Teemu

[Krice]

The automap actually works as a look command for terrain at the same time, because I made it show what the thing is at the cursor. The number of tiles is growing, it's now 33. I'm going through level themes and checking if the automap is missing something and it always seems to. Like the hatch which currently shows as a door (because it is a door object), but it just doesn't look right so I have to make a frame for that, too. Also, each special portal type really needs a dedicated frame. On top of that I have to use the ancient 0.85 version of Brick Atelier to draw these, because my inability to finish the new version.

 

[Krice]

I think it would be a good idea to pack (and encrypt) the text data on files, but I'm not sure how to do it. For text data some kind of dictionary based routine would work, but I'm wondering how they determine the pieces of text. I guess the easiest way would be create an entry for each word, but some things like "It is" or "You are" should be one entry, because they repeat often. I bet there are libraries for this, but I'd hate to add another dependency for this.

 

[Krice]

The next difficult feature I'm trying to finish is the shopkeeper. There is no support for that in the "engine", so it's interesting to figure out how to implement it. My plan is different from a "typical" shop where you walk inside the shop, pick up items and buy them. I'm planning to make the shopkeeper pick up items and give them to the player.

 

[Tavernking]

Will this be a paid game or free?

 

[Krice]

Free, but if someone knows how I could monetize roguelike development then let me know.

 

[Tavernking]

Free, but if someone knows how I could monetize roguelike development then let me know.

It isn't like any other game? Put it on Steam? Though then you have to worry about marketing - knowing a game is free will attract a lot of players automatically.
I only ask because ascii games aren't my thing at all, but I'd try it if there was a demo or it was free.

 

[Krice]

It's a "lite" roguelike with crappy graphics + ascii, no sounds. Do people really ask money for that kind of stuff? Then again, some people ask money when they "create content" on youtube...

 

[Krice]

I'm reverting my opinion about using "free" C-style functions in a OOP based code. It does "work" obviously, but it's not a good idea after all. One of the ideas of OOP is the rigid structure of class hierarchy and while some people think it's a bad thing it's just not. I think it's easier to keep track of everything when you have to put stuff in a class. It's quite difficult to create good class hierarchies but if you get it right it's just going to work and keep the code clean. I'm actually having fun reverting functions back to classes.

 

[Krice]

the links are broken

Only the github link is broken (removed the link), because I removed the source code from github. And to be honest, I'm probably not going to release the source for the latest version unless it becomes a popular game which I doubt very much. There will be a 64-bit Windows version and that's it.

 

[Krice]

It should work, it's just a normal web page. Maybe your country/government/browser has restricted it, who knows these days.

 

[Arbaces]

No, I can't access it either. It was up a couple weeks ago. Also, please consider a Linux build at least if you're not going to release source code. For a roguelike to not run on *nix is major decline.

 

[Krice]

Or maybe mbnet is restricting traffic from some countries. However, it's a problem to solve when the new version is released. I could do some kind of linux build, but I have no idea how to do it. Does linux actually have binary builds? I've always thought they compile everything from source, because linux distrubutions are not compatible?

 

[Arbaces]

Building portable Linux binaries isn't black magic, but it's not common knowledge either. The short version is: just dynamically link against old libraries. If your libraries are newer, then it won't run on less cutting edge distros. The biggest catch is usually glibc. My friend used to build all his releases on Slackware for this reason.

 

[rusty_shackleford]

the shorter version is build a statically linked binary

 

[Arbaces]

the shorter version is build a statically linked binary

Please God, no! There's a reason flatpack, appimage, and all the others exist. It's because static linkage sucks. It doesn't even solve the problem of portability-- you're still at the mercy of a lot of moving parts in the system. On top of that, if something major changes (like X11 to Wayland migration) you'll have to recompile when you would have had no problem if you dynamically linked the middleware. I've run dynamically linked Linux binaries from the 90s with no issue except for OSS having been replaced by ALSA/Pulseaudio. I doubt a statically linked program would be usable for a decade, let alone two.

 

[rusty_shackleford]

the shorter version is build a statically linked binary

Please God, no! There's a reason flatpack, appimage, and all the others exist. It's because static linkage sucks. It doesn't even solve the problem of portability-- you're still at the mercy of a lot of moving parts in the system. On top of that, if something major changes (like X11 to Wayland migration) you'll have to recompile when you would have had no problem if you dynamically linked the middleware. I've run dynamically linked Linux binaries from the 90s with no issue except for OSS having been replaced by ALSA/Pulseaudio. I doubt a statically linked program would be usable for a decade, let alone two.

there's statically linked nethack binaries from the early 90s that still work just fine
the dynamically linked ones don't because the X11 libs it depends on don't even exist anymore, btw

 

[Krice]

How about the directory structure? This game has 'data' and 'savegame' folders in the root of the game, so no fancy user directory stuff. Is it fine in linux? I think saving some data in user folder sucks, it's sometimes happening in Windows as well in Users (or in finnish Käyttäjät) ... directories. Ech.. I don't like it, because then everything programs put in there stays there forever unless you remove them manually. Ech.

 

[Krice]

I'm trying a new kind of strategy to reach the new version. The plan is to map (unfinished) content to classes and keep working only on those classes. I think it will help me focus on the important stuff, but we'll see how this works.

 

[Krice]

I already knew this, but it was funny to have a data about this, which is that out of 250 total classes only 23 has unfinished stuff and most of them are related to rpg system. I need to cheese it out somehow.

 

[Krice]

Finally updated to Visual Studio 2022 (you can still keep 2019 which is cool) and this is the first project moved to 2022. I was watching a "review" of 2022 where the dude said this is faster than 2019, because it's 64-bit. You know where this is going. Yeah, 2022 is noticeably slower than 2019, at least with C++. Compiling is slower, the editor feels more sluggish etc. What I don't get why some random youtube "reviewers" lie about this, it's not like this is some hot stuff. It's an IDE and the community version is free. It's not that bad and you can turn off lot of features you don't need. The funny thing is that it's actually the same as 2019, but they somehow managed to make it slower. I haven't found any new features that would blow my mind.

 

[Krice]

I have some kind of idea for the rpg system, well, finally... Each damage type has base damage and extra damage. They are linked to items mostly. Combat skills would improve the probability of getting the max from the dice. I'm not sure yet how to implement accuracy or how often you hit or miss, maybe zero damage roll would be a miss? It could be funny to increase extra damage if you are drunk, but make it otherwise negative for hitting probability.

 

[Krice]

It looks like MBnet is ending so Teemu's homepage need to be relocated, I think this time on a real, working internet web page. It's just such a pain trying to find something that doesn't cost a lot, but still works. Ech.

 

[Krice]

I made a decision to focus only on this project to get it ready, a decision that can fail... For some reason this is the funniest project to work on, because there is more freedom to do crazy stuff. Lately I've been refactoring symbol decisions for items and other stuff, because it was quite random. With ascii you run out of glyphs quite fast, but it is a limit you have to work with. From non-letters only ; : ` and , are unused, others are set to one or more object. Unlike in many roguelikes some letters are for terrain objects like T for tree and Y for palm tree. Colors are another story and I think some of the colors are maybe too close to each other, because I didn't want to use too bright palette.

 

[Krice]

I had to add a new game object type for powder and liquid containers. There is an option to use existing objects, but they would then have some extra features added which is always a bad idea. I have done that mistake before and it just doesn't work well when you have "empty" data for some objects etc. First I thought this would be painful, because OOP, but it was actually quite easy in this case and took only 32 minutes. The engine is somewhat data-driven and got in a new object type without too much hassle. Objects are stored and handled (in display map etc.) by their base class handle, so that's also one reason why it was easy.

 

[Krice]

The source code is getting so good, that I can no longer fix it.. I have to finish the rpg system. It's pretty much the only unfinished part anyway.