inkcpp runtime errors and some confusion on my side... (implementing inkcpp into a c++ game engine)

[Lykdraft]

I hope the code is self-explanatory. If not, you gladly could ask me.
(Its inkcpp implemented in a not-so-well-known Game Engine)

#include <ink/story.h>
#include <ink/runner.h>
#include <ink/choice.h>
#include <iostream>

using namespace ink::runtime;


Init() {
//// Load ink binary story, generated from the inkCPP compiler
	story* myStory = story::from_file("test2.bin");
	runner thread = myStory->new_runner();

	while (thread->can_continue())
	{
		rLine = thread->getall(); // --> Attention: Here 'thread->getline()' and 'getall()' works.
		Log::message("Start -> %s \n", rLine.c_str());
	} 
	
		if (thread->has_choices()) {

			Log::message("=== CHOICES ===\n");
			int lineCount = 0;
			for (const choice& c : *thread) {
				Log::message("%d: %s\n", lineCount, c.text());
				lineCount++;
			}
	
		rLine = thread->getline();
		Log::message("68: getLine() - : %s \n", rLine.c_str());
		Log::message("=== MAKING CHOICE 1 ===\n");

		const choice* choice0 = thread->get_choice(1); // -> tried this for debugging. Passes/Compiles.
		Log::message("choice -> %s \n", choice0->text());

		thread->choose(0); // Not ANY story text after this line ("thread->choose(0) or thread->choose(1) ;")

		//rLine = thread->getline(); // --> when enabled, causes runtime exception...
		Log::message("68: getLine() - : %s \n", rLine.c_str());

		auto choice1 = thread->begin(); // -> ... a debug test.
		auto choice2 = thread->end(); // -> ...another debug test. 
		Log::message("79: choice1 -> %s \n", choice1->text());
		Log::message("80: choice2 -> %s \n", choice2->text());
	}

		if (!thread->can_continue()) {
			Log::message("Warning: thread cannot continue.\n"); // --> Doesn't trigger! So 'canContinue()' is true.
			return 1;
		}


		//if (thread->can_continue() && thread.is_valid()) --> Does pass, but...
		//{
		//	rLine = thread->getall(); // --> ...STILL GETS A RUNTIME ERROR HERE!!
		//	Log::message("getLine() - 3 - : %s \n", rLine.c_str());
		//} 

		// Debug story state after choice
		Log::message("After choice - can_continue: %s\n", thread->can_continue() ? "true" : "false"); // prints "true"
		Log::message("After choice - has_choices: %s\n", thread->has_choices() ? "true" : "false"); // prints "false"

		if (thread->can_continue() && thread.is_valid()) { // Does pass...
			try {
				Log::message("trying getLine...\n"); // still prints to console
				rLine = thread->getline(); // --> another RUNTIME exception here... --> Attention: Here 'thread->getline()' dOESN't work!
			}
			catch (const std::exception& e) { // --> DOESN't catch... 
				Log::message("catching exception...\n"); // --> doesn't print...
				if (std::string(e.what()).find("divert") != std::string::npos ||
					std::string(e.what()).find("knot") != std::string::npos) {
					Log::message("story ended on first - stop\n");
				}
				if (std::string(e.what()).find("divert value") != std::string::npos)
				{
					Log::message("story ended - on second stop");
					break;      
				}
				throw; // real error – propagate. --> It seems the Code doesn't get here... Not sure...
			}
			Log::message("End Line called: %s\n", rLine.c_str()); // Doesn't get called anymore if Runtime exception.
                           //  If I out-commented the lines, it prints, but it's empty... 
		}
	return 1;
}

Also, I don't get any text printed after I select a choice via thread->choose(0);. Nothing happens...
Can you help me here? That would be very appreciated.

Here is the test.ink that I compiled in test.bin, if this helps:

> Once upon a time.
> In another time.
> 
>     * There is the first choice.
>     Indeed, the first choice.
>         ** and another choice.
>         ->targetKnotName
>     
>     * There is a second choice.
>     - Of course, there is a second choice.
>         ** to try the same targetKnot...
>         -> targetKnotName
>     
>     - They lived happily ever after.
>     -> DONE
> 
> === targetKnotName ===
> This is the content of the knot.
> Yeah, this knot.
> - and that's it...
> -> endOfStory
> 
> === endOfStory ===
> There it ends.
> Really.
> -> DONE

Addendum: rLine is --> std::string rLine And before I tried const char* rLine. The same errors appear in any case.

[JBenda]

Hey,

I replaced the Log with std::cout to test your code and it runs as expected after some fixes:

#include <ink/system.h>
#include <ink/choice.h>
#include <ink/runner.h>
#include <ink/story.h>

#include <iostream>

using namespace ink::runtime;

int main()
{


	story*      myStory = story::from_file("issue120.ink.bin");
	runner      thread  = myStory->new_runner();
	std::string rLine;

	while (thread->can_continue()) {
		rLine = thread->getall(); // --> Attention: Here 'thread->getline()' and 'getall()' works.
		std::cout << "Start -> " << rLine.c_str() << std::endl;
	}

	if (thread->has_choices()) {

		std::cout << "=== CHOICES ===" << std::endl;
		int lineCount = 0;
		for (const choice& c : *thread) {
			std::cout << lineCount << ": " << c.text() << std::endl;
			lineCount++;
		}

		rLine = thread->getline();
		std::cout << "68: getLine() - : " << rLine.c_str() << std::endl;
		std::cout << "=== MAKING CHOICE 1 ===" << std::endl;

		const choice* choice0 = thread->get_choice(1); // -> tried this for debugging. Passes/Compiles.
		std::cout << "choice -> " << choice0->text() << std::endl;

		thread->choose(0
		); // Not ANY story text after this line ("thread->choose(0) or thread->choose(1) ;")

		rLine = thread->getall(); // --> when enabled, causes runtime exception...
		std::cout << "68: getLine() - : " << rLine.c_str() << std::endl;

		auto choice1 = thread->begin(); // -> ... a debug test.
		auto end     = thread->end();   // -> ...another debug test.
		for (auto choice = thread->begin(); choice != thread->end(); ++choice) {
			std::cout << "xx: choice1 -> " << choice->text() << std::endl;
		}

		thread->choose(0);
		rLine = thread->getall();
		std::cout << "xx: getLine() - " << rLine.c_str() << std::endl;
	}
}

results in

Start -> Once upon a time.
In another time.

=== CHOICES ===
0: There is the first choice.
1: There is a second choice.
68: getLine() - :
=== MAKING CHOICE 1 ===
choice -> There is a second choice.
68: getLine() - : There is the first choice.
Indeed, the first choice.

xx: choice1 -> and another choice.
xx: getLine() - and another choice.
This is the content of the knot.
Yeah, this knot.
and that's it...
There it ends.
Really.

I had no problem with your line //rLine = thread->getline(); // --> when enabled, causes runtime exception... but with the access to the choices thereafter, since getline() is not enough to reach the next choices thread->begin() == thread->end() therefore there are no choices which leads to a runtime error if you try to access them despite.

thread->end() is never a valid pointer and works like the C++ default iterators.

I hope thit helps, If I missed something please point me to the direction ^^

[Lykdraft]

Well...

I had no problem with your line //rLine = thread->getline(); // --> when enabled, causes runtime exception...

Mmhh... Still get the runtime exceptions at the same places I've commented in the original code.
When I out-comment these lines that cause the runtime exception, I get
this text in the console on successful run:

Start -> Once upon a time.
In another time.

=== CHOICES ===
0: There is the first choice.
1: There is a second choice.
68: getLine() - :
=== MAKING CHOICE 1 ===
choice -> There is a second choice.
68: getLine() - :
xx: choice1 ->

You can see, somehow I don't get any text in the 68: getLine() and in the xx:choice1 ->
I have no idea why. As this is Unigine (C++ based) and the inkCL works fine:

I really don't know what's wrong here. Maybe it's the engine..? But I doubt that. But it could be after all.
I will try it this week with Wicked Engine. But that might take a little time.
I need to figure out how to do that first and how to bind it to the Wicked Engine's immediate UI.

Best,
tL.

[JBenda]

That xx:choice1 -> has no output is because at this point there are no choices. Try running thread->has_choices() before; it should return false. And it has no choices because you just chose one before and did not reach the next.

68: getLine() returns an empty string, because you thread->can_continue() is false at this point (I recon) and therefore you cannot get any further and then getLine() returns an empty string.

Hope this helps

[Lykdraft]

I understand. Because I've commented out the
rLine = thread->getall(); // --> when enabled, causes runtime exception...
and the:
rLine = thread->getall();
at the end, no?

That's why there is no text, right?

But when I enable these lines, it gives me a runtime exception on these specific lines.
Co-Pilot Debigging says it is because the runner isn't correctly initialized. But this seems a bit shady.
So what should I do now? :) Between a rock and a hard place it seems... lol...

(Maybe it's the thread->cancontinue(), like you said. I can't check it right now,
I am on a different system, but I will tonight.)

[Lykdraft]

(Maybe it's the thread->cancontinue(), like you said. I can't check it right now

No, it's not the issue. I'll stop it right here.
If there's any news on this, I will follow up.