Flower's Dev Diary (Week of October 23rd)

[FlowerChild]

All right, decided to go straight to 1.4 and after about 6 hours work am at the half-way mark in integrating my base-class edits (that's the tough part).

In the end I figured with the MCP name changes and such that were bound to occur between versions, I just may as well kill two birds with one stone and both refactor for Fernflower and port to 1.4 at the same time.

Personally, I'm really not digging Fernflower. The code it generates seems far sloppier than what I was getting with JAD. For some reason it prefixes every call to all member functions with "this." which is really annoying (and useless), and also tends to inline massive conditionals as part of return statements in ways that makes them rather annoying to figure out.

So far I've yet to see any advantages to it, but I guess I'll just take it on faith that there are some...somewhere ;)

Anyways, the update is moving along nicely. I suspect I'll probably be done with at least this first pass tomorrow, and then can start considering what's going on with ML.

[PatrickSJ]

Personally, I'm really not digging Fernflower. The code it generates seems far sloppier than what I was getting with JAD. For some reason it prefixes every call to all member functions with "this." which is really annoying (and useless), and also tends to inline massive conditionals as part of return statements in ways that makes them rather annoying to figure out.

For decompiling Minecraft + mods in one go Fernflower supposedly produces a lot less errors. JAD doesn't always play nice with nested classes and enum, and I think I recall it having minor issues with generics.

However, I've never hit these errors when decompiling just Minecraft with JAD.

[FlowerChild]

A little code discovery while porting:

Skeletons drop coal now? Wtf?

No...that simply will not do :)

EDIT: Wait, those must be the nether skeletons...hmmmm.

[Graphite]

Maybe it's just preparation for christmas, considering they've been bad, bad boys :P

[walker_boh_65]

EDIT: Wait, those must be the nether skeletons...hmmmm.

Yes don't worry, it is JUST the wither skeletons. Still odd, but not completely game breaking.

[FlowerChild]

Yes don't worry, it is JUST the wither skeletons. Still odd, but not completely game breaking.

Yup, that's what it appears to be. I just brain farted due to both the nether and regular skeletons now using the same class, and the function to test which is which not being clearly named (which is generally the case with all new functions that come with a new release).

It did remind me I have to take care of the potato nonsense on zombies though ;)

[Battlecat]

It did remind me I have to take care of the potato nonsense on zombies though ;)

Good call. I'd rather have to explore to find them instead of suddenly finding them in my mob trap.

[FlowerChild]

Good call. I'd rather have to explore to find them instead of suddenly finding them in my mob trap.

Yeah, for sure. It's also just another "get everything no matter what you do" design change I'd rather not have in the mod.

Having unique methods of acquiring specific resources is cool. It encourages diversity in player activities and potentially exploring systems a player might never experience otherwise.

I want people to work for their potatoes damn it! :)

[walker_boh_65]

I want people to work for their potatoes damn it! :)

From what I have seen potatoes (and just because of the way things works prior to this, carrots as well) can already be found in village farms instead of wheat.

Just as an FYI in case you didn't know.

[FlowerChild]

From what I have seen potatoes (and just because of the way things works prior to this, carrots as well) can already be found in village farms instead of wheat.

Yup, I'm aware, which is a nice little bonus for finding a village, which is really what I'm talking about.

[Six]

From what I have seen potatoes (and just because of the way things works prior to this, carrots as well) can already be found in village farms instead of wheat.

Yup, I'm aware, which is a nice little bonus for finding a village, which is really what I'm talking about.

The rather drastic impact this will have on BTW though, is no pig breeding until you find a village. I look forward to some awesome overworld exploration with a stack of the larger maps :)

[FlowerChild]

The rather drastic impact this will have on BTW though, is no pig breeding until you find a village. I look forward to some awesome overworld exploration with a stack of the larger maps :)

You're forgetting kibble.

[Six]

Oh wow, I never knew kibble activated breeding and was an all animal food. I'd always assumed it was just an alternate wolf food for making poop.

[Eriottosan]

Oh wow, I never knew kibble activated breeding and was an all animal food. I'd always assumed it was just an alternate wolf food for making poop.

Pigs only, IIRC, but yeah (:

[Foxy Boxes]

That's kinda cool. It makes villages like slimes were, a rare treat that enables something early on that if you fail to find you have to wait for a late-game equivalent.

[pgelinas]

Personally, I'm really not digging Fernflower. The code it generates seems far sloppier than what I was getting with JAD. For some reason it prefixes every call to all member functions with "this." which is really annoying (and useless), and also tends to inline massive conditionals as part of return statements in ways that makes them rather annoying to figure out.

Ok, I'm glad to know that I'm not the only one that had problems with Fernflower. The code I ended up with when decompiling Minecraft was rather hideous, but at that time I thought it was because of the compiler optimizations (and, well, it's decompiled code after all!). Although I don't recall seeing "this." that often, maybe I'm just too used to the keyword I don't see it anymore. And those conditionals are a pain in the butt, you end up with 10 nested conditionnals and you have no freaking idea where you're at...

Also, I don't know if it's Fernflower-related or not, but when decompiling two same classes with a minor difference in the source code I ended up with two very different source files due to 1- variable scoping and 2- variable names (like, using var21 instead or var15 and shit like that).

As for the advantages, one of them is that it's easier to use under Linux: with JAD you needed wine because the version they used was not available under Linux. There's none of this with Fernflower.

[FlowerChild]

Woot! Two pieces of good news:

First, someone at Mojang must have bought a book on OOP, as I just saw a bunch of instanceof checks converted into a call to a single function in a base class!

Joy of joys! :)

This might not seem like a big deal, but I think this is the first time ever that a change on Mojang's part has allowed me to eliminate a base-class edit, rather than add one ;)

I'll keep my fingers crossed this trend will continue into the future.

For modders, you can check out this 8th wonder of the world in SpawnerAnimals.creatureSpecificInit().

Secondly, I'm down to only 10 base-classes left to convert over to Fernflower & 1.4 out of the 110 I had to do!

Woot!

<dances>

[CumuloNimbus10]

It sounds like the 1.4 update process is going along quite nicely. I'm glad you aren't having the same problems that came up in the 1.3 update. Let's hope this trend continues. :)

[walker_boh_65]

-snip-

Sweet. That that look like a release of. . .sometime before the end of next year. ;)

No really. I am glad this update sounds a LOT smoother than the port to 1.3 was.

[odranoel]

amazing progress mate. but for you, that seems to be the norm lately. so im sure from your point of view this is more or less a walk in the park :p (albeit a boring and un-entertaining walk)

one thing i just realized, im sure we will start seeing creative use of all those new sounds mojang put in :p you already did an excellent job with the limited ones you had up until now. getting an expanded pallet is always awesome for a creator :)

[FlowerChild]

Excellent! Many of you are probably already aware, but ML just updated, and I only have 5 base-classes left to go :)

So, false alarm on my earlier plans to release without ML. The timing is perfect.

I still have quite a bit more work to do mind you, as I'll still need to both integrate with the server code, and spend some time axing some of the new vanilla features, but this is definitely good news :)

[Battlecat]

That's excellent news FC! I'll bet that's a significant concern cleared off the board!

[FlowerChild]

Indeed it is. I'll probably try to wipe out these remaining 4 files tonight (world.class is always a big one, which is what I just finished), then set into integrating with ML tomorrow.

I think I've been at it around 11 or so hours today. Probably time to start wrapping it up :)

[TheYaMeZ]

Great news! Glad to hear everything has worked out with modloader and the porting.
As always, thanks for the time you invest into making minecraft more amazing week by week.

[FlowerChild]

All right, done with the base-class port. Time to crash and watch me some Walking Dead :)