I would like to pass a std::map to chaiscript. However, I'm not sure how to do this. My code is currently as follows:
#include <map>
#include <string>
#include <chaiscript/chaiscript.hpp>
int main(int argc, char* argv[]) {
chaiscript::ChaiScript chai;
auto data = std::map<std::string, std::string>{
{ "key1", "val1"},
{ "key2", "val2"},
};
chai.add_global(chaiscript::var(&data), "data");
chai.eval(R"(
print(data["key1"]);
)");
}
However, this code crashes with an exception saying, that chaiscript doesn't know what to do with the bracket [] operator. How can I fix this?
I could tell chaiscript what the right function is, but I would prefer it, if the map is compatible with chaiscripts internal Map type!
Update:
I found a bit in the documentation, which explains that the chaiscript map type supports arbitrary input. Looking at the code, this seems to be done by the Boxed_Value type. However, this probably means that it is fundamentally impossible to directly insert std::map into scripts.
I'm now thinking about either writing a custom type, or a conversion function to solve the problem. Keeping you posted...
As far as I can see, out-of-the-box-chaiscript only provides you with the std::map<std::string,chaiscript::Boxed_Value> map type. Therefore, if you want to add your own map to a script, you need to either provide chaiscript with a new type, or convert to the given one. Thus, I see the following solutions:
Case 1: You only need to get out a map from chaiscript to c++
This case can be found in the documentation. You need to supply a conversion function, and off you go.
chai.add(chaiscript::map_conversion<std::map<std::string, std::string>>());
auto map = chai.boxed_cast<std::map<std::string, std::string>>(chai.eval("data"));
Case 2: You only need to supply a map to chaiscript from c++
This is basically the same as Case 1, but you have to supply the conversion function yourself.
auto convert = [](const std::map<std::string, std::string>& std_map) {
auto chai_map = std::map<std::string, chaiscript::Boxed_Value>{};
for (auto&& entry : std_map)
chai_map.emplace(entry.first, chaiscript::var(entry.second));
return chai_map;
};
chai.add(chaiscript::var(convert(data)),"data");
Case 3: You want to share a global value between chaiscript and c++
This case is rather tricky. You either have to supply chaiscript with a get_map() and send_map() function, which handle the data synchronization:
chai.eval(R"(
data = get_map();
data["key1"] = "val1";
send_map(data);
)");
Or you have to add a custom data type, which handles the synchronization in the background.
My Solution:
Fortunately for my case, I don't really need a shared state between chaiscript and c++, and therefore can rely on the solution for Case 2.
Related
I want to use the nlohmann JSON library in order to read options from a JSON file. Specifying options is optional, as reflected in the constructor in my code example. I'm assuming the JSON structure is an object in its root.
Unfortunately, I'm unable to use these options, because it is unclear to me how I can force the JSON structure to be an object. What is worse, merely initializing a member variable with a JSON object {} (magically?) turns it into an array [{}].
#include <cstdlib>
#include <iostream>
#include <nlohmann/json.hpp>
class Example {
public:
explicit Example(const nlohmann::json& options = nlohmann::json::object())
: m_options{options}
{
std::clog << options << '\n' << m_options << '\n';
}
private:
nlohmann::json m_options;
};
auto main() -> int
{
Example x;
Example y{nlohmann::json::object()};
return EXIT_SUCCESS;
}
This results in the following output. Notice that we have to perform some ceremony in order to use an empty object as the default value (= empty settings), with = nlohmann::json::object(). Also notice that the settings object changes its type as soon as we initialize the member value (!):
{}
[{}]
My use use case is quite straightforward, but I'm unable to extract settings, unless I explicitly check whether the settings are an array or an object.
Another thing that worries me is that incorrect code compiles without warning, e.g., code in which I use x.value("y") on a JSON array x containing an object with key "y". Only at run time do I discover that I should have done x.at(0).value("y") instead.
In brief, the whole situation is quite surprising to me. I must be missing something / I must be using this library in an unintended way?
nlohman is a very "modern" library, it uses a lot of features in C++. And that might make it harder to read and understand the code. But it is very flexible.
This short introduction might help
Introduction to nlohmann json
Parse text to json object is done like
constexpr std::string_view stringJson = R"({"k1": "v1"})";
nlohmann::json j = nlohmann::json::parse( stringJson.begin(), stringJson.end() );
I have a function that returns a custom class structure, but how should I handle the cases where I wish to inform the user that the function has failed, as in return false.
My function looks something like this:
Cell CSV::Find(std::string segment) {
Cell result;
// Search code here.
return result;
}
So when succesful, it returns the proper result, but how should I handle the case when it could fail?
I thought about adding a boolean method inside Cell to check what ever Cell.data is empty or not (Cell.IsEmpty()). But am I thinking this issue in a way too complicated way?
There are three general approaches:
Use exceptions. This is what's in Bathsheba's answer.
Return std::optional<Cell> (or some other type which may or may not hold an actual Cell).
Return bool, and add a Cell & parameter.
Which of these is best depends on how you intend this function to be used. If the primary use case is passing a valid segment, then by all means use exceptions.
If part of the design of this function is that it can be used to tell if a segment is valid, exceptions aren't appropriate, and my preferred choice would be std::optional<Cell>. This may not be available on your standard library implementation yet (it's a C++17 feature); if not, boost::optional<Cell> may be useful (as mentioned in Richard Hodges's answer).
In the comments, instead of std::optional<Cell>, user You suggested expected<Cell, error> (not standard C++, but proposed for a future standard, and implementable outside of the std namespace until then). This may be a good option to add some indication on why no Cell could be found for the segment parameter passed in, if there are multiple possible reasons.
The third option I include mainly for completeness. I do not recommend it. It's a popular and generally good pattern in other languages.
Is this function a query, which could validly not find the cell, or is it an imperative, where the cell is expected to be found?
If the former, return an optional (or nullable pointer to) the cell.
If the latter, throw an exception if not found.
Former:
boost::optional<Cell> CSV::Find(std::string segment) {
boost::optional<Cell> result;
// Search code here.
return result;
}
Latter:
as you have it.
And of course there is the c++17 variant-based approach:
#include <variant>
#include <string>
struct CellNotFound {};
struct Cell {};
using CellFindResult = std::variant<CellNotFound, Cell>;
CellFindResult Find(std::string segment) {
CellFindResult result { CellNotFound {} };
// Search code here.
return result;
}
template<class... Ts> struct overloaded : Ts... { using Ts::operator()...; };
template<class... Ts> overloaded(Ts...) -> overloaded<Ts...>;
void cellsAndStuff()
{
std::visit(overloaded
{
[&](CellNotFound)
{
// the not-found code
},
[&](Cell c)
{
// code on cell found
}
}, Find("foo"));
}
The C++ way of dealing with abject failures is to define an exception class of the form:
struct CSVException : std::exception{};
In your function you then throw one of those in the failure branch:
Cell CSV::Find(std::string segment) {
Cell result;
// Search code here.
if (fail) throw CSVException();
return result;
}
You then handle the fail case with a try catch block at the calling site.
If however the "fail" branch is normal behaviour (subjective indeed but only you can be the judge of normality), then do indeed imbue some kind of failure indicator inside Cell, or perhaps even change the return type to std::optional<Cell>.
If you can use C++17, another approach would be to use an std::optional type as your return value. That's a wrapper that may or may not contain a value. The caller can then check whether your function actually returned a value and handle the case where it didn't.
std::optional<Cell> CSV::Find(std::string segment) {
Cell result;
// Search code here.
return result;
}
void clientCode() {
auto cell = CSV::Find("foo");
if (cell)
// do stuff when found
else
// handle not found
}
A further option is using multiple return values:
std::pair<Cell, bool> CSV::Find(std::string segment) {
Cell result;
// Search code here.
return {result, found};
}
// ...
auto cell = CSV::Find("foo");
if (cell->second)
// do stuff with cell->first
The boolean flag says whether the requested Cell was found or not.
PROs
well known approach (e.g. std::map::insert);
quite direct: value and success indicator are return values of the function.
CONs
obscureness of first and second which requires to always remember the relative positions of values within the pairs. C++17 structured bindings / if statement with initializer partially resolve this issue:
if (auto [result, found] = CSV::Find("foo"); found)
// do stuff with `result`
possible loss of safety (the calling code has to check if there is a result value, before using it).
Details
Returning multiple values from functions in C++
C++ Error Handling - downside of using std::pair or std::tuple for returning error codes and function returns
For parsing, it is generally better to avoid std::string and instead use std::string_view; if C++17 is not available, minimally functional versions can be whipped up easily enough.
Furthermore, it is also important to track not only what was parsed but also the remainder.
There are two possibilities to track the remainder:
taking a mutable argument (by reference),
returning the remainder.
I personally prefer the latter, as in case of errors it guarantees that the caller has in its hands a unmodified value which is useful for error-reporting.
Then, you need to examine what potential errors can occur, and what recovery mechanisms you wish for. This will inform the design.
For example, if you wish to be able to parse ill-formed CSV documents, then it is reasonable that Cell be able to represent ill-formed CSV cells, in which case the interface is rather simple:
std::pair<Cell, std::string_view> consume_cell(std::string_view input) noexcept;
Where the function always advances and the Cell may contain either a proper cell, or an ill-formed one.
On the other hand, if you only wish to support well-formed CSV documents, then it is reasonable to signal errors via exceptions and that Cell only be able to hold actual cells:
std::pair<std::optional<Cell>, std::string_view> consume_cell(...);
And finally, you need to think about how to signal end of row conditions. It may a simple marker on Cell, though at this point I personally prefer to create an iterator as it presents a more natural interface since a row is a range of Cell.
The C++ interface for iterators is a bit clunky (as you need an "end", and the end is unknown before parsing), however I recommend sticking to it to be able to use the iterator with for loops. If you wish to depart from it, though, at least make it work easily with while, such as std::optional<Cell> cell; while ((cell = row.next())) { ... }.
I have recently started to use boost::program_options and found it to be highly convenient. That said, there is one thing missing that I was unable to code myself in a good way:
I would like to iterate over all options that have been collected in a boost::program_options::variables_map to output them on the screen. This should become a convenience function, that I can simply call to list all options that were set without the need to update the function when I add new options or for each program.
I know that I can check and output individual options, but as said above, this should become a general solution that is oblivious to the actual options. I further know that I can iterate over the contents of variables_map since it is simply an extended std::map. I could then check for the type containd in the stored boost::any variable and use .as<> to convert it back to the appropriate type. But this would mean coding a long switch block with one case for each type. And this doesn't look like good coding style to me.
So the question is, is there a better way to iterate over these options and output them?
As #Rost previously mentioned, Visitor pattern is a good choice here. To use it with PO you need to use notifiers for your options in such a way that if option is passed notifier will fill an entry in your set of boost::variant values. The set should be stored separately. After that you could iterate over your set and automatically process actions (i.e. print) on them using boost::apply_visitor.
For visitors, inherit from boost::static_visitor<>
Actually, I made Visitor and generic approach use more broad.
I created a class MyOption that holds description, boost::variant for value and other options like implicit, default and so on. I fill a vector of objects of the type MyOption in the same way like PO do for their options (see boost::po::options_add()) via templates. In the moment of passing std::string() or double() for boosts::variant initialization you fill type of the value and other things like default, implicit.
After that I used Visitor pattern to fill boost::po::options_description container since boost::po needs its own structures to parse input command line. During the filling I set notifyer for each option - if it will be passed boost::po will automatically fill my original object of MyOption.
Next you need to execute po::parse and po::notify. After that you will be able to use already filled std::vector<MyOption*> via Visitor pattern since it holds boost::variant inside.
What is good about all of this - you have to write your option type only once in the code - when filling your std::vector<MyOption*>.
PS. if using this approach you will face a problem of setting notifyer for an option with no value, refer to this topic to get a solution: boost-program-options: notifier for options with no value
PS2. Example of code:
std::vector<MyOptionDef> options;
OptionsEasyAdd(options)
("opt1", double(), "description1")
("opt2", std::string(), "description2")
...
;
po::options_descripton boost_descriptions;
AddDescriptionAndNotifyerForBoostVisitor add_decr_visitor(boost_descriptions);
// here all notifiers will be set automatically for correct work with each options' value type
for_each(options.begin(), options.end(), boost::apply_visitor(add_descr_visitor));
It's a good case to use Visitor pattern. Unfortunately boost::any doesn't support Visitor pattern like boost::variant does. Nevertheless there are some 3rd party approaches.
Another possible idea is to use RTTI: create map of type_info of known types mapped to type handler functor.
Since you are going to just print them out anyway you can grab original string representation when you parse. (likely there are compiler errors in the code, I ripped it out of my codebase and un-typedefed bunch of things)
std::vector<std::string> GetArgumentList(const std::vector<boost::program_options::option>& raw)
{
std::vector<std::string> args;
BOOST_FOREACH(const boost::program_options::option& option, raw)
{
if(option.unregistered) continue; // Skipping unknown options
if(option.value.empty())
args.push_back("--" + option.string_key));
else
{
// this loses order of positional options
BOOST_FOREACH(const std::string& value, option.value)
{
args.push_back("--" + option.string_key));
args.push_back(value);
}
}
}
return args;
}
Usage:
boost::program_options::parsed_options parsed = boost::program_options::command_line_parser( ...
std::vector<std::string> arguments = GetArgumentList(parsed.options);
// print
I was dealing with just this type of problem today. This is an old question, but perhaps this will help people who are looking for an answer.
The method I came up with is to try a bunch of as<...>() and then ignore the exception. It's not terribly pretty, but I got it to work.
In the below code block, vm is a variables_map from boost program_options. vit is an iterator over vm, making it a pair of std::string and boost::program_options::variable_value, the latter being a boost::any. I can print the name of the variable with vit->first, but vit->second isn't so easy to output because it is a boost::any, ie the original type has been lost. Some should be cast as a std::string, some as a double, and so on.
So, to cout the value of the variable, I can use this:
std::cout << vit->first << "=";
try { std::cout << vit->second.as<double>() << std::endl;
} catch(...) {/* do nothing */ }
try { std::cout << vit->second.as<int>() << std::endl;
} catch(...) {/* do nothing */ }
try { std::cout << vit->second.as<std::string>() << std::endl;
} catch(...) {/* do nothing */ }
try { std::cout << vit->second.as<bool>() << std::endl;
} catch(...) {/* do nothing */ }
I only have 4 types that I use to get information from the command-line/config file, if I added more types, I would have to add more lines. I'll admit that this is a bit ugly.
this is my first question after long time checking on this marvelous webpage.
Probably my question is a little silly but I want to know others opinion about this. What is better, to create several specific methods or, on the other hand, only one generic method? Here is an example...
unsigned char *Method1(CommandTypeEnum command, ParamsCommand1Struct *params)
{
if(params == NULL) return NULL;
// Construct a string (command) with those specific params (params->element1, ...)
return buffer; // buffer is a member of the class
}
unsigned char *Method2(CommandTypeEnum command, ParamsCommand2Struct *params)
{
...
}
unsigned char *Method3(CommandTypeEnum command, ParamsCommand3Struct *params)
{
...
}
unsigned char *Method4(CommandTypeEnum command, ParamsCommand4Struct *params)
{
...
}
or
unsigned char *Method(CommandTypeEnum command, void *params)
{
switch(command)
{
case CMD_1:
{
if(params == NULL) return NULL;
ParamsCommand1Struct *value = (ParamsCommand1Struct *) params;
// Construct a string (command) with those specific params (params->element1, ...)
return buffer;
}
break;
// ...
default:
break;
}
}
The main thing I do not really like of the latter option is this,
ParamsCommand1Struct *value = (ParamsCommand1Struct *) params;
because "params" could not be a pointer to "ParamsCommand1Struct" but a pointer to "ParamsCommand2Struct" or someone else.
I really appreciate your opinions!
General Answer
In Writing Solid Code, Steve Macguire's advice is to prefer distinct functions (methods) for specific situations. The reason is that you can assert conditions that are relevant to the specific case, and you can more easily debug because you have more context.
An interesting example is the standard C run-time's functions for dynamic memory allocation. Most of it is redundant, as realloc can actually do (almost) everything you need. If you have realloc, you don't need malloc or free. But when you have such a general function, used for several different types of operations, it's hard to add useful assertions and it's harder to write unit tests, and it's harder to see what's happening when debugging. Macquire takes it a step farther and suggests that, not only should realloc just do _re_allocation, but it should probably be two distinct functions: one for growing a block and one for shrinking a block.
While I generally agree with his logic, sometimes there are practical advantages to having one general purpose method (often when operations is highly data-driven). So I usually decide on a case by case basis, with a bias toward creating very specific methods rather than overly general purpose ones.
Specific Answer
In your case, I think you need to find a way to factor out the common code from the specifics. The switch is often a signal that you should be using a small class hierarchy with virtual functions.
If you like the single method approach, then it probably should be just a dispatcher to the more specific methods. In other words, each of those cases in the switch statement simply call the appropriate Method1, Method2, etc. If you want the user to see only the general purpose method, then you can make the specific implementations private methods.
Generally, it's better to offer separate functions, because they by their prototype names and arguments communicate directly and visibly to the user that which is available; this also leads to more straightforward documentation.
The one time I use a multi-purpose function is for something like a query() function, where a number of minor query functions, rather than leading to a proliferation of functions, are bundled into one, with a generic input and output void pointer.
In general, think about what you're trying to communicate to the API user by the API prototypes themselves; a clear sense of what the API can do. He doesn't need excessive minutae; he does need to know the core functions which are the entire point of having the API in the first place.
First off, you need to decide which language you are using. Tagging the question with both C and C++ here makes no sense. I am assuming C++.
If you can create a generic function then of course that is preferable (why would you prefer multiple, redundant functions?) The question is; can you? However, you seem to be unaware of templates. We need to see what you have omitted here to tell if you if templates are suitable however:
// Construct a string (command) with those specific params (params->element1, ...)
In the general case, assuming templates are appropriate, all of that turns into:
template <typename T>
unsigned char *Method(CommandTypeEnum command, T *params) {
// more here
}
On a side note, how is buffer declared? Are you returning a pointer to dynamically allocated memory? Prefer RAII type objects and avoid dynamically allocating memory like that if so.
If you are using C++ then I would avoid using void* as you don't really need to. There is nothing wrong with having multiple methods. Note that you don't actually have to rename the function in your first set of examples - you can just overload a function using different parameters so that there is a separate function signature for each type. Ultimately, this kind of question is very subjective and there are a number of ways of doing things. Looking at your functions of the first type, you would perhaps be well served by looking into the use of templated functions
You could create a struct. That's what I use to handle console commands.
typedef int (* pFunPrintf)(const char*,...);
typedef void (CommandClass::*pKeyFunc)(char *,pFunPrintf);
struct KeyCommand
{
const char * cmd;
unsigned char cmdLen;
pKeyFunc pfun;
const char * Note;
long ID;
};
#define CMD_FORMAT(a) a,(sizeof(a)-1)
static KeyCommand Commands[]=
{
{CMD_FORMAT("one"), &CommandClass::CommandOne, "String Parameter",0},
{CMD_FORMAT("two"), &CommandClass::CommandTwo, "String Parameter",1},
{CMD_FORMAT("three"), &CommandClass::CommandThree, "String Parameter",2},
{CMD_FORMAT("four"), &CommandClass::CommandFour, "String Parameter",3},
};
#define AllCommands sizeof(Commands)/sizeof(KeyCommand)
And the Parser function
void CommandClass::ParseCmd( char* Argcommand )
{
unsigned int x;
for ( x=0;x<AllCommands;x++)
{
if(!memcmp(Commands[x].cmd,Argcommand,Commands[x].cmdLen ))
{
(this->*Commands[x].pfun)(&Argcommand[Commands[x].cmdLen],&::printf);
break;
}
}
if(x==AllCommands)
{
// Unknown command
}
}
I use a thread safe printf pPrintf, so ignore it.
I don't really know what you want to do, but in C++ you probably should derive multiple classes from a Formatter Base class like this:
class Formatter
{
virtual void Format(unsigned char* buffer, Command command) const = 0;
};
class YourClass
{
public:
void Method(Command command, const Formatter& formatter)
{
formatter.Format(buffer, command);
}
private:
unsigned char* buffer_;
};
int main()
{
//
Params1Formatter formatter(/*...*/);
YourClass yourObject;
yourObject.Method(CommandA, formatter);
// ...
}
This removes the resposibility to handle all that params stuff from your class and makes it closed for changes. If there will be new commands or parameters during further development you don't have to modifiy (and eventually break) existing code but add new classes that implement the new stuff.
While not full answer this should guide you in correct direction: ONE FUNCTION ONE RESPONSIBILITY. Prefer the code where it is responsible for one thing only and does it well. The code whith huge switch statement (which is not bad by itself) where you need cast void * to some other type is a smell.
By the way I hope you do realise that according to standard you can only cast from void * to <type> * only when the original cast was exactly from <type> * to void *.
I'm using luabind 0.9.1 from Ryan Pavlik's master distribution with Lua 5.1, cygwin on Win XP SP3 + latest patches x86, boost 1.48, gcc 4.3.4. Lua and boost are cygwin pre-compiled versions.
I've successfully built luabind in both static and shared versions.
Both versions pass all the tests EXCEPT for the test_object_identity.cpp test which fails in both versions.
I've tracked down the problem to the following issue:
If an entry in a table is created for NON built-in class (i.e., not int, string, etc), the value CANNOT be retrieved.
Here's a code piece that demonstrates this:
#include "test.hpp"
#include <luabind/luabind.hpp>
#include <luabind/detail/debug.hpp>
using namespace luabind;
struct test_param
{
int obj;
};
void test_main(lua_State* L)
{
using namespace luabind;
module(L)
[
class_<test_param>("test_param")
.def_readwrite("obj", &test_param::obj)
];
test_param temp_object;
object tabc = newtable(L);
tabc[1] = 10;
tabc[temp_object] = 30;
TEST_CHECK( tabc[1] == 10 ); // passes
TEST_CHECK( tabc[temp_object] == 30 ); // FAILS!!!
}
tabc[1] is indeed 10 while tabc[temp_object] is NOT 30! (actually, it seems to be nil)
However, if I use iterate to go over tabc entries, there're the two entries with the CORRECT key/value pairs.
Any ideas?
BTW, overloading the == operator like this:
#include <luabind/operator.hpp>
struct test_param
{
int obj;
bool operator==(test_param const& rhs) const
{
return obj == rhs.obj;
}
};
and
module(L)
[
class_<test_param>("test_param")
.def_readwrite("obj", &test_param::obj)
.def(const_self == const_self)
];
Doesn't change the result.
I also tried switching to settable() and gettable() from the [] operator. The result is the same. I can see with the debugger that default conversion of the key is invoked, so I guess the error arises from somewhere therein, but it's beyond me to figure out what exactly the problem is.
As the following simple test case show, there're definitely a bug in Luabind's conversion for complex types:
struct test_param : wrap_base
{
int obj;
bool operator==(test_param const& rhs) const
{ return obj == rhs.obj ; }
};
void test_main(lua_State* L)
{
using namespace luabind;
module(L)
[
class_<test_param>("test_param")
.def(constructor<>())
.def_readwrite("obj", &test_param::obj)
.def(const_self == const_self)
];
object tabc, zzk, zzv;
test_param tp, tp1;
tp.obj = 123456;
// create new table
tabc = newtable(L);
// set tabc[tp] = 5;
// o k v
settable( tabc, tp, 5);
// get access to entry through iterator() API
iterator zzi(tabc);
// get the key object
zzk = zzi.key();
// read back the value through gettable() API
// o k
zzv = gettable(tabc, zzk);
// check the entry has the same value
// irrespective of access method
TEST_CHECK ( *zzi == 5 &&
object_cast<int>(zzv) == 5 );
// convert key to its REAL type (test_param)
tp1 = object_cast<test_param>(zzk);
// check two keys are the same
TEST_CHECK( tp == tp1 );
// read the value back from table using REAL key type
zzv = gettable(tabc, tp1);
// check the value
TEST_CHECK( object_cast<int>(zzv) == 5 );
// the previous call FAILS with
// Terminated with exception: "unable to make cast"
// this is because gettable() doesn't return
// a TRUE value, but nil instead
}
Hopefully, someone smarter than me can figure this out,
Thx
I've traced the problem to the fact that Luabind creates a NEW DISTINCT object EVERY time you use a complex value as key (but it does NOT if you use a primitive one or an object).
Here's a small test case that demonstrates this:
struct test_param : wrap_base
{
int obj;
bool operator==(test_param const& rhs) const
{ return obj == rhs.obj ; }
};
void test_main(lua_State* L)
{
using namespace luabind;
module(L)
[
class_<test_param>("test_param")
.def(constructor<>())
.def_readwrite("obj", &test_param::obj)
.def(const_self == const_self)
];
object tabc, zzk, zzv;
test_param tp;
tp.obj = 123456;
tabc = newtable(L);
// o k v
settable( tabc, tp, 5);
iterator zzi(tabc), end;
std::cerr << "value = " << *zzi << "\n";
zzk = zzi.key();
// o k v
settable( tabc, tp, 6);
settable( tabc, zzk, 7);
for (zzi = iterator(tabc); zzi != end; ++zzi)
{
std::cerr << "value = " << *zzi << "\n";
}
}
Notice how tabc[tp] first has the value 5 and then is overwritten with 7 when accessed through the key object. However, when accessed AGAIN through tp, a new entry gets created. This is why gettable() fails subsequently.
Thx,
David
Disclaimer: I'm not an expert on luabind. It's entirely possible I've missed something about luabind's capabilities.
First of all, what is luabind doing when converting test_param to a Lua key? The default policy is copy. To quote the luabind documentation:
This will make a copy of the parameter. This is the default behavior when passing parameters by-value. Note that this can only be used when passing from C++ to Lua. This policy requires that the parameter type has an accessible copy constructor.
In pratice, what this means is that luabind will create a new object (called "full userdata") which is owned by the Lua garbage collector and will copy your struct into it. This is a very safe thing to do because it no longer matters what you do with the c++ object; the Lua object will stick around without really any overhead. This is a good way to do bindings for by-value sorts of objects.
Why does luabind create a new object each time you pass it to Lua? Well, what else could it do? It doesn't matter if the address of the passed object is the same, because the original c++ object could have changed or been destroyed since it was first passed to Lua. (Remember, it was copied to Lua by value, not by reference.) So, with only ==, luabind would have to maintain a list of every object of that type which had ever been passed to Lua (possibly weakly) and compare your object against each one to see if it matches. luabind doesn't do this (nor do I think should it).
Now, let's look at the Lua side. Even though luabind creates two different objects, they're still equal, right? Well, the first problem is that, besides certain built-in types, Lua can only hold objects by reference. Each of those "full userdata" that I mentioned before is actually a pointer. That means that they are not identical.
But they are equal, if we define an __eq meta operation. Unfortunately, Lua itself simply does not support this case. Userdata when used as table keys are always compared by identity, no matter what. This actually isn't special for userdata; it is also true for tables. (Note that to properly support this case, Lua would need to override the hashcode operation on the object in addition to __eq. Lua also does not support overriding the hashcode operation.) I can't speak for the authors of Lua why they did not allow this (and it has been suggested before), but there it is.
So, what are the options?
The simplest thing would be to convert test_param to an object once (explicitly), and then use that object to index the table both times. However, I suspect that while this fixes your toy example, it isn't very helpful in practice.
Another option is simply not to use such types as keys. Actually, I think this is a very good suggestion, since this kind of light-weight binding is quite useful, and the only other option is to discard it.
It looks like you can define a custom conversion on your type. In your example, it might be reasonable to convert your type to a Lua number which will behave well as a table index.
Use a different kind of binding. There will be some overhead, but if you want identity, you'll have to live with it. It sounds like luabind has some support for wrappers, which you may need to use to preserve identity:
When a pointer or reference to a registered class with a wrapper is passed to Lua, luabind will query for it's dynamic type. If the dynamic type inherits from wrap_base, object identity is preserved.