Iv'e been wondering, is there a way to read/write to a text file using spellscript? If not could you implement it if it is possible please, it would be really useful for certain things (donators etc).
I haven't tried it, but you should have access to all the standard Java file stuff. You'll have to tell me if you can get it working
You'll also have to be careful of client/server discrepancies since scripts are executed on both ends. Iow, don't accidentally place files on your clients' computers (or give script editing permission to anyone irresponsible.) This can be checked with SERVER.isClient() and SERVER.isRemote().
I played with warstomp, it was my computer because it works great now.
Yeah, I figured as much, as I had tested it and couldn't get any issues. Good to know you got it working.
Also, the ripple-y War Stomp that ripples out was originally what I wanted, but I thought Shockwave was both cooler and easier to implement Based on prior testing, however, something like War Stomp may be very laggy, as physic-ized blocks tend to be so. Would at least half to cap it so at higher levels, players can't grief servers with it, lol. But now that there's an entire skill tree, I don't see why I can't throw it in there
As for summoning stuff, I was hesitant in the past to add them due to AI problems and added server lag (I AM NECROMANCER WITH 50 SKELLINGTONS), but something like familiars may be possible in the future.
What if in the JSON we could specify where all the skills go with a coordinate, for example
#These are the coordinates for the skill.
"skillSet": [
"warrior_leap"(1,0),
"warrior_bash"(3,0),
"warrior_leash"(2,0),
"warrior_charge"(2, 1),
"warrior_whirlwind"(4, 0),
"warrior_shockwave"(3, 1)
],
#These are the coordinates for the lines, and the information for the Prereqs. They are read in left to right order, and the lines hook in in said order.
"skillPrereqs&drawLines": {
"warrior_shockwave":["warrior_whirlwind"];(4,0);(4,1);(3,1),
"warrior_charge": ["warrior_leash"];(2,0);(2,1),
}
The coordinates correspond with a given point on the chart. So, the coordinates after "warrior_leap"((1,0)) mean Leap will be at spot (1,0)
Here is a chart, for reference:
The code given, using the coordinates given would give out a Skill Tree as follows:
Thoughts?
First off, JSON is a strict format: unfortunately, coords would likely have to be input as a string or a list of numbers (assuming the user understands every two numbers constitute a coordinate). That seems pretty messy, though. Also, imposing a grid format on the tree is a good idea since it'd be easier for users to reason about, but I'm not sure if its worth limiting the choice of possible locations.
Doing it this way, we'd just be adding coords as additional elements of lists. We could also include the coords inside strings (using the fact that the innards of strings can be formatted however we want), but this seems a touch cleaner.
I'll have to think about it a good bit-- it might be better to let the user specify exact pixels rather than grid slots; then again, I could even give the user options either way. I do want things to be as self-evident and legible in hsmConfig as possible, so it may be better to let users specify them in a completely separate property.
1. How do we import our java classes? And, are we able to call upon the methods and globals of spellscript in said java classes?
2. I'm getting an error; I'm not seeing the blocks moving up and down for shockwave (warstomp was a cooler name ).
3. Here's an idea for edited skill tree GUI:
Changes:
Straight lines, to avoid confusion.
Numbers, to help tell you what level they are at, 0 meaning no levels.
Red 0's mean that it has a prereq
(in my opinion, it's easier to read through)
EDIT: oops, messed up, got a line going the wrong way.
EDIT AGAIN: Fixed it!
You currently can only import Java classes that are in the Minecraft/HSM project already-- this is part of the reason I want to put out another update very soon. Allowing users to use whatever classes they want will involve transferring jars over the network, and some other internal guts which will require some work. (But not too much work.)
EDIT: This means everything in java.lang, Google's guava, and a lot more are available.
Shockwave ain't working? Well this is what I get for rushing :/ That's a very strong reason for a fast update.
Skill numbers next to the icon are a strong possibility, I'll just need a day to experiment with it (and other things) so nothing looks cluttered. Didn't have time yesterday to neaten the skill tree GUI much, unfortunately. Yet another reason for a fast update!
The skill trees are dynamically generated based on hsmConfig, which means square-ish lines won't cut it for all possible skill trees (not without changing the approach to the problem). However, finding any good solution for straight or curved lines for a given tree is a theoretically challenging Computer Science problem, one which would be very nice to avoid For now, though, the skills are just sorted in a semi-intelligent fashion and connected directly. This could be avoided by not generating the lines and allowing skill-makers to specify them, but this seems tedious, unsightly, and unnecessary. I suppose providing the option would be nice for you perfectionists out there But I'd rather make the generation smarter than maintain more code.
On the bright side, any strangeness from the lines will probably become less noticeable once there are more spells, so, we shall see...
EDIT: If you or anyone for that matter has an idea for an alternate, simpler method of generating the lines, though, I'd definitely love to hear it.
Yo frizzil love the update! The skill tree is awesome and the new additions like the swings per second and crit chance etc. Just wondering, but how does the skill tree effect the spells when you level them up? iv'e looked through the spells code and i cannot find anything that looks like the skill tree improves the spells.
If you mean spells having an impact on other spells, this is actually a plan, though I don't think I've said anything about it XD Perhaps it's such an inherent part of skill trees that everyone expects it?
There are some things like this I'm aching to do, but I'm not sure if I should do another update early this week or wait until the weekend to let people settle into the new update. I might have to wait for feedback on that one.
EDIT: There's also some bugs I didn't have time to address (I absolutely forced myself to get this update out last night, which wound up being at 5 AM), so their existence will likely be a deciding factor.
To install, you'll additionally want to place "sslib" alongside "spells" (remember to delete your old "spells", or start migrating it all if you've customized anything!)
I'll update the OP tomorrow and hopefully get a video of new stuff up as well
This update's taking a while, but I still plan on getting it out later tonight. I'm doing a lot more balance work than I expected (Spellscript has distracted me from some glaring issues-- well that, and I added the ability to invest levels into each of your skills, which has taken some work). I probably won't be updating the front page until tomorrow when I wake up, and the video's definitely gonna have to til tomorrow at the earliest.
I'll make one tomorrow, hopefully along with a video I hear the last video I made helped people with Spellscript immensely, so I'll definitely make another.
Im currently making an RPG mod for my server only atm. It will be based on a LARP. But no stats. I wanted to ask if you could answer some questions.
1. How can I implement variable melee ranges for different melee weapons?
2. I want classes activated by sashes. Someone also told me items have a hard time interacting with stuff like classes (Warrior, Bard, Wizard, etc.) I wanted spells to cost in materials, and a way to have a pool "mana" or other essence for magic classes. How would I implement these?
3. Levels will come with new abilities and weapons and armor. How would I restrict what I person can wear by class?
If you're using some framework like Forge or MCP, it's a whole 'nother ballpark than what I've been doing for Hack/Mine. You'd have to do some fairly complicated things to get all that working, since it's relatively invasive, which means compatibility is difficult to impossible. You'll have to get creative if this is the case.
EDIT: Iow, you'll likely need to get familiar with bytecode injection (specifically with ASM) and Java's reflection utilities in order design things in a sane manner. Well, if you can call that sane XD
However, if you're not using those tools, then the changes are pretty easy to make with straight-up class edits. This is assuming you're familiar with basic software engineering concepts If not, you should teach yourself using all the free resources available round the interwebz (not to assume whether or not you're already educated.)
EDIT: An idea: do absolutely everything through the "Sash" item. This will minimize any invasiveness dramatically. So for example, whenever you need to know the player's class, scan his armor for the sash. Mana pool? Make it a property on the sash, and persist it with metadata. Of course, you'll want a bunch of static helper methods to reduce the boilerplate (including down-casting the sash itself), but this would probably be easier to maintain than bytecode injecting methods onto EntityPlayer (and probably friendlier to other mods as well.)
Why not try you generating the dungeon itself as you go along it? I do not know how you coded your generation (probably at a call of a forge hook I assume?) but given that you are generating 16x16 chunks, you start at a room that is only 16x16, you go north and and you expand another 16x16 on the north end.
I could help you integrate it if you are willing to get help for a fellow modder as I created a code that could generate "any structure" no matter how big it is that I am willing to share to this great mod.
Sorry, perhaps I didn't explain so clearly.
The problem is imposing higher-level structure. Generating a dungeon "as you go along it" is easy, regardless of size. The problem is that doing so results in a dungeon with almost no organization; it just sprawls without a sense of direction. Hence, in order to introduce structure, you'd have to pre-generate the entire dungeon and cache the result, then lay down the pre-generated chunks of it as world generation occurs (or at least that's one possible solution).
This is not a trivial thing, given that dungeon data would need to be persisted at a level higher than chunks, then referenced in a very efficient manner-- this is part of the reason the Area System was introduced.
That's a damn shame! I'm pretty darn sure that huge multi-chunk generation is possible tho. I believe Forge has some stuffs for worldgen, might be worth looking at that once you port the mod ^^
Multi-chunk generation is possible even in vanilla Minecraft (take mineshafts for example), but they have an even worse problem-- you can't generate structures at a "high level". In other words, you couldn't give the structure a kind of "flow" as you wander through it, or have keys that unlock doors in other areas in such, etc. Generation in one chunk can't retroactively affect the structure in another, since nothing is guaranteed-- every chunk has to be independent in a sense. Thus, you wind up with massive, "wandery" experiences, and an inability to make anything interesting of it. It'd permanently cripple that dungeon's enjoyability.
Spellscript has fully functioning access to generic types at this point (including generic type inference!), as well as import resolution, aliasing, and finally, access to static methods. *However*, generic/static methods have yet to be implementable in the language. So you can access such features introduced from Java, but I still need to add ways to create static and generic methods of your own. It's mostly a matter of deciding on the syntax in a way that's as easy to use as possible, so wish me luck. Static methods will be easy, generics might have to wait until later though...
I've also got a nifty maze-generation script working from the console All thanks to Java integration, my friends.
Also, there's some delay because I'm working on job applications atm. The real world is calling my name, and I must answer! Glad to have Hack/Mine on the resume, though. I'm basically shooting for the stars in the game industry at this point and hoping I get in
This weekend's feeling pretty good for an update. I just gotta make sure everything is seamless, and that I've addressed all the bugs posted since the last update. However, my applications come first, since finding a job is kinda crazy important. Apart from that, though, Spellscript is 97% ready for this update. Woohoo!
Some suggestions I thought of while designing a city on a server:
There should be a certain block (preferably a new, creative only block) that stops mobs from being able to spawn within the chunk that said block is placed in. This hopefully will keep mobs from turning potential cities into a medieval Stalingrad.
An additional block should be a spawn block for NPCs. These NPCs should be able to take all the races and classes that the player can take (albeit weaker), along with a generic "guard" class. These NPCs would attack all hostile mobs, and (if more back story to the races is added) possibly races that are at odds with the NPC's race. The guards behavior would depend on the class, with Warriors and Guards moving forward to attack enemies, and Mages and priests staying at a distance from the mobs to support others. The Ranger should be given more complex tactics, rather than just running away (YES I AM BIASED IN FAVOR OF RANGERS), as there are more tactical possibilities with the traps and ability to run up walls. If not fighting, chasing, or running away from mobs, they should return to standing on top of their spawn block, and only 1 NPC should be alive at one time. This means that organized forces can be formed, and areas will not be infested by NPCs. To keep things from looking bad, the spawn blocks should be able to have different textures, perhaps dependent on the block underneath it or something.
A different kind of NPC should be added: mercenaries. Essentially the same as guard NPCs, but paying them certain amounts of money (dependent on weapons and skills) will bind them to the player for a set amount of time.
The trading system should be changed so that a store keeper can buy more items if in the vicinity of a chest (or similar). Preferably, a similar type of spawn block should be set that spawns and keeps back the NPCs so that shops can be made without them going crazy.
Actually, this is essentially turning into a rant on NPCs, so why not just include a whole bunch of different professions and similar, so a miniature society can be made.
Add different kinds of dungeons. The towers and underground labyrinths are cool, but I'm hoping for something more along the lines of a walled in village filled with traps and corridors leading up to a boss fight that results in either death or high quality items. I'm sure there are plenty of people (me included, provided I can find the time) that would be willing to design dungeons for you.
Welp, that seems like everything for now. Dunno if any of these have been mentioned before, or if you guys think they are awful, but I'll post them anyway
Those are actually some solid ideas, though a few things: firstly, how is a non-mob block better than the existing method of disabling mobs through the command line? (see the wiki for how) While a block is easier to place, a block is also much hard to undo because you'd have to find the danged thing
Also, NPC stuff is planned, just a matter of time. Races will definitely be given to them, though classes are a tad different. I'm not sure what use NPCs would have of physically being different classes-- you only really care about that to the extent of appearance and dialogue, so you can effectively emulate it in other ways. Though I imagine a "caste" system or such in order to at least design/script them on a categorical basis would be in order.
As for guards that follow you around... eeehhhhh. How much value did NPC companions add in the Diablo series? Maybe a little in II, kinda silly in III. So, maybe not *guards* per say, but maybe something more colorful and interesting. Like MY LITTLE PONY EPIC MOUNTS!!! Or something, ya know. Just tossin' ideas around. Crazy, random companions would have the capacity for far more entertaining dialog and gameplay imo, while not detracting so much from HSM's feel, which is already surreal/zany. Thing is, epic mounts wouldn't be mercenaries so much as things attached to your character that give you special abilities, spouting out dialog from time to time.
Bosses and new mobs are very close on the agenda. Once Spellscript is out and stable, I plan on working on the mobs a great deal and adding some models people have donated (which there are plenty of.) Not to mention implementing some ideas of my own I get excited about implementing, say, a ghost knight mini boss with floaty bits of armor...
New dungeons are also close on the agenda. Clearly they're important to the mod's enjoyability, the question is, are they more or less so than new mobs/bosses? Not to group the two together, necessarily. I'm honestly not sure.
The other problem with dungeons is that there are some serious limitations on generating them, since unlike many other mods, this occurs real-time as the game is running, rather than through some external program that edits existing maps. Hence, as you're running around, even medium-end systems can encounter metric tons of lag if dungeons aren't designed carefully. I'm dreaming lately of a different approach to generating them that'd allow their development to be unhindered by efficiency concerns-- however, I'm not sure yet if it'd work out, and I may only be able to know after having implemented it and experimented with it.
In short, the procedure would be to generate and queue up dungeons on a completely separate thread in some abstract representation, then "copy in" all the data when a new dungeon is requested. It's an iffy design issue.
Example limitation: wouldn't it be nice if towers branched off on either side into a new structure? Or if towers connected in an interesting way to form a castle? Well, too bad! Towers currently have to generate in exactly one chunk, while that chunk is being generated, in order not to lag the crap outta your system.
I'm excited... but does that mean we won't get to try it out soon? Hopefully it's still planned for this weekend?
I'm trying seriously hard to have it out this weekend, but I can't guarantee. Skill trees will be in there, but there won't be new spells yet. Well, except for one or two
The thing that made me delay mostly was the introduction of new spells, but I think it'll be alright to postpone that a little, just so the update is out and it can be tested.
I also want to go ahead and make sure anything that breaks old scripts is introduced now rather than later, so as few people as possible have to edit their old scripts.
Im trying to spawn iron golems using the spellscript block but it won't work, the code I used is as follows:
for int i in 100:
world.spawnLiving("golem", block.getX(), block.getY() + 4, block.getZ())
Also I hope we get capes soon, and i still cant figure out what that book is for in the second inventory screen.
Try "iron golem" and "snowman" I might need to make it refer to "golem" as well for convenience... but then again, it technically ain't the entity's "LivingID".
I should seriously add the capes in at some point XD But not now.
Well don't you worry, the book's gonna be used in this next update, hehehe...
Soooo... I've kinda spent the last 2 or 3 days integrating Java with Spellscript. As in, you can now use any Java class you want in Spellscript. Yyyyyeah, I'm really excited for this.
For example, the following code is valid:
# Prints 'HELLO FRIEND HOW ARE YOU'
var words = ["hello", "friend", "how", "are", "you"]
var sb = StringBuilder()
for var word in words:
sb.append(word).append(" ")
SERVER.log(sb.toString().toUpperCase())
What's awesome is, you'll be able to write your own libraries in Java, Scala, etc, then include them as part of the mod if you want (while still only allowing users to import an exact set of packages you specify). It'll send over the network and everything O.o
That is to say, the library will be transmitted-- as for individual objects, non-serializable ones will be set to null. But that's probably the kinda behavior you'd want anyhow, since not designing your classes for serialization is very likely to result in errors.
Another limitation is that, although you'll be able to implement Java interfaces, you won't be able to extend abstract classes. This may change later if somehow, I discover a way to compile Spellscript to bytecode while still having Spellscript do everything it already does O.o This seems pretty danged unlikely though (part of the reason I decided to make Spellscript to begin with was because of bytecode's unsandboxability.)
But expect extreme awesomeness with this next update, as Spellscript's gettin' a major upgrade.
Love this mod, I would enjoy to see an easier way to play multi-player...
As soon as the Bukkit port is made, you'll be able to simply download the jar and double-click it, then yer set. It might even auto-download in the Technic launcher
Also, updating to 1.5 might help, as your client will come with the open-to-LAN version of the server.
I need a tiny bit of help with creating a spell lol, Basically im trying to make an area spell that does damage to mobs inside of it. Basically i got the area part right, but i cant seem to give it an timed effect to things inside that area. Can someone help me a tiny bit?
Well if you try again and have some code, post it and we'll try to help you out
Hi. I've been playing this mod for quite some time, and I really enjoy it.
I found this mod one day while looking at the Technic launcher after I was getting a bit bored of Red matter hammering a 3x3 grid for 1000 blocks.
I was playing Technic pack, and EE2 was a bit too awesome.
Fell in love with this mod at first glance.
I really am enjoying the changes in the new version.
I look forward to any newer updates.
Also, if anyone knows of good Let's plays of this mod that don't use any texture packs, let me know. There are very few texture packs that don't make me nausious for some reason, and I wish I was making that up.
My brain is wired oddly.
Also, I hope that you do add eventually different types of dungeons than the current towers and the current dungeons.
And to prevent anyone yelling at me, I have read the last 50 pages.
So I do know about the discussion of mazes.
I just am bad at finding my way through mazes. The current dungeons in this mod are confusing enough.
Also, I recently found the Rusty Coathanger.
I feel unclean using it.
Thanks for the flattery And the towers and dungeons can always use improvement, it's true.
If you feel unclean using that, then you definitely wouldn't have liked it before I toned it down a bit... *shivers*. Let's just it had to do with, um, contraception in the 1930's.
So I was surprised to learn this afternoon that Hack/Mine is over 5.3 million downloads! At least 3.7 million of those are distinct. My word, there are a lot of you O.o
As far as updates go, I've been doing a lot to the innards of Spellscript. Pretty soon, Java classes may be presented in Spellscript just as they are in Java! This is already the case with errors and exceptions. This means you'll be able to implement interfaces and such, then pass them back to Minecraft to do powerful things. It also means that fairly soon, you'll be able to do just about anything you can do in Java with Spellscript (but more productively cause the syntax ain't 20 years old!) So rejoice! Rejoice I say!
I've also been thinking a bit about adding aspects of functional programming to the language (such as optional Referential Transparency (RT) as part of function declarations.) Still pondering it, though.
What I mean is a way for players to interact with spellscript.
Hmm. Well, with the next update, admins will be able to create scripts that anyone can run from the console. So in a sense, that's a way users will at least benefit almost directly.
As for actually programming in Spellscript, I'll still have to think about it a lot more. Just a matter of sandboxing it well enough that there would be *zero* security concerns.
I keep end of streaming when I try to connect to the official server, I know I installed the newest alpha(0.6.4.3) Anyone have any suggestions?
Uhhh... the official hasn't been in use for weeks now. I think I mentioned that on the OP... I think. I may not have edited the original linking of the IP, though. Lemme do that right quick...
And as a note, we're currently on 0.6.4.2 (3 will be out soon!) But whatevs, alpha links are confusing.
Yes, all clients are using the latest stable Technic Launcher (build #224).
No, but if that is the case it is a very common global configuration
Four so far:
- 2 macs running different OS X builds (though both are java 1.6 as this is required by technic launcher...1.7 + OS X has some issues apparently).
- 1 ubuntu linux box running a fairly recent java 1.7
- 1 windows box with unknown configuration
Ok yeah I can do that. Might take a weekend or two but I'll let you know what find out.
Alright, awesome. That'll help confirm whether it's MC or HSM causing the issue, if not the exact thing that's causing it. And it may be a global configuration issue, the point though is that it's probably a Minecraft issue given that the issue presented in vanilla 1.2.3 in addition to HSM (at least in the two cases I recall, not sure about yours.) Which means if I want to fix it, I'll either have to fix a Minecraft problem or update to the latest version (assuming they've fixed it, anyhow. Surely they have. But you never know...)
The Macs should be totally fine with regards to Java, as I use a MacBook with java 1.6 to develop HSM on (for similar reasons.) And assuming the linux box is the server, it's probably not the culprit of the client lag. It'd be interesting to know what's on the windows machine(s), but alas, I understand if users are being stingy
Other relevant question: this is just SMP, correct? Doesn't occur in single player? Just making sure, if you know off the top of your head.
Oh, and btw: you wouldn't happen to have any admins fiddling with Spellscript, would you? Because someone could easily create a script causing lag (scripts are executed on both client AND server side.) Just throwing ideas out.
But any additional info you can give is always appreciated, if not to solve your case, then to help diagnose others. Lemme know if/when you've profiled jank
EDIT: I should mention, I haven't experienced or heard reports of such lag with the 0.6 alphas, at which point I was testing servers strenuously (hence the public server.) If it's not system-dependent, then the issue was presented since then, which seems unlikely to me given that I haven't fiddled with anything network-intensive since then. But it could be that we didn't aggravate the server in *just* the right way. But it's hard to say.
That first Entity should be entity (for anyone else trying this).
Hmmm..tilde doesn't seem to open script scheduler on the mac (tried many key variations too...definitely opped..tried creative/not as well just in case).
Not sure what you mean by console (minecraft server console doesn't understand and technic launcher console is read only).
In any case I went into creative and did it via a spell block. Noticed that the editor could use scrolling or word-wrapping
I was able to verify via the minecraft debug screen that entities were zapped, but the severe lag persisted (see attached screenshot). It only seems to happen when I move the view around. Often if I stay still and stare at the floor I can get it down into the green.
Indeed
A bukkit port would be awesome.
No option to *attach* a screenshot (maybe my account's too new) so here's an inline link instead:
Fixed the typo If it's any consolidation to folks, I'm including the kill script as an ss command in the next patch.
Ah, emphasis on client side lag... why can't I learn to read, yeesh. Sorry bout that.
Well, a pertinent question is, do your clients have Optifine installed? (If in Technic, the answer is yes.) This tends to have an impact of its own, which might even be system dependent (though if its on all your clients, that's pretty doubtful.)
I've actually experienced the exact problem you're describing on my old desktop, come to think of it-- thing is, it was in Minecraft 1.2.3, not just HSM. The game would run fine for a few minutes, then hardcore lag until I force-quit the program. I never could fix it, and just assumed it was because my CPU wasn't powerful enough for that version of MC. *However*, since I've gotten a new computer, someone reported having the issue I just described, and fixing it by updating to the correct version of Java 7, with the right number of bits in his case.
Come to think of it... I was running 32-bit windows on a 64-bit machine (don't ask)... maybe being forced to run 32-bit Java on a 64-bit *system* (not OS) was my initial issue? Hmm. Are you absolutely certain your configuration(s) aren't the issue? How many machines are we talking about here? Sorry to reemphasize, it's just historically been the issue.
If you want, you could download VisualVM (or another JVM profiler if you prefer) and sample Hack/Mine on your machine while running the client. A screenie (or copy/paste if it lets you) of the whole method listing would be fantastic. I'd be happy to help you get that set up if you feel like testing
1
I haven't tried it, but you should have access to all the standard Java file stuff. You'll have to tell me if you can get it working
You'll also have to be careful of client/server discrepancies since scripts are executed on both ends. Iow, don't accidentally place files on your clients' computers (or give script editing permission to anyone irresponsible.) This can be checked with SERVER.isClient() and SERVER.isRemote().
Yeah, I figured as much, as I had tested it and couldn't get any issues. Good to know you got it working.
Also, the ripple-y War Stomp that ripples out was originally what I wanted, but I thought Shockwave was both cooler and easier to implement
As for summoning stuff, I was hesitant in the past to add them due to AI problems and added server lag (I AM NECROMANCER WITH 50 SKELLINGTONS), but something like familiars may be possible in the future.
0
First off, JSON is a strict format: unfortunately, coords would likely have to be input as a string or a list of numbers (assuming the user understands every two numbers constitute a coordinate). That seems pretty messy, though. Also, imposing a grid format on the tree is a good idea since it'd be easier for users to reason about, but I'm not sure if its worth limiting the choice of possible locations.
So a possible format might be like this:
"useCoords": true, "skillSet": [ "warrior_leap", 1, 0, "warrior_bash", 3, 0, "warrior_leash", 2, 0, "warrior_charge", 2, 1, "warrior_whirlwind", 4, 0, "warrior_shockwave", 3, 1 ], "skillPrereqs": { "warrior_shockwave":[["warrior_whirlwind", 4,0, 4,1, 3,1], ["warrior_another", 3,0, 3,1]], "warrior_charge": [["warrior_leash", 2,0, 2,1]] }Doing it this way, we'd just be adding coords as additional elements of lists. We could also include the coords inside strings (using the fact that the innards of strings can be formatted however we want), but this seems a touch cleaner.
I'll have to think about it a good bit-- it might be better to let the user specify exact pixels rather than grid slots; then again, I could even give the user options either way. I do want things to be as self-evident and legible in hsmConfig as possible, so it may be better to let users specify them in a completely separate property.
0
You currently can only import Java classes that are in the Minecraft/HSM project already-- this is part of the reason I want to put out another update very soon. Allowing users to use whatever classes they want will involve transferring jars over the network, and some other internal guts which will require some work. (But not too much work.)
EDIT: This means everything in java.lang, Google's guava, and a lot more are available.
Shockwave ain't working? Well this is what I get for rushing :/ That's a very strong reason for a fast update.
Skill numbers next to the icon are a strong possibility, I'll just need a day to experiment with it (and other things) so nothing looks cluttered. Didn't have time yesterday to neaten the skill tree GUI much, unfortunately. Yet another reason for a fast update!
The skill trees are dynamically generated based on hsmConfig, which means square-ish lines won't cut it for all possible skill trees (not without changing the approach to the problem). However, finding any good solution for straight or curved lines for a given tree is a theoretically challenging Computer Science problem, one which would be very nice to avoid
On the bright side, any strangeness from the lines will probably become less noticeable once there are more spells, so, we shall see...
EDIT: If you or anyone for that matter has an idea for an alternate, simpler method of generating the lines, though, I'd definitely love to hear it.
0
If you mean spells having an impact on other spells, this is actually a plan, though I don't think I've said anything about it XD Perhaps it's such an inherent part of skill trees that everyone expects it?
There are some things like this I'm aching to do, but I'm not sure if I should do another update early this week or wait until the weekend to let people settle into the new update. I might have to wait for feedback on that one.
EDIT: There's also some bugs I didn't have time to address (I absolutely forced myself to get this update out last night, which wound up being at 5 AM), so their existence will likely be a deciding factor.
0
SHAZAM: Hack/Mine v0.6.4.3
To install, you'll additionally want to place "sslib" alongside "spells" (remember to delete your old "spells", or start migrating it all if you've customized anything!)
I'll update the OP tomorrow and hopefully get a video of new stuff up as well
0
This update's taking a while, but I still plan on getting it out later tonight. I'm doing a lot more balance work than I expected (Spellscript has distracted me from some glaring issues-- well that, and I added the ability to invest levels into each of your skills, which has taken some work). I probably won't be updating the front page until tomorrow when I wake up, and the video's definitely gonna have to til tomorrow at the earliest.
Hang onto your pants...
0
I'll make one tomorrow, hopefully along with a video
If you're using some framework like Forge or MCP, it's a whole 'nother ballpark than what I've been doing for Hack/Mine. You'd have to do some fairly complicated things to get all that working, since it's relatively invasive, which means compatibility is difficult to impossible. You'll have to get creative if this is the case.
EDIT: Iow, you'll likely need to get familiar with bytecode injection (specifically with ASM) and Java's reflection utilities in order design things in a sane manner. Well, if you can call that sane XD
However, if you're not using those tools, then the changes are pretty easy to make with straight-up class edits. This is assuming you're familiar with basic software engineering concepts
EDIT: An idea: do absolutely everything through the "Sash" item. This will minimize any invasiveness dramatically. So for example, whenever you need to know the player's class, scan his armor for the sash. Mana pool? Make it a property on the sash, and persist it with metadata. Of course, you'll want a bunch of static helper methods to reduce the boilerplate (including down-casting the sash itself), but this would probably be easier to maintain than bytecode injecting methods onto EntityPlayer (and probably friendlier to other mods as well.)
0
Sorry, perhaps I didn't explain so clearly.
The problem is imposing higher-level structure. Generating a dungeon "as you go along it" is easy, regardless of size. The problem is that doing so results in a dungeon with almost no organization; it just sprawls without a sense of direction. Hence, in order to introduce structure, you'd have to pre-generate the entire dungeon and cache the result, then lay down the pre-generated chunks of it as world generation occurs (or at least that's one possible solution).
This is not a trivial thing, given that dungeon data would need to be persisted at a level higher than chunks, then referenced in a very efficient manner-- this is part of the reason the Area System was introduced.
0
Multi-chunk generation is possible even in vanilla Minecraft (take mineshafts for example), but they have an even worse problem-- you can't generate structures at a "high level". In other words, you couldn't give the structure a kind of "flow" as you wander through it, or have keys that unlock doors in other areas in such, etc. Generation in one chunk can't retroactively affect the structure in another, since nothing is guaranteed-- every chunk has to be independent in a sense. Thus, you wind up with massive, "wandery" experiences, and an inability to make anything interesting of it. It'd permanently cripple that dungeon's enjoyability.
0
Spellscript has fully functioning access to generic types at this point (including generic type inference!), as well as import resolution, aliasing, and finally, access to static methods. *However*, generic/static methods have yet to be implementable in the language. So you can access such features introduced from Java, but I still need to add ways to create static and generic methods of your own. It's mostly a matter of deciding on the syntax in a way that's as easy to use as possible, so wish me luck. Static methods will be easy, generics might have to wait until later though...
I've also got a nifty maze-generation script working from the console
Also, there's some delay because I'm working on job applications atm. The real world is calling my name, and I must answer! Glad to have Hack/Mine on the resume, though. I'm basically shooting for the stars in the game industry at this point and hoping I get in
This weekend's feeling pretty good for an update. I just gotta make sure everything is seamless, and that I've addressed all the bugs posted since the last update. However, my applications come first, since finding a job is kinda crazy important. Apart from that, though, Spellscript is 97% ready for this update. Woohoo!
Those are actually some solid ideas, though a few things: firstly, how is a non-mob block better than the existing method of disabling mobs through the command line? (see the wiki for how) While a block is easier to place, a block is also much hard to undo because you'd have to find the danged thing
Also, NPC stuff is planned, just a matter of time. Races will definitely be given to them, though classes are a tad different. I'm not sure what use NPCs would have of physically being different classes-- you only really care about that to the extent of appearance and dialogue, so you can effectively emulate it in other ways. Though I imagine a "caste" system or such in order to at least design/script them on a categorical basis would be in order.
As for guards that follow you around... eeehhhhh. How much value did NPC companions add in the Diablo series? Maybe a little in II, kinda silly in III. So, maybe not *guards* per say, but maybe something more colorful and interesting. Like MY LITTLE PONY EPIC MOUNTS!!! Or something, ya know. Just tossin' ideas around. Crazy, random companions would have the capacity for far more entertaining dialog and gameplay imo, while not detracting so much from HSM's feel, which is already surreal/zany. Thing is, epic mounts wouldn't be mercenaries so much as things attached to your character that give you special abilities, spouting out dialog from time to time.
Bosses and new mobs are very close on the agenda. Once Spellscript is out and stable, I plan on working on the mobs a great deal and adding some models people have donated (which there are plenty of.) Not to mention implementing some ideas of my own
New dungeons are also close on the agenda. Clearly they're important to the mod's enjoyability, the question is, are they more or less so than new mobs/bosses? Not to group the two together, necessarily. I'm honestly not sure.
The other problem with dungeons is that there are some serious limitations on generating them, since unlike many other mods, this occurs real-time as the game is running, rather than through some external program that edits existing maps. Hence, as you're running around, even medium-end systems can encounter metric tons of lag if dungeons aren't designed carefully. I'm dreaming lately of a different approach to generating them that'd allow their development to be unhindered by efficiency concerns-- however, I'm not sure yet if it'd work out, and I may only be able to know after having implemented it and experimented with it.
In short, the procedure would be to generate and queue up dungeons on a completely separate thread in some abstract representation, then "copy in" all the data when a new dungeon is requested. It's an iffy design issue.
Example limitation: wouldn't it be nice if towers branched off on either side into a new structure? Or if towers connected in an interesting way to form a castle? Well, too bad! Towers currently have to generate in exactly one chunk, while that chunk is being generated, in order not to lag the crap outta your system.
0
I'm trying seriously hard to have it out this weekend, but I can't guarantee. Skill trees will be in there, but there won't be new spells yet. Well, except for one or two
The thing that made me delay mostly was the introduction of new spells, but I think it'll be alright to postpone that a little, just so the update is out and it can be tested.
I also want to go ahead and make sure anything that breaks old scripts is introduced now rather than later, so as few people as possible have to edit their old scripts.
Try "iron golem" and "snowman"
I should seriously add the capes in at some point XD But not now.
Well don't you worry, the book's gonna be used in this next update, hehehe...
0
For example, the following code is valid:
# Prints 'HELLO FRIEND HOW ARE YOU' var words = ["hello", "friend", "how", "are", "you"] var sb = StringBuilder() for var word in words: sb.append(word).append(" ") SERVER.log(sb.toString().toUpperCase())What's awesome is, you'll be able to write your own libraries in Java, Scala, etc, then include them as part of the mod if you want (while still only allowing users to import an exact set of packages you specify). It'll send over the network and everything O.o
That is to say, the library will be transmitted-- as for individual objects, non-serializable ones will be set to null. But that's probably the kinda behavior you'd want anyhow, since not designing your classes for serialization is very likely to result in errors.
Another limitation is that, although you'll be able to implement Java interfaces, you won't be able to extend abstract classes. This may change later if somehow, I discover a way to compile Spellscript to bytecode while still having Spellscript do everything it already does O.o This seems pretty danged unlikely though (part of the reason I decided to make Spellscript to begin with was because of bytecode's unsandboxability.)
But expect extreme awesomeness with this next update, as Spellscript's gettin' a major upgrade.
As soon as the Bukkit port is made, you'll be able to simply download the jar and double-click it, then yer set. It might even auto-download in the Technic launcher
Also, updating to 1.5 might help, as your client will come with the open-to-LAN version of the server.
Well if you try again and have some code, post it and we'll try to help you out
0
Thanks for the flattery
If you feel unclean using that, then you definitely wouldn't have liked it before I toned it down a bit... *shivers*. Let's just it had to do with, um, contraception in the 1930's.
0
So I was surprised to learn this afternoon that Hack/Mine is over 5.3 million downloads! At least 3.7 million of those are distinct. My word, there are a lot of you O.o
As far as updates go, I've been doing a lot to the innards of Spellscript. Pretty soon, Java classes may be presented in Spellscript just as they are in Java! This is already the case with errors and exceptions. This means you'll be able to implement interfaces and such, then pass them back to Minecraft to do powerful things. It also means that fairly soon, you'll be able to do just about anything you can do in Java with Spellscript (but more productively cause the syntax ain't 20 years old!) So rejoice! Rejoice I say!
I've also been thinking a bit about adding aspects of functional programming to the language (such as optional Referential Transparency (RT) as part of function declarations.) Still pondering it, though.
Hmm. Well, with the next update, admins will be able to create scripts that anyone can run from the console. So in a sense, that's a way users will at least benefit almost directly.
As for actually programming in Spellscript, I'll still have to think about it a lot more. Just a matter of sandboxing it well enough that there would be *zero* security concerns.
Uhhh... the official hasn't been in use for weeks now. I think I mentioned that on the OP... I think. I may not have edited the original linking of the IP, though. Lemme do that right quick...
And as a note, we're currently on 0.6.4.2 (3 will be out soon!) But whatevs, alpha links are confusing.
Alright, awesome. That'll help confirm whether it's MC or HSM causing the issue, if not the exact thing that's causing it. And it may be a global configuration issue, the point though is that it's probably a Minecraft issue given that the issue presented in vanilla 1.2.3 in addition to HSM (at least in the two cases I recall, not sure about yours.) Which means if I want to fix it, I'll either have to fix a Minecraft problem or update to the latest version (assuming they've fixed it, anyhow. Surely they have. But you never know...)
The Macs should be totally fine with regards to Java, as I use a MacBook with java 1.6 to develop HSM on (for similar reasons.) And assuming the linux box is the server, it's probably not the culprit of the client lag. It'd be interesting to know what's on the windows machine(s), but alas, I understand if users are being stingy
Other relevant question: this is just SMP, correct? Doesn't occur in single player? Just making sure, if you know off the top of your head.
Oh, and btw: you wouldn't happen to have any admins fiddling with Spellscript, would you? Because someone could easily create a script causing lag (scripts are executed on both client AND server side.) Just throwing ideas out.
But any additional info you can give is always appreciated, if not to solve your case, then to help diagnose others. Lemme know if/when you've profiled jank
EDIT: I should mention, I haven't experienced or heard reports of such lag with the 0.6 alphas, at which point I was testing servers strenuously (hence the public server.) If it's not system-dependent, then the issue was presented since then, which seems unlikely to me given that I haven't fiddled with anything network-intensive since then. But it could be that we didn't aggravate the server in *just* the right way. But it's hard to say.
0
Fixed the typo
Ah, emphasis on client side lag... why can't I learn to read, yeesh. Sorry bout that.
Well, a pertinent question is, do your clients have Optifine installed? (If in Technic, the answer is yes.) This tends to have an impact of its own, which might even be system dependent (though if its on all your clients, that's pretty doubtful.)
I've actually experienced the exact problem you're describing on my old desktop, come to think of it-- thing is, it was in Minecraft 1.2.3, not just HSM. The game would run fine for a few minutes, then hardcore lag until I force-quit the program. I never could fix it, and just assumed it was because my CPU wasn't powerful enough for that version of MC. *However*, since I've gotten a new computer, someone reported having the issue I just described, and fixing it by updating to the correct version of Java 7, with the right number of bits in his case.
Come to think of it... I was running 32-bit windows on a 64-bit machine (don't ask)... maybe being forced to run 32-bit Java on a 64-bit *system* (not OS) was my initial issue? Hmm. Are you absolutely certain your configuration(s) aren't the issue? How many machines are we talking about here? Sorry to reemphasize, it's just historically been the issue.
If you want, you could download VisualVM (or another JVM profiler if you prefer) and sample Hack/Mine on your machine while running the client. A screenie (or copy/paste if it lets you) of the whole method listing would be fantastic. I'd be happy to help you get that set up if you feel like testing