Smart-Overhaul

Printable View

Show 40 post(s) from this thread on one page

05-31-2013, 10:49 PM
Brandon

Quote:

Originally Posted by Kevin

I understand that only you would see that change as I mentioned it as a 'personal only draw.'

The point of that would be for only 1 option I can currently imagine:
Purely for the fun of viewing it and the available animations that you could create and show as an interaction (that would not be interacting with you, as it is purely aesthetic for the script user).

I only asked because I wasn't sure if it would be something that would require time of you, or whether there'd be available access natively via the api you're adding, or whether we will simply have get access. It's not a big deal as it's purely a mostly useless aesthetic "play-around" function.

Yeah it wouldn't be difficult to add but I didn't see a use for such a thing which was asked before on the OpenGL thread so I didn't bother adding it. Only use I thought of at the time was for anti-leech where you can spawn a model or move a model to some random location on screen.

And yeah its a lot more work to add OpenGL stuff to the plugin atm because not only would I have to export for Lape, but I'd have to write wrappers and export those for Pascal-Script too. You can see that if you type GLIteratorPointer~ in Simba's function box. It should show one for iterating Ints, Floats, Raw, short, etc.. :( Lape however can do that natively.

Not to mention the limitations I have of passing things to Simba via SharedMemory/Serialization. In the past I found ways around those though. And I also had to write the Simba include as well. Perhaps I might just start working on the plugin again and hopefully finish the walking I started with Kasi.. Though I don't remember where I saved those files at all or how far I got or anything. IIRC, we worked on the text recognition with his good pattern finding skills as well.

TBH, the only time I do anything related to SRL/Smart/OpenGL/Bug-fixing/Internet is when I'm asked to or if I'm tagged/skyped/emailed/msg'd or if I release something here or one the web or anywhere that needs fixing and yeah..
06-01-2013, 02:27 AM
Flight

Ah there we go, received a fully colored bitmap from the OGL buffer for Simba. :thumbsup: I quickly tested SPS; it works like a charm. Nice to know we can still walk in OGL mode. Very good job Brandon. AFAIK all that's left now is, as you said before, drawing the mouse on the SMART canvas but it sounds like you have some ideas towards that. Keep up the great work!
06-01-2013, 02:31 AM
rj

Quote:

Originally Posted by Flight

Ah there we go, received a fully colored bitmap from the OGL buffer for Simba. :thumbsup: I quickly tested SPS; it works like a charm. Nice to know we can still walk in OGL mode. Very good job Brandon. AFAIK all that's left now is, as you said before, drawing the mouse on the SMART canvas but it sounds like you have some ideas towards that. Keep up the great work!

Have you tested the object finding with OpenGL yet?

@Brandon I have a couple questions about open gl:

1) how accurate is object finding, I understand you have to use some sort of tolerance? A I wrong here?

2) how fast is object finding?

3) do we have to update model ids after every update, every login, or whenever the model is changed ie graphical updates)

Thanks in advance
06-01-2013, 02:47 AM
Brandon

Quote:

Originally Posted by Officer Barbrady

Have you tested the object finding with OpenGL yet?

@Brandon I have a couple questions about open gl:

1) how accurate is object finding, I understand you have to use some sort of tolerance? A I wrong here?

2) how fast is object finding?

3) do we have to update model ids after every update, every login, or whenever the model is changed ie graphical updates)

Thanks in advance

1) 100% There is no mistakes for object finding. It's not going to find a random object with a specific ID. You give it a bad ID, it gives you nothing. You give it a valid Object ID, it returns all information about that object. Do what you want with the information. Click it, print it, read it, save it, etc.. It would only make a mistake if my math is off or I made a mistake implementing it.

2) Instant but then again, this depends on Simba. See, C++ side finds the model instantly. As fast as runescape can draw, that is how fast it is found. The second it is drawn on screen, it is recognized and saved. On the other hand, it sends it to Simba/PS/Lape for processing. If Simba's interpreter is slow, it will be slow. Fortunately, I haven't had any "slow" reactions in Simba. This includes text recognition and all that stuff. IIRC, it has chooseoption and all of that 100% accurate as well. Same with recognizing text.

You can see that the FPS is the same with the plugin loaded as without. Well for me it is. Press the Tilde (~) key on your keyboard and type displayfps to turn it on/off and see.

3) No. This isn't Java reflection/injection. There is no hooks involved. The only time an update is required is if Jagex updates their "Engine" to use different OpenGL calls (such as OpenGL 4.X+). And of course graphical updates would change an ID but does not require me to update Smart/Plugins. All it requires is the scripter update their Object-ID. Also colour is an option as well. So best of BOTH worlds are involved here. If you want to use colour, you can. If you want to use OpenGL, you can. If you want both, it's there. Also, this doesn't block Reflection/Injection so if Smart ever gets those back, you can also use those in combination with these options as well or entirely separate.
It's an option not just the only choice which is good.

The include currently has a file with a list of collected ID's I was using and AFAIK, it still works even though I wrote it near mid 2012. You can test it by checking ChooseOptions, GetSkillLevel, etc..

http://i.imgur.com/TZJZ514.png
06-01-2013, 02:50 AM
rj

Quote:

Originally Posted by Brandon

1) 100% There is no mistakes for object finding. It's not going to find a random object with a specific ID. You give it a bad ID, it gives you nothing. You give it a valid Object ID, it returns all information about that object. Do what you want with the information. Click it, etc..

2) Instant but then again, this depends on Simba. See, C++ side finds the model instantly. As fast as runescape can draw, that is how fast it is found. The second it is drawn on screen, it is recognized and saved. On the other hand, it sends it to Simba/PS/Lape for processing. If Simba's interpreter is slow, it will be slow. Fortunately, I haven't had any "slow" reactions in Simba. This includes text recognition and all that stuff. IIRC, it has chooseoption and all of that 100% accurate as well. Same with recognizing text.

3) No. This isn't Java reflection/injection. There is no hooks involved. The only time an update is required is if Jagex updates their "Engine" to use different OpenGL calls (such as OpenGL 4.X+). And of course graphical updates would change an ID but does not require me to update Smart/Plugins. All it requires is the scripter update their Object-ID. Also colour is an option as well. So best of BOTH worlds are involved here. If you want to use colour, you can. If you want to use OpenGL, you can. If you want both, it's there. Also, this doesn't block Reflection/Injection so if Smart ever gets those back, you can also use those in combination with these options as well or entirely separate.

The include currently has a file with a list of collected ID's I was using and AFAIK, it still works even though I wrote it near mid 2012. You can test it by checking ChooseOptions, GetSkillLevel, etc..

http://i.imgur.com/TZJZ514.png

Well, I know what I'm doing when I get home!

The reason I thought I would have to change model ids is because I saw a script on tribot were the user had to enter their model I'd each time they used the script
06-01-2013, 02:50 AM
Flight

Quote:

Originally Posted by Officer Barbrady

Have you tested the object finding with OpenGL yet?

Yes of course. Objects are projected as on-screen models and they all have specific ID's.

Quote:

Originally Posted by Officer Barbrady

@Brandon I have a couple questions about open gl:

1) how accurate is object finding, I understand you have to use some sort of tolerance? A I wrong here?

2) how fast is object finding?

3) do we have to update model ids after every update, every login, or whenever the model is changed ie graphical updates)

Thanks in advance

1) 100%. Each model on the screen is stored with the variables: ID, Stride, SX, SY and TriangleCount. The SX/SY is (always) the exact center of the model; so when a specific model is found at that very instant the middle of it is exactly found and you cannot 'miss it' unless it has changed position on your screen (NPC/Player walking and/or camera movement). Tolerance... I don't understand that part.

2) Well I just did a test at the fishing place at the barbarian village and this is the speed test:

Code:

ALL models found in 16 ms. 1 fishing spots found in 0 ms.

Tested 2 more times for nothing over 40ms.

3) No. The model I got for a fly-fishing spot (there's two different models for this one particularly) was back when this didn't even exist; I found this ID back when we only had Silab's original OGL interception, well over a year ago if I'm not mistaken. Only some models will change IDs and triangle counts if Jagex chooses to update their actual in-game appearance.

Ah darn it, I've been ninja'd, agh...
06-01-2013, 02:53 AM
rj

Quote:

Originally Posted by Flight

Yes of course. Objects are projected as on-screen models and they all have specific ID's.

1) 100%. Each model on the screen is stored with the variables: ID, Stride, SX, SY and TriangleCount. The SX/SY is (always) the exact center of the model; so when a specific model is found at that very instant the middle of it is exactly found and you cannot 'miss it' unless it has changed position on your screen (NPC/Player walking and/or camera movement). Tolerance... I don't understand that part.

2) Well I just did a test at the fishing place at the barbarian village and this is the speed test:

Code:

ALL models found in 16 ms. 1 fishing spots found in 0 ms.

Tested 2 more times for nothing over 40ms.

3) No. The model I got for a fly-fishing spot (there's two different models for this one particularly) was back when this didn't even exist; I found this ID back when we only had Silab's original OGL interception, well over a year ago if I'm not mistaken. Only some models will change IDs and triangle counts if Jagex chooses to update their actual in-game appearance.

Ah darn it, I've been ninja'd, agh...

Ahh we'll I thought it generated an I'd based off Angles and some complicated math stuff, and I figured if our camera angle changes then the object angles would be different making a different Id so then it would ned s tolerance or something lol. I just didn't fully understand how OpenGL works
06-01-2013, 02:55 AM
Brandon

Quote:

Originally Posted by Officer Barbrady

Well, I know what I'm doing when I get home!

The reason I thought I would have to change model ids is because I saw a script on tribot were the user had to enter their model I'd each time they used the script

Ohh I know what you're talking about. You mean this? https://tribot.org/forums/index.php?...ndon#entry3657

I thought they fixed that a while back? It used to be that way before but I came up with this idea to ignore the first few scan-lines on images. This doesn't affect models at all and Inventory/Bank/Interfaces are found perfectly as well. Angles don't affect it either because all triangles/pointers/quads are the same regardless of the angle you're looking at. Only difference is their location on screen which can be used for animation but its really hard to get that down to an art.

I remember when it "looked" like this:

http://i.imgur.com/vahpxSY.png
http://i.imgur.com/j0rcUlU.png
http://i.imgur.com/GGVfQ.png
http://i.imgur.com/syvgs.png

Perfection! Lol.. Dunno where I put that code though and I was sure it had a bug so I removed that. But yeah it still gives you 100% accurate Model info.
@Officer Barbrady; There is other pictures above and on OP as well. Once you try it, you'll see it for yourself ;)

You can see more pics here: http://villavu.com/forum/showthread.php?t=86230
06-01-2013, 04:30 PM
Sjoe

Just wanna say no more lag! Works great ;)
I'll see if a Six hour reload works for me.
06-01-2013, 04:34 PM
rj

Quote:

Originally Posted by Brandon

Ohh I know what you're talking about. You mean this? https://tribot.org/forums/index.php?...ndon#entry3657

I thought they fixed that a while back? It used to be that way before but I came up with this idea to ignore the first few scan-lines on images. This doesn't affect models at all and Inventory/Bank/Interfaces are found perfectly as well. Angles don't affect it either because all triangles/pointers/quads are the same regardless of the angle you're looking at. Only difference is their location on screen which can be used for animation but its really hard to get that down to an art.

I remember when it looked like this:

http://i.imgur.com/vahpxSY.png

Perfection! Lol.. Dunno where I put that code though and I was sure it had a bug so I removed that. But yeah it still gives you 100% accurate Model info.

Can you show me some more pictures of this OpenGL stuff? I'm getting pumped to arrive home haha
06-01-2013, 06:17 PM
Ashaman88

Quote:

Originally Posted by Officer Barbrady

Can you show me some more pictures of this OpenGL stuff? I'm getting pumped to arrive home haha

http://puu.sh/361fO.png
06-01-2013, 06:20 PM
rj

Quote:

Originally Posted by Ashaman88

http://puu.sh/361fO.png

Wools was going to add thanks to that comment but I hit "add" instead of the bar

Are you going to convert your Lrc to ope gl?

Or maybe a Lrc fisher? :)
06-01-2013, 06:28 PM
Ashaman88

Quote:

Originally Posted by Officer Barbrady

Wools was going to add thanks to that comment but I hit "add" instead of the bar

Are you going to convert your Lrc to ope gl?

Or maybe a Lrc fisher? :)

Gonna try out OGL LRC first, already have it able to identify present deposits or not :)
06-01-2013, 06:30 PM
rj

Quote:

Originally Posted by Ashaman88

Gonna try out OGL LRC first, already have it able to identify present deposits or not :)

Very nice, can it see models that are not on the screen? We wouldn't want to be like powerbot were they run 200 tiles to a gold ore deposit
06-01-2013, 06:32 PM
Ashaman88

Quote:

Originally Posted by Officer Barbrady

Very nice, can it see models that are not on the screen? We wouldn't want to be like powerbot were they run 200 tiles to a gold ore deposit

Not sure - I would probably just have it work like mine does now, but 100% detection and such
06-01-2013, 07:11 PM
Itankbots

Mine still just gets stuck on the "loading... Please wait" screen. and Never loads the actual smart clients...

I have no idea why...If anyone could help me would be much appreciated. Message me on here or skype peter_knight2191
06-01-2013, 11:57 PM
Brandon

Quote:

Originally Posted by Itankbots

Mine still just gets stuck on the "loading... Please wait" screen. and Never loads the actual smart clients...

I have no idea why...If anyone could help me would be much appreciated. Message me on here or skype peter_knight2191

You can send me a PM with your TV ID and pass if you like or if someone else is willing to help you then that's good too.

@Everyone else. If you find that you sometimes get awkward displays or can't find an ID because there is many on top of it, play around with the Stride. GLHSetStride takes the values: 12, 16, 20, 24, 36, 40. Some strides are only available in HD mode such as 28 and above. Other strides display different ID's.

12 Displays NPC's, whereas 24 displays Ground Items.. etc. So yeah Don't panic if you see a crap ton of ID's at once. GLHEnableFilter would also hide specific ID's or display only specific ones. It takes time to get used to the include. If you want, you can re-write it to your liking because what I wrote was fairly simple and just an example of how an include would be.

I think I covered all that in the tutorial or something.. Meh.. I'll probably go clean up a lot of the OpenGL api and make things a lot simpler (Currently converting a lot of it to C++x11 and Documenting it (finally)).
06-02-2013, 02:13 AM
rj

Couple more suggestions: maybe a button that opens up the Simba scripts folder :)

I was thinking maybe a button that could run Simba scripts automaticity but then I thought well that requires editing Simba and this isn't powerbot but it would still be cool
06-02-2013, 02:27 AM
Brandon

Quote:

Originally Posted by Officer Barbrady

Couple more suggestions: maybe a button that opens up the Simba scripts folder :)

I was thinking maybe a button that could run Simba scripts automaticity but then I thought well that requires editing Simba and this isn't powerbot but it would still be cool

That would be nice but that would limit Smart to only a specific Controller/Client. For example, I don't run Smart with Simba alone. I have a C++x11 Library that runs Smart using C++ compiled code. So I double click, enter my username/pass, it runs Smart and plays my script.

Adding a button like that would also not be helpful if Scar-divi adopts this Smart.

But yeah. Not a bad idea just don't think that would be an option for "everyone". I guess you could argue that they don't have to click it but meh.. What do you think after reading the above?
06-02-2013, 02:29 AM
rj

Quote:

Originally Posted by Brandon

That would be nice but that would limit Smart to only a specific Controller/Client. For example, I don't run Smart with Simba alone. I have a C++x11 Library that runs Smart using C++ compiled code. So I double click, enter my username/pass, it runs Smart and plays my script.

Adding a button like that would also not be helpful if Scar-divi adopts this Smart.

But yeah. Not a bad idea just don't think that would be good for "everyone".

You mean Simba? :)

I has another couple ideas but I forgot ffs, if I remember ill post here later ;)

Quote:

Originally Posted by Officer Barbrady

You mean Simba? :)

I has another couple ideas but I forgot ffs, if I remember ill post here later ;)

Well not specifically Simba. See, Smart wasn't just made for only Simba and I don't wanna take away the original intent of Smart in which Ben designed it. He still is and always will be the rightful owner of Smart and can do anything he wishes with it at any time.

Ex:

Scar can load smart and Simba can load Smart and any language that uses the same calling convention can load Smart and use it. Adding a button that is specific to Simba would imply that Smart is made for Simba alone. Besides that's what we have the start menu for. Now if Simba came up with a Script repository that needed to be maintained by Smart for whatever reason, then I'd have no problem incorporating that (though it can be done through a plugin.. Same with the button thing).

Ex2(Smart can be loaded by any language with the right calling conventions):
C++ Code:
typedef int (*ptr_SpawnProcess)(char* RemotePath, char* JavaArgs, char* Plugins);
typedef int (*ptr_SpawnClient)(int PID, char* Root, char* Parameters, int Width, int Height, char* InitSequence, char* UserAgent);



int main()
{
    tstring Bits = _T(sizeof(void*) == 8 ? "64" : "32");
    tstring Library = _T("LibSmartRemote") + Bits;
    
    #if defined _WIN32 || defined _WIN64
    HANDLE Lib = LoadLibrary((Library + _T(".dll")).c_str());
    #else
    void* Lib = dlopen((Library + _T(".so")).c_str(), RTLD_LAZY);
    #endif
    
    #if defined _WIN32 || defined _WIN64
    ptr_SpawnProcess SpawnProcess = reinterpret_cast<ptr_SpawnProcess>(GetProcAddress(Lib, _T("SmartSpawnProcess")));
    ptr_SpawnClient SpawnClient = reinterpret_cast<ptr_SpawnClient>(GetProcAddress(Lib, _T("SmartSpawnClient")));
    #else
    ptr_SpawnProcess SpawnProcess = reinterpret_cast<ptr_SpawnProcess>(dlsym(Lib, _T("SmartSpawnProcess")));
    ptr_SpawnClient SpawnClient = reinterpret_cast<ptr_SpawnClient>(dlsym(Lib, _T("SmartSpawnClient")));
    #endif
    
    SpawnProcess(...);
    SpawnClient(....);
        RunScript(...);
}

Quote:
Originally Posted by Brandon
Well not specifically Simba. See, Smart wasn't just made for only Simba and I don't wanna take away the original intent of Smart in which Ben designed it. He still is and always will be the rightful owner of Smart and can do anything he wishes with it at any time.

Ex:

Scar can load smart and Simba can load Smart and any language that uses the same calling convention can load Smart and use it. Adding a button that is specific to Simba would imply that Smart is made for Simba alone. Besides that's what we have the start menu for. Now if Simba came up with a Script repository that needed to be maintained by Smart for whatever reason, then I'd have no problem incorporating that (though it can be done through a plugin.. Same with the button thing).

Ex2(Smart can be loaded by any language with the right calling conventions):
C++ Code:
typedef int (*ptr_SpawnProcess)(char* RemotePath, char* JavaArgs, char* Plugins);
typedef int (*ptr_SpawnClient)(int PID, char* Root, char* Parameters, int Width, int Height, char* InitSequence, char* UserAgent);



int main()
{
    tstring Bits = _T(sizeof(void*) == 8 ? "64" : "32");
    tstring Library = _T("LibSmartRemote") + Bits;
    
    #if defined _WIN32 || defined _WIN64
    HANDLE Lib = LoadLibrary((Library + _T(".dll")).c_str());
    #else
    void* Lib = dlopen((Library + _T(".so")).c_str(), RTLD_LAZY);
    #endif
    
    #if defined _WIN32 || defined _WIN64
    ptr_SpawnProcess SpawnProcess = reinterpret_cast<ptr_SpawnProcess>(GetProcAddress(Lib, _T("SmartSpawnProcess")));
    ptr_SpawnClient SpawnClient = reinterpret_cast<ptr_SpawnClient>(GetProcAddress(Lib, _T("SmartSpawnClient")));
    #else
    ptr_SpawnProcess SpawnProcess = reinterpret_cast<ptr_SpawnProcess>(dlsym(Lib, _T("SmartSpawnProcess")));
    ptr_SpawnClient SpawnClient = reinterpret_cast<ptr_SpawnClient>(dlsym(Lib, _T("SmartSpawnClient")));
    #endif
    
    SpawnProcess(...);
    SpawnClient(....);
        RunScript(...);
}
Oo I never knew that I thought it was Simba specific nifty

06-02-2013, 10:57 PM
Lalaji

Can I suggest an additional feature I don't think smart has right now?

Add an option to start smart minimized, whether its the first time a script starts, or the sixhourfix starting a new client.

Thanks,
Let me know what you think :)
06-03-2013, 04:21 PM
Ashaman88

Well got LRC script working using OGL detection for the deposits and banking - minimal edits to script. This is pretty cool! I'll post a version of it once I transfer more of the script over to OGL
06-03-2013, 05:06 PM
Rich

Quote:

Originally Posted by Lalaji

Can I suggest an additional feature I don't think smart has right now?

Add an option to start smart minimized, whether its the first time a script starts, or the sixhourfix starting a new client.

Thanks,
Let me know what you think :)

Simple idea but definitely a good one. It'd be nicer if SMART just started in the background.
06-04-2013, 01:52 AM
Sawyer

Quote:

Originally Posted by Ashaman88

Well got LRC script working using OGL detection for the deposits and banking - minimal edits to script. This is pretty cool! I'll post a version of it once I transfer more of the script over to OGL

I am VERY interested in OGL scripting. Where can I learn more of it?
06-04-2013, 02:03 AM
Flight

Quote:

Originally Posted by Sawyer

I am VERY interested in OGL scripting. Where can I learn more of it?

http://villavu.com/forum/showthread.php?t=88242
^ That's the development thread but don't use the OpenGL32.dll or the GLHook.dll from that one, it's outdated. Use the one on the OP of this thread.

I don't think I'll be working with OGL as I've found over time I get hit with an intense amount of lag. I wonder if there's a memory leak somewhere. I've found the more models that are on the MS the slower everything runs. It's actually disappointing. :(
06-04-2013, 02:11 AM
Brandon

Quote:

Originally Posted by Flight

http://villavu.com/forum/showthread.php?t=88242
^ That's the development thread but don't use the OpenGL32.dll or the GLHook.dll from that one, it's outdated. Use the one on the OP of this thread.

I don't think I'll be working with OGL as I've found over time I get hit with an intense amount of lag. I wonder if there's a memory leak somewhere. I've found the more models that are on the MS the slower everything runs. It's actually disappointing. :(

Only if you display the drawings I found this. Such as displaying a ton of ID's at once. No leaks though because I didn't use any pointers. I'm actually re-writing the DLL from scratch and trying to get Smart to do the drawing instead (most likely impossible so I gotta find another way). Also cut down on the massive amount of code that draws the ID's.. It took me 27 lines previously to draw one ID and now 5 lines.

Other than that, I think the above suggested Ideas are good. I might just make it minimize to tray or something. Iunno. We'll see.
06-04-2013, 02:17 AM
Flight

Then this might be a clue as I only switched on the overlay once in the beginning while acquiring some model IDs, after that I turned it off, or atleast I thought, yet still things slowed over time. Do you think perhaps something in the debugging isn't being switched off when it's supposed to? Only a thought...
06-04-2013, 02:31 AM
Le Jingle

I know Brandon will either snipe me and/or post a much more detailed and complete post, but try the stride function that was mention via him (Brandon) a couple posts back in this thread. You can display different items via different strides, and also customize which groups/specifics you want painted too.

Hope this helps,
Lj
06-04-2013, 02:55 AM
Sawyer

Quote:

Originally Posted by Flight

http://villavu.com/forum/showthread.php?t=88242
^ That's the development thread but don't use the OpenGL32.dll or the GLHook.dll from that one, it's outdated. Use the one on the OP of this thread.

I don't think I'll be working with OGL as I've found over time I get hit with an intense amount of lag. I wonder if there's a memory leak somewhere. I've found the more models that are on the MS the slower everything runs. It's actually disappointing. :(

Alright Thanks :) I'll read up on it.
06-04-2013, 02:58 AM
Ashaman88

Quote:

Originally Posted by Sawyer

Alright Thanks :) I'll read up on it.

And just read through his include for a place to start
06-04-2013, 09:32 PM
Sawyer

Quote:

Originally Posted by Ashaman88

And just read through his include for a place to start

What Lape include am I supposed to install?
06-04-2013, 09:40 PM
rj

What would be the difference between directx intercepting and OpenGL?
06-04-2013, 10:40 PM
NKN

Quote:

Originally Posted by Officer Barbrady

What would be the difference between directx intercepting and OpenGL?

One is intercepting DirectX and the other is OpenGL?
06-04-2013, 10:44 PM
rj

Quote:

Originally Posted by NKN

One is intercepting DirectX and the other is OpenGL?

...is one faster then the other?
06-04-2013, 10:45 PM
NKN

Quote:

Originally Posted by Officer Barbrady

...is one faster then the other?

http://i511.photobucket.com/albums/s...8/facedesk.jpg
06-05-2013, 12:45 AM
Flight

Quote:

Originally Posted by Officer Barbrady

...is one faster then the other?

I doubt speed would differ that much. If I had to guess I'd say the difference between intercepting the two is very little; Brandon could probably tell you more about it.

Also, on that note, @Brandon, do you know how Mato was able to use a series of mathematical procedures, in his DX interception, to accurately calculate the player's current tile position? Could the same be done in OpenGL?
06-05-2013, 12:52 AM
euphemism

Quote:

Originally Posted by Flight

do you know how Mato was able to use a series of mathematical procedures, in his DX interception, to accurately calculate the player's current tile position? Could the same be done in OpenGL?

He mentioned this quite a while back in his GLHook thread, yes it is possible. Brandon was working on it with someone, but as I recall it was quite difficult and his partner-in-crime left and so Brandon stopped working on it.
06-05-2013, 02:13 AM
Brandon

Quote:

Originally Posted by Flight

I doubt speed would differ that much. If I had to guess I'd say the difference between intercepting the two is very little; Brandon could probably tell you more about it.

Also, on that note, @Brandon, do you know how Mato was able to use a series of mathematical procedures, in his DX interception, to accurately calculate the player's current tile position? Could the same be done in OpenGL?

Quote:

Originally Posted by euphemism

He mentioned this quite a while back in his GLHook thread, yes it is possible. Brandon was working on it with someone, but as I recall it was quite difficult and his partner-in-crime left and so Brandon stopped working on it.

This. My partner in crime is still alive and talks to me on skype and stuff almost every other day. I can do the walking by myself but it'll take quite some time and probably would be 100% accurate but it'd work like SPS assuming I can get the code down.. Its HARD though! Believe me when I say its hard.

Uhh Mato's walking uses the compass angle and an exact point given by the scripter. If the user is not on that EXACT spot, it will not work.. Even if one pixel off, it will not work.

And yeah, OpenGL is not different from directx at all. Infact, I'm fairly certain that I can look at the DirectX calls and write a plugin that hooks that too. The only thing I hate is that I'd have to use all Microsoft API's.. I used directx before but never actually "hooked" it.. A debugger would be nice I guess if I knew one? Meh for OpenGL, I use 3 from best to worst: gDebugger, OllyDbg, glIntercept..

I can get a map from OpenGL and figure out the player's exact position within that map. Thing is, I don't have a way of knowing whether the map is varrock or what not. Its a local position and to walk, using unique points, you need a "Global" position. A hash of the map is difficult because many maps have the same average RBG colour.

My other idea was to use Assembly or a debugger and find where the RS Camera info is stored or write a shader emulator. If I get that Cam info, walking will be as accurate as reflection/injection.

Show 40 post(s) from this thread on one page

All times are GMT. The time now is 08:03 AM.