtr el, unsigned int index) {
if (el) {
const TokenList &x = el->Tokens();
TokenPtr token = el->KeyToken();
ERR_FAIL_COND_V(!token, nullptr);
if (index >= x.size()) {
ERR_FAIL_V_MSG(nullptr, "missing token at index: " + itos(index) + " " + String(token->StringContents().c_str()));
}
return x[index];
}
return nullptr;
}
// ------------------------------------------------------------------------------------------------
// wrapper around ParseTokenAsDim() with print_error handling
size_t ParseTokenAsDim(const TokenPtr t) {
const char *err;
const size_t i = ParseTokenAsDim(t, err);
if (err) {
print_error(String(err) + " " + String(t->StringContents().c_str()));
}
return i;
}
// ------------------------------------------------------------------------------------------------
// wrapper around ParseTokenAsFloat() with print_error handling
float ParseTokenAsFloat(const TokenPtr t) {
const char *err;
const float i = ParseTokenAsFloat(t, err);
if (err) {
print_error(String(err) + " " + String(t->StringContents().c_str()));
}
return i;
}
// ------------------------------------------------------------------------------------------------
// wrapper around ParseTokenAsInt() with print_error handling
int ParseTokenAsInt(const TokenPtr t) {
const char *err;
const int i = ParseTokenAsInt(t, err);
if (err) {
print_error(String(err) + " " + String(t->StringContents().c_str()));
}
return i;
}
// ------------------------------------------------------------------------------------------------
// wrapper around ParseTokenAsInt64() with print_error handling
int64_t ParseTokenAsInt64(const TokenPtr t) {
const char *err;
const int64_t i = ParseTokenAsInt64(t, err);
if (err) {
print_error(String(err) + " " + String(t->StringContents().c_str()));
}
return i;
}
} // namespace FBXDocParser