TVP 7.4 Develop & TVP 7.7 Stable Full Distribution

[Olddies]

Oh I thought you were hosting Fusion32's server, that's why I was kinda confused seeing it so stable right out of the gate. But I didnt mean anything bad! I was just surprised if the public RE sources worked out of the box because even Fusion mentions that they're probably not the best for use in a live production out of the box.

Try the fusion32 repo and create some dummy bots to login and do some actions ingame, i did with 422 dummy bots as players doing basic things, moving, attacking, and some doing spells the server didnt crash

 

[Ianoyo]

Try the fusion32 repo and create some dummy bots to login and do some actions ingame, i did with 422 dummy bots as players doing basic things, moving, attacking, and some doing spells the server didnt crash

It's amazing how many issues/crashes 500 real players can discover compared to 500 bots programmed to run the same methods over and over.

However, the engine could still be solid and few problems exist, it's just unlikely that there are 0 issues.

 

[Olddies]

It's amazing how many issues/crashes 500 real players can discover compared to 500 bots programmed to run the same methods over and over.

However, the engine could still be solid and few problems exist, it's just unlikely that there are 0 issues.

Im just talking by myself you guys can do any stress test that you want or need to figure out whatever you feel simple as that

 

[Jpstafe]

 

[Evil Mark]

Try the fusion32 repo and create some dummy bots to login and do some actions ingame, i did with 422 dummy bots as players doing basic things, moving, attacking, and some doing spells the server didnt crash

I see, well wish you the best still. I wont try it, we have our own engine and I trust in Neptunos very capable hands more than anyone elses. He's the first one to reverse engineer the engine and he knows it like the back of his hand due to the fact that he did not reconstruct the entire thing with AI

 

[M4rcin]

Im just talking by myself you guys can do any stress test that you want or need to figure out whatever you feel simple as that

The problem is, it’s not exactly easy to do exploratory testing for 500 players when you’re just one person
You can’t really get around that - everyone’s perception is limited in some way, and anyone who thinks otherwise is probably even more limited.
You can stress test it, but it will likely be only a couple of execution paths.
This could check if your resources are managed correctly (in that exact execution path), but that's pretty much it.

 

[Olddies]

The problem is, it’s not exactly easy to do exploratory testing for 500 players when you’re just one person
You can’t really get around that - everyone’s perception is limited in some way, and anyone who thinks otherwise is probably even more limited.
You can stress test it, but it will likely be only a couple of execution paths.
This could check if your resources are managed correctly (in that exact execution path), but that's pretty much it.

in that case i just have to wait

 

[benner]

Hey guys, how’s it going?

I’m kinda “new” to the community even though this account is pretty old. I’ve always liked OTServers since the good old scuffed days, but I kinda stopped for a long time. Still, I’m always coming and going, reading threads, checking ideas and projects from the shadows. I don’t work with this, don’t plan to make money with it, it’s just a hobby to scratch the nostalgia itch.

Right now I know how to compile, mess with some basic stuff and solve simpler issues. But just to be clear: memory leaks, heavy optimization and that hardcore low-level stuff? I have zero clue. And I can only compile on Windows, learned that not long ago, so I’m pretty limited when it comes to environment too.

Anyway, the reason for this post is a more “conceptual” question:
at the end of the day, what’s the best base to use as a reference for developing old school servers in the 7.4 (or even 7.1) to 7.7 range?

What are the limitations of each base people are using nowadays?

Which base gives you more freedom to mess around, create systems, customize stuff without feeling too restricted?

And which one is more “locked”, more limited, that tends to cause more headache or kills creativity?

I’m not sure if I’m asking this in the right thread, because I saw it started with a specific purpose, so if I’m off-topic, sorry about that. But I think these questions are valid, and a lot of people in the same situation as me (hobby, learning, curiosity) probably wonder the same things.


Thanks to anyone who can shed some light on this, any opinion already helps a lot. I’m more lost than a newbie in Rookgaard without a map.

 

[Tofame]

You can't say you got edge over TFS, if you removed lua scripting for some custom rl engine based one.
Both that rl engine and TFS are inferior to modern standards - which is to use sol for lua bridge to C++.
It's unquestionable, as I'm speaking facts here (benchmarked facts).

 

[M4rcin]

You can't say you got edge over TFS, if you removed lua scripting for some custom rl engine based one.
Both that rl engine and TFS are inferior to modern standards - which is to use sol for lua bridge to C++.
It's unquestionable, as I'm speaking facts here (benchmarked facts).

Although I agree with Lua having an advantage over .dat, TFS Lua api won't be a bottleneck in any probable use case.
Also, if you are speaking facts, why not link them?

Hey guys, how’s it going?

I’m kinda “new” to the community even though this account is pretty old. I’ve always liked OTServers since the good old scuffed days, but I kinda stopped for a long time. Still, I’m always coming and going, reading threads, checking ideas and projects from the shadows. I don’t work with this, don’t plan to make money with it, it’s just a hobby to scratch the nostalgia itch.

Right now I know how to compile, mess with some basic stuff and solve simpler issues. But just to be clear: memory leaks, heavy optimization and that hardcore low-level stuff? I have zero clue. And I can only compile on Windows, learned that not long ago, so I’m pretty limited when it comes to environment too.

Anyway, the reason for this post is a more “conceptual” question:
at the end of the day, what’s the best base to use as a reference for developing old school servers in the 7.4 (or even 7.1) to 7.7 range?

What are the limitations of each base people are using nowadays?

Which base gives you more freedom to mess around, create systems, customize stuff without feeling too restricted?

And which one is more “locked”, more limited, that tends to cause more headache or kills creativity?

I’m not sure if I’m asking this in the right thread, because I saw it started with a specific purpose, so if I’m off-topic, sorry about that. But I think these questions are valid, and a lot of people in the same situation as me (hobby, learning, curiosity) probably wonder the same things.


Thanks to anyone who can shed some light on this, any opinion already helps a lot. I’m more lost than a newbie in Rookgaard without a map.

I think your question is perfectly valid.
There seem to be four ways currently, each has pros and cons.

TVP (current thread):
Branched out of early TFS, close to the 7.X, but far from perfect, badly written, has Lua API, maybe even revscripts? so you can do a lot using just that. Mechanic completeness would be 5/10.
It is unstable (which has been proved by many projects), edited by someone very unprofessional, and his poor changes make C++ extendability even worse.
Taking that path in its current state, without a good understanding of C++, I consider cutting corners, which will ultimately lead to burnout.
A ton of ninja crashes, issues you will have a hard time understanding and debugging, especially as a newcomer.

Reversed RealOT:
Cipsoft engine from 2005, the code is questionable, but there are enthusiasts.
Although there might be different perceptions about code quality, it is a fact that, outside the regular game mechanics, you will have a terrible time extending it, given that you are not familiar with C++, and familiarity alone might still not be enough.
It is the most complete distribution you can find mechanicwise, in the end, it is the engine Cipsoft used back in 7.7 times.
Stabilitywise, we dont know yet - it might be fine, or it might have issues - definitely more stable than TVP, but as the Reverse-engineered project is pretty new, noone had actual players play it yet.

TFS Main:
Path I took as a rookie, and I don't regret it, so I might be a little biased toward it.
Most extendable of all your options, very capable Lua interface, decent support, and actively developed.
The catch is that it is targeting Tibia 12~, so you gotta do quite some work to introduce 7.X mechanics that differ in many scenarios.
Although that task is objectively the easiest - all you gotta do is realise what you are missing, change sprites, and make adjustments - no rewrites, no huge headaches, and actual protocol downgrade is not required at all, as long as you use OTC (which again, is the most extendable tibia client).
If you are uncertain about a mechanic, you can look it up in Reversed RealOT, if it crashes, I am sure you will find people ready to help you, all you gotta do is create an Issue on the github, and make sure you are verbose and follow all rules.

Create your own engine:
Some developers argue that if you don't write your own engine, you are not a true developer.
But developers argue about a ton of stupid stuff, you don't always have to listen to them.
Although that experience would probably be the most beneficial to you - it will be a long time till you will see the effect of your work.

Big DO's:
Any path you choose, make sure you ask questions, have an open mind, and accept that we make mistakes, remember shit backwards, and sometimes make bad decisions, so don't bash yourself if you make one too.
A big part of OT development is about the journey, not about the rabbit.

Big DON'Ts:
No canary, no mehah OTC, and no TVP (in its current state). Each of these repos is, or used to be, a circle of mediocre programmers who will ban you from their Discord group if you point out bad practices that ultimately make each a timesink they are.

TL; DR
If you want 7.7 and no new mechanics (but custom old mechanics), go with RealOT and Cip client.
If you want to create a new definition of 7.X, TFS Main and OTCV8 with changed sprites, a couple of months of adjustments and you will be quite close. You get to learn C++ slowly and end up with the most complete Lua API.

TVP is not a real option since it is unstable right as is, although this might change in the near future.

 

[Error 502]

TFS Main:
Path I took as a rookie, and I don't regret it, so I might be a little biased toward it.
Most extendable of all your options, very capable Lua interface, decent support, and actively developed....

tfs main is dead on GitHub.


and I don't know of any server that uses the main repo 1.6 or 1.7

 

[M4rcin]

Aha, I hurt someone’s feelings
It’s not dead, it’s being actively contributed to, with contributions even today.



Anyway, I dont think it changes much if you don’t know something @Error 502.
I guess there is a ton of users unaware of various things, yet they can realise that no one asked.

Your post gives out an attitude of someone who got baited into one of the distributions I mentioned, hope it works out well for you regardless.

 

[Beatss]

Although I agree with Lua having an advantage over .dat, TFS Lua api won't be a bottleneck in any probable use case.
Also, if you are speaking facts, why not link them?


I think your question is perfectly valid.
There seem to be four ways currently, each has pros and cons.

TVP (current thread):
Branched out of early TFS, close to the 7.X, but far from perfect, badly written, has Lua API, maybe even revscripts? so you can do a lot using just that. Mechanic completeness would be 5/10.
It is unstable (which has been proved by many projects), edited by someone very unprofessional, and his poor changes make C++ extendability even worse.
Taking that path in its current state, without a good understanding of C++, I consider cutting corners, which will ultimately lead to burnout.
A ton of ninja crashes, issues you will have a hard time understanding and debugging, especially as a newcomer.

Reversed RealOT:
Cipsoft engine from 2005, the code is questionable, but there are enthusiasts.
Although there might be different perceptions about code quality, it is a fact that, outside the regular game mechanics, you will have a terrible time extending it, given that you are not familiar with C++, and familiarity alone might still not be enough.
It is the most complete distribution you can find mechanicwise, in the end, it is the engine Cipsoft used back in 7.7 times.
Stabilitywise, we dont know yet - it might be fine, or it might have issues - definitely more stable than TVP, but as the Reverse-engineered project is pretty new, noone had actual players play it yet.

TFS Main:
Path I took as a rookie, and I don't regret it, so I might be a little biased toward it.
Most extendable of all your options, very capable Lua interface, decent support, and actively developed.
The catch is that it is targeting Tibia 12~, so you gotta do quite some work to introduce 7.X mechanics that differ in many scenarios.
Although that task is objectively the easiest - all you gotta do is realise what you are missing, change sprites, and make adjustments - no rewrites, no huge headaches, and actual protocol downgrade is not required at all, as long as you use OTC (which again, is the most extendable tibia client).
If you are uncertain about a mechanic, you can look it up in Reversed RealOT, if it crashes, I am sure you will find people ready to help you, all you gotta do is create an Issue on the github, and make sure you are verbose and follow all rules.

Create your own engine:
Some developers argue that if you don't write your own engine, you are not a true developer.
But developers argue about a ton of stupid stuff, you don't always have to listen to them.
Although that experience would probably be the most beneficial to you - it will be a long time till you will see the effect of your work.

Big DO's:
Any path you choose, make sure you ask questions, have an open mind, and accept that we make mistakes, remember shit backwards, and sometimes make bad decisions, so don't bash yourself if you make one too.
A big part of OT development is about the journey, not about the rabbit.

Big DON'Ts:
No canary, no mehah OTC, and no TVP (in its current state). Each of these repos is, or used to be, a circle of mediocre programmers who will ban you from their Discord group if you point out bad practices that ultimately make each a timesink they are.

TL; DR
If you want 7.7 and no new mechanics (but custom old mechanics), go with RealOT and Cip client.
If you want to create a new definition of 7.X, TFS Main and OTCV8 with changed sprites, a couple of months of adjustments and you will be quite close. You get to learn C++ slowly and end up with the most complete Lua API.

TVP is not a real option since it is unstable right as is, although this might change in the near future.

Funny how some people repeat opinions as if they were universal truths, while the numbers right in front of them tell a completely different story. The table speaks for itself: the projects that are actually alive today are the ones following the current Tibia versions, maintaining continuous development, and evolving instead of staying fossilized in 7.x. It’s no coincidence that Canary 3.0 sits at the top — real stability, active support, and an engine that doesn’t try to survive stuck in 2005.


This whole idea of “TFS Main being active” is basically urban legend at this point. The GitHub repo is practically dead, PRs rotting for months, minimal activity. Calling that “active development” in 2025 requires a level of imagination even Cipsoft doesn’t use in their trailers.


The claim that “Canary shouldn’t be an option because it’s unstable” was true many years ago. Today, people repeat it out of habit, not because they have any technical basis. The codebase is organized, refactored, modern, performance-oriented — not a pile of rushed patches or chaotic commits.


And just to be absolutely clear: our focus has always been the current Tibia versions. We never promised official support for 7.x. If someone wants 7.4, 7.6, 7.7, etc., they are free to fork, adapt, and go their own way — just like any modern open-source project. The code is open for exactly that reason. But if someone joins the Discord demanding support for something we never offered, ignores documentation, invents theories, and ends up banned… well, that’s on them. Tantrums are not a substitute for technical arguments.


While some people live off nostalgia, others brag about engine “purity” without having a single real server running, we stay focused on what actually matters: stability, continuous development, modern C++, and an ecosystem that moves forward instead of freezing in time. If you want to work with old versions, you can — on your own. If you want a project that is alive and evolving, you already know where to find it.


Of course, there will always be those who prefer complaining instead of reading, demanding instead of contributing, and then acting surprised when that behavior has consequences. At that point, there’s not much I can do for you.

 

[M4rcin]

If your PR is bad and you never improve it, yes, it will never be merged.
That’s the main difference between TFS and Canary

Anyway, this is a topic of 7.X distrib, and user asked for 7.X distrib - you go on rant about people stuck in past, and I am not reading past.



Keep using canary, no one here is going to stop you and I really hope you don’t get to learn the truth the hard way … but please stop the offtop.

 

[Firemac]

Although I agree with Lua having an advantage over .dat, TFS Lua api won't be a bottleneck in any probable use case.
Also, if you are speaking facts, why not link them?


I think your question is perfectly valid.
There seem to be four ways currently, each has pros and cons.

TVP (current thread):
Branched out of early TFS, close to the 7.X, but far from perfect, badly written, has Lua API, maybe even revscripts? so you can do a lot using just that. Mechanic completeness would be 5/10.
It is unstable (which has been proved by many projects), edited by someone very unprofessional, and his poor changes make C++ extendability even worse.
Taking that path in its current state, without a good understanding of C++, I consider cutting corners, which will ultimately lead to burnout.
A ton of ninja crashes, issues you will have a hard time understanding and debugging, especially as a newcomer.

Reversed RealOT:
Cipsoft engine from 2005, the code is questionable, but there are enthusiasts.
Although there might be different perceptions about code quality, it is a fact that, outside the regular game mechanics, you will have a terrible time extending it, given that you are not familiar with C++, and familiarity alone might still not be enough.
It is the most complete distribution you can find mechanicwise, in the end, it is the engine Cipsoft used back in 7.7 times.
Stabilitywise, we dont know yet - it might be fine, or it might have issues - definitely more stable than TVP, but as the Reverse-engineered project is pretty new, noone had actual players play it yet.

TFS Main:
Path I took as a rookie, and I don't regret it, so I might be a little biased toward it.
Most extendable of all your options, very capable Lua interface, decent support, and actively developed.
The catch is that it is targeting Tibia 12~, so you gotta do quite some work to introduce 7.X mechanics that differ in many scenarios.
Although that task is objectively the easiest - all you gotta do is realise what you are missing, change sprites, and make adjustments - no rewrites, no huge headaches, and actual protocol downgrade is not required at all, as long as you use OTC (which again, is the most extendable tibia client).
If you are uncertain about a mechanic, you can look it up in Reversed RealOT, if it crashes, I am sure you will find people ready to help you, all you gotta do is create an Issue on the github, and make sure you are verbose and follow all rules.

Create your own engine:
Some developers argue that if you don't write your own engine, you are not a true developer.
But developers argue about a ton of stupid stuff, you don't always have to listen to them.
Although that experience would probably be the most beneficial to you - it will be a long time till you will see the effect of your work.

Big DO's:
Any path you choose, make sure you ask questions, have an open mind, and accept that we make mistakes, remember shit backwards, and sometimes make bad decisions, so don't bash yourself if you make one too.
A big part of OT development is about the journey, not about the rabbit.

Big DON'Ts:
No canary, no mehah OTC, and no TVP (in its current state). Each of these repos is, or used to be, a circle of mediocre programmers who will ban you from their Discord group if you point out bad practices that ultimately make each a timesink they are.

TL; DR
If you want 7.7 and no new mechanics (but custom old mechanics), go with RealOT and Cip client.
If you want to create a new definition of 7.X, TFS Main and OTCV8 with changed sprites, a couple of months of adjustments and you will be quite close. You get to learn C++ slowly and end up with the most complete Lua API.

TVP is not a real option since it is unstable right as is, although this might change in the near future.

Out of curiosity, why do you give this new player who is genuinely interested in getting help a 100% copy pasted answer from Chat GPT? You are aware that your AI answer has several flaws since you do not even know how to use any prompts but just plainly ask it?

How much real experience do you have in reality and how much of what you do and write are you basing on AI?

 

[M4rcin]

Out of curiosity, why do you give this new player who is genuinely interested in getting help a 100% copy pasted answer from Chat GPT? You are aware that your AI answer has several flaws since you do not even know how to use any prompts but just plainly ask it?

How much real experience do you have in reality and how much of what you do and write are you basing on AI?

No LLM was used to write either of my posts.
And to answer your last question - All of it, and none of it

 

[Evil Mark]

No LLM was used to write either of my posts.
And to answer your last question - All of it, and none of it

You can ignore Firemac. He does everything to increase his post count. He also believes anyone able to talk and write properly in english is using AI. Don't waste your energy.

 

[Zyzz immortal]

You can ignore Firemac.

This is the guy who was attacking Kay that he is jealous of Ezz success with Tibiara xD

 

[Evil Mark]

I've added the git repo to the main thread.

 

[Danihellx]

I've added the git repo to the main thread.

@Evil Mark
Hello Mark, first of all, thank you for making the TVP base available to the community.


I noticed that today you created a GitHub repository for the community with the TVP Item Editor and TVP Monster Editor, but after a few minutes you deleted that repository and created another one called “TVP-OTland.”


However, you didn’t upload the item and monster editors to the new GitHub repository. From what I can tell, the editors there aren’t the same ones that are attached directly to the forum with download links.


My question is: are you planning to release those files, or will you keep them private?