HIVE

Ok makes more sense now.

I think it shall work if it works on patch 1.26 (since it is the only version of War3 I currently manage to make work on Linux).
It would be so amazing, it would be possible to have dynamic allocation! Then you have to do a compiler to some other languages and you can write JASS bytecode instead!

So you'd really be able to do amazing stuff if you bypass the JASS step.

Yeah, well, that AddressDetection.w3m doesn't work for me on the latest patch (compiler complains about type casting).

You should strive to make 1.29 working. It introduced a host of new natives (not all of those are new, it's just a diff): https://www.hiveworkshop.com/pastebin/53ff76b1d7058f617f758256730ca25119122/

Oh yeah no more need for Damage Detection and such stuff woha

But actually it is impossible to make it work so stuck with 1.26 I guess haha

Yeah, setting tooltips and icons is so dope. And you can set/get cooldown! No more deleting+adding an ability to make cooldown reset abilities.

Ah, that's unfortunate. I guess you can have something like static if GAME_VER_LT_129 then define stub/supplementary functions for the natives so you can still write code that uses them, in case you submit something to Hive for example, or if some newer patch starts working for Linux.

Yeah some of these new natives look dope!
Like some dream becoming true

There's a variable for the game version?
That sounds lit.
However since I won't be able to test it that sounds kind of underwhelming haha

The bytecode stuff is working on my warcraft 3
Guess I can create function of doom to allocate dynamically memory in addition to normal vJASS scripting.

Sounds very interesting/promising.

Guess I'll have to reorganize my workflow to include all of these files and start to sandbox this shit.
I hope it will be able to do it on latest warcraft III too, so we can generate JASS bytecodes with languages such as Wurst or vJASS and get some amazing stuff going on.

Nah, but you can make one: constant boolean PATCH_LT_129 = JASS_MAX_ARRAY_SIZE <= 8192.
Are you sure you won't ever be able to get a newer patch working on Linux? Perhaps the Classic team will try to make Reforged conform with Wine or something.

ARRAY SIZE IS BIGGER THAN 8192 NOW?
I feel like I'm still stuck in prehistory x)
I don't know if there is a JASS_MAX_ARRAY_SIZE variable in my version either.

Anyway a global isn't too hard to configure in those cases.

Right now it is impossible since the unified launcher or it seems and I did not find anything helping me.
So for now I'm stuck in 1.26.

By the way are you sure the bytecode stuff is not working on the latest patches? Sounds sad because I won't be able to publish stuff using it if it does not work :/

Worst case I can tinker with it for fun but it's quite limited.

Yeah, like 65536 or something.

There are direct launchers at C:\Program Files\Warcraft III\x86[_64], did you try those?

I guess it'd work if this is a thing: https://www.hiveworkshop.com/threads/jass-hot-code-reload.313811/, which I haven't tried because I'm too lazy, lol.
Seems like the author is furtive on how it works, but assures it'll work in future.

Oh great!

Yup and it did not work don't worry I've tried basically everything ^^

Oh yeah this thing looks incredible too. I guess the bytecode stuff might still work otherwise it does not make any sense.
However I wonder what is used to make it work.
Modding turned out way different during my hiatus

Have you tried installing Wine from their site directly, and from experimental branch? https://wiki.winehq.org/Download Every update has a change of making it work I bet.

Here's what it inserts with 'init' command: https://www.hiveworkshop.com/attachments/jhcr_war3map-j.329720/
Judging by a cursory glance, seems like he has written a bytecode interpreter. So doesn't seem like making the JASS VM run bytecode. :/

Trust me I've got my fair share of wine tries ^^
Right now I just can't make it work >_<

Ah seems interesting too but it is not what I really expect.
Anyway I've read a ton of stuff about JASS bytecode today (instead of working I've to admit) and I'm really hyped about this.
Tonight I'll just try to sleep more than last night (the last message I sent you was after getting out of bed..) so maybe for the coming days/weekend

I'm very skeptical of using bytecode. Even if there was a way to run arbitrary bytecode on the latest patch, I'd imagine Blizzard rushing to fix it upon notice. Sorry for the pessimism, I'm just afraid it'd be vain effort to write in bytecode. :/

Naaaaaaah I'm a mad lad you know me.

Once a mad lad, forever a mad lad!

To be honest, if it works in 1.26 I'll keep doing crazy shit in 1.26 haha

Well, I'll be looking forward to what you can stir up in the fathoms of your genius. Keep me posted!

Ahaha such kind words for a mad lad I guess.

Anyway I have some ideas of projects and within these ideas there is the bytecode to analyze and see if I can produce some stuff. So it depends what I'll do first, and knowing myself I'll probably do a lot of stuff at the same time.

So we'll see how it goes!

I'm doing this Python school course.. Damn, list comprehension, slices, f-strings... Why can't C++ be cool like that >.< Next up to amaze me is Haskell.

I'll do one last stab at persuading you to try Wurst; it has tuples, varargs, lambda closures, and just look at the switch case combining and looping possiblities: https://wurstlang.org/manual.html#statements
It's whitespace sensitive like Python, so you must like it. :)

By the way, what do you think about assembling a list of vJASS idioms? Extends array, unit indexing, module initializer, FirstOfGroup loop, etc. That could make a great addition to tutorials section.

Actually it's good for Python because it aims to be really close to pseudocode. But for C++ I think there is already too much going on. But you could implement all these operators to mimic the Python syntax sugar ^^
I really hate Haskell and all this "functional programming" trend tbh.

My complain with Wurst is not the language, it looks particularly cool (except this indentation based which is something I actually really don't like about Python). On the other hand, I really don't like the fact that it is really not clear how it will translate to JASS later and how much it will hurt the performances.

For instance there are a lot of vJASS features I'll never use because of the drop in performances because of how it is translated later on.


What do you mean by "assembling a list of vJASS idioms"?

Well, we're getting std::format(), essentially a rip-off of Python's format; std::format("{}", 42) is the same as "{}".format(42), and you know what? I don't mind that at all! ^^ Maybe I like extensive standard library.

Hmm.. what if there was a compiler explorer for Wurst, akin to godbolt.org?

Like this: https://en.wikibooks.org/wiki/More_C++_Idioms
Example: https://www.hiveworkshop.com/pastebin/012a9d919a98fd0210b80f3da56baff319140/