Finally my humble efforts to port Lemmings to Java have reached a state in which I think I can release it for a (more or less) public alpha/beta test.
I came up with the name "Lemmini" which is the scientific "tribe" name for lemmings since it not in direct conflict with the trademark "Lemmings". Sounds a little silly in my mother tongue, but maybe not as much in yours ;)

Please note that you need a (legal) copy of "Lemmings for Windows" to actually play the game. Lemmini is just a game engine which needs the resources of "Lemmings for Windows to run" ... mainly for legal reasons.
I'd recommend to keep the CD-ROM in reach or to copy the WINLEMM folder from the CD-ROM to your HDD since future updates of Lemmini might need to extract the resources again to reflect changes in the resources.

Secondly you'll need a Java runtime environment (JRE). I'd recommend to download the latest version "JRE1.5 Update 6" from here:

http://javashoplm.sun.com/ECom/docs/Welcome.jsp?StoreId=22&PartDetailId=jre-1.5.0_06-oth-JPR&SiteId=JSC&TransactionId=noreg

After installing Java and having your copy of "Lemmings for Windows" ready to hand, you are prepared to visit my Lemmini homepage:

http://home.arcor.de/0xdeadbeef/lemmini.htm

As described there, you might either run/install Lemmini as a Webstart application by clicking to the according link on the site (and answering all the following questions with "yes") or you can download the file "lemmini.jar" directly, copy it somewhere to your local HDD and start it with a double click. The advantage of Java Webstart is that this will automatically check for updates and download the latest version of Lemmini. Then again if you don't like that or don't have a permanent internet connection, you might want to stay with the Jar.

I'd call Lemmini a playable Alpha at this point, because a few game features are missing. Most noticeably you can play all levels from the beginning, the level codes are given, but you can't enter them anywhere and there's no kind of "game over" screen whatsoever.
Then again, all Levels of the Windows version (+5 additional levels) are in, the levels are completely playable, also music, sound and user interface are finished or at least nearly finished.
I tested quiet a lot during the last days on five different systems and had no crashes lately, so I hope it's more or less stable. Then again, I could only test on WinXP systems. So it might crash, refuse to work or behave strangely on different operating systems. Last time I tried in Linux, it crashed while loading resources. But that was some months ago and since then I reworked all the resource names to avoid conflicts with uppercase/lowercase file names. So it might work under Linux, but it's kinda unlikely.

About resources: the game needs almost no CPU-time during normal gameplay on the five systems I tested, where the lowest one was an 1.6MHz Centrino Notebook. It eats much more resources in "fast forward" mode, but this is a different story (limitations in the current Java runtime under Windows). The RAM consumption is about 60MB, so I guess if you have 100MB free, you shouldn't have problems.

The intention of this public alpha/beta test is to get feedback on the following topics:

- Does the game work (well) on your system (especially Windows versions before XP, Linux and Mac)?
- Is the timing ok (release rate, movement of lemmings, animation speed of lemmings, level objects like traps etc.)?
- Is the behaviour of lemmings ok in detail (height of obstacles, falling height, digging speed etc.)?
- Do the Traps and other level objects work as they should (also entries, exits)
- Are the level names, order, start position ok? Are all levels fully playable and "winnable" with the default solution?
- Anything else strange/wrong compared to the Amiga/DOS/Windows version?
- Are there glitches, bugs, crashes, freezes etc. ?

If this is ok with the Moderators, I'd suggest to post comments in this thread. Please don't spam me with emails or PMs for every little comment or issue. Also I'd ask you to keep this beta test internal to this forum for the moment.
And PLEASE don't ask for "Lemmings for Windows"! It's still on sale and I ask you to respect the rights of the copyright holders. I bought my copy at eBay for just 5€ as I started this project, so I think this should not be an issue.

Anyway, I can neither promise that Lemmini will work (well) on your system nor that I will consider/comment any issue discovered during the test. Please be aware that this is a spare time project any my interests may and will change from time to time.

Hope you enjoy it...

That's entirely reasonable, but on the other hand I hope you agree that it's entirely reasonable of me not to want to pay for another version of Lemmings when I already have Lemmings for Mac and Cheapo. Pity; I'm very interested in your project!

A couple of things I've noticed:

• CPU usage is higher at the level intro screen than when playing the level. :o
• The wrong password is displayed -- the password for the NEXT level should be displayed, not the password for the completed level.
• Some sounds are chopped off (e.g. "Oing" and the tool selection sound.)
• Floaters seem to fall too fast, but this may just be in my head.
• The digging animation looks like it loops before it completes.

All in all, it's very well done. (Especially the mouse handling.)  ;D

Now for feature requests: :P

• Full-screen mode, if possible.
• Explosions similar to Lemmings or Cheapo

Hm, on my machine it's between 0 and 2% in any intro/briefing/debriefing/game screen. Only fast forward needs more.


Oops, you're right. I'll change that in the next release...


Did you try to select another Mixer in the Sound Menu? Also make sure that your using JRE1.5 Update5 or above. In earlier versions of the JRE, short sounds were not played correctly. The sound code itself is correct and works flawlessly on several machines.


They should be accurate on pixel/frame basis, but if anyone can confirm, I'll check again...


Hm, looks ok for me. Can anybody confirm?


The mouse handling is a bit jerky since the mouse cursor is software drawn to be "in sync" with the game. And indeed I already doubled the frame rate.
BTW:  I thought you'd praise the music ;)
And you should really read the scrolltext :)

I restarted the game and the problem disappeared.


Doesn't seem jerky to me. :/

And the music is good too. :D

OK, thanks to some files that ccexplore supplied I'm able to test this out!

I've only tried one level (1 Tricky: This should be a doddle) so far, and noticed the following things:

* On the intro screen the level number is "0" for some reason.
* I'd much rather have a music system like the Mac version -- music fixed for each level -- so that I don't always have to hear the Cancan! It's OK, but there are much nicer musics down the line.
* A big one. When two lemmings are under the cursor, the one already doing a skill has priority, as it should. BUT if the skill you're trying to assign is inapplicable, that priority doesn't get overridden! E.g. the first lemming builds over the gap, the second lemming comes up behind and needs to be given a builder also. You can't do it because the first lemming has priority but he can't be given another builder skill.
* Changing the release rate caused the game to crash.

Ok, this and the wrong levelcode will be fixed in the next release.


The music is fixed for each level. And the first level "Just dig!" has CANCAN as fixed music like the Amiga had. So I don't really get the problem.


Indeed there's no (intentional) priority handling at all. Indeed, just by chance, the "youngest" (which entered the level later) lemming has priority at the moment. I must admit I'm not sure if there was any designed priority handling in other versions. At least I did not had the impression when playing the windows version and can't really remember for the Amiga version.
It might be a little confusing however, since the "type" of the "lemming under the cursor" is displayed and if there are 4 different types, which one's name should be printed. At this moment, the name is that of the one who will get the skill
However I could implement something like that if there's demand for it. I put it on my list.


Hm, this doesn't sound too good. Any more more input on this issue? Like: Does the game just exit or is a error message displayed? If so, could you post it? Does this happen always or sometimes, at first press or after N presses etc.?
Does it also happen if you use the keyboard shortcuts ("+", "-")?

I played it on my father's Windows 98 and compared it with WinLemm. There were a few problems:

*Walking speed too fast
*The terrain height that causes the lemmings to turn around is too high(the lemmings were walking something like halfway up the right side of the top chamber in Just Dig)
*No music whatsoever(most likely computer-specific, but still...)

How much too fast? On my system, Lemmings for Windows run too slow. Indeed on minute game time takes much longer than a minute. Please check how long a minute game time takes for Lemmings and Lemmini on your system.


That's a slope, not an edge.


Did you turn it on? I think it's disabled by default.

Yep, I'm playing it on Win98 SE and, well...

* The Lemmings, objects and toolbar are constantly flashing (that's a little bit annoying) and I think it causes the game to lag
* Fast Forward Mode doesn't work
* The game hangs after a while and requires to be closed using Cttl + Alt + Delete. And after, then, the music repeats a tiny part of the song constantly and it's pretty annoying. I need to restart my computer to avoid that! :( The only thing I did is to increase the release rate to 99 and when almost all Lemmings enter the exit: Crash! without any reasons.

Music works and I don't think Lemmings are walking too fast. I didn't tested the terrain height problem, but it seems to be ok too.

That's bizzare since all the drawing is done offscreen and then always the complete screen is copied to the front buffer at once.
I will have to think about how this could happen at all...
What kind of machine are we talking about??? What kind of videocard do you have? Do you know if it supports DirectX?


What does that mean? Is the speed of lemmings and timer not influenced by pressing the button (or "f") or is the icon not even "pressed"? What about CPU load when playing and when using FF mode?


If it hangs: does the menu still react? When you say it crashes, did it exit or freeze? No error message?


Something at least...

Well, actually the game it quite playable despite of this and none of my private test players noticed that while they noticed changes in framerate at once. But of course they are no hardcore lemmings freaks.
I will put it in if there's demand for it and there is time for it (there seem to be more urgent issues like crashes and redraw problems on Win98). However I'm a little perplexed by the phrasing "not negotiable". I mean, at the end I'll have to decide what I can or want to do and what I can't/don't.

Right, and specifically on the problem I pointed out, if there are two lemmings under the cursor and one cannot take the skill (e.g. "builder" is selected and one of the two is already a builder) then the priority MUST pass over to the other lemming. This is absolutely vital to the gameplay. (Yes, the game is still playable, but there are situations where you can't do something you want to and should be able to.)

About the music problem, if you have fixed the music for each level, then level 17 Fun should have music #17, level 18 should cycle back to music #1 (Cancan), and so the level I was playing, 1 Tricky, should have music #14, no?

Calm down guys, I said it's on the list. However if someone gives me the rules for priority handling, it's more likely it'll make it into  a release soon.


You can have a look in the resource directoy in  "levels/orig/orig.ini" (and "levels/ohno/ohno.ini"). There the available music is defined as well as the definition which music to play for which level.
At the moment, level Fun 17 has the following definition line:
# Fun 17 - Easy when you know how (based on "Compression Method 1"/lvl0027.ini)
fun_16 = lvl0027b.ini,CAKHLFLBDU,16

The music is "16" which is defined as:
music_16 = mountain.mod

If there's something wrong with the music order, playe have a look in the ini file and tell me what do change.
Indeed you can change the ini file(s) yourself to test for yourself.

And one general thing for those who observe crashes: please download "lemmini.jar" from my site, put it somewhere and start it from the command line (console) with

   java -jar lemmini.jar

If the game crashes, please have a look at the console if there is any error text printed out by the virtual machine.

I had a first look a it, and I quite like it. The name Lemmini doesn't sound that bad IMO (I'm also German)
However there are some problems, or differences to what I'm used to I noticed:

- object animations continue when the game is paused or when fast forwarded, in same speed; that causes for traps, with fast forward more lemmings can pass one trap, it might cause a potential backroute for 'Compression method 1' (Taxing 6) and perhaps also for other levels; pressing pause again can cause that some more lemmings than intended get killed by a trap
- already the first lemmig coming out is affected by the RR, actually the first lemming should come out not depending on the RR
- the climbers seem far too fast to me (twice the speed they actually should have?)
- a climber hitting his head does the 'vaulter' animation (transitioning to walker) at the beginning if his fall; a climber who should actually do this animation (when having finished climbing an object) doesn't do it however
- changing the RR influences the outcoming of the lemmings instantly, at least in the DOS version the RR changes effectively only after outcoming of the next lemming (I actually don't have any problem with it, also, in Cheapo it works the same way, but anyway)
- the 'jumper', a lemming climbing a step of a height 3 - 6 px doesn't have his own animation which causes a skip of a few frames (this problem also appears in the WinLemm version) - actually each 2 px of vertical movement should take one (I think) extra frame *image (jumper) below
- when building on a miner's tunnel or a 45 degree slope (away from it, not into), the following lemmings cannot access the bridge of the builder, might be fixed having the bridge start by 1 px shifted
- it is impossible to assign a builder to a lemming when he would hit his head before setting his first brick
- I did one level with 97% saved (100% needed), and get the "Rock Bottom!" comment, which is (afaik) only meant for 0% saved

Some things I'd like to suggest:
- the scrolling seems a bit slow to me (fortunately there's still the mini-map to scroll), might get a little faster
- the in/decreasing of the RR works also a bit too slow for my taste
- I'd like to have the Enter key for fast forward, as it is used for Lemmings 2 and Cheapo
- the exit in the hell style doesn't have its horns, in the brick style the exit is not completely displayed (the area where the trigger area is located is black), this should be corrected
- it seems a bit strange to me that the numbers of skills/RR is displayed above the button (I don't know about the Amiga version however)

The digger animation, as it was already said, seems a bit strage also to me.
*) here's an image how the jumper looks like in e.g. the DOS version:

I'm perfectly calm. The emphasis was just to distinguish what was the main point I was making and what wasn't.


OK, I'll do that tomorrow (earliest I can do it, when I'm on my home computer again).

Indeed I liked the idea to let the animations run because it looks a little silly if e.g. the flames stop to flicker just because pause was pressed. Then again, I agree it might be a problem with traps. I think, I will change that. Also the particles of an explosion are not influenced by pause mode. Will change that too, I guess, though it doesn't influence game play.


I know. It's already on my todo list, but at some point, I wanted to release the Beta. Will be fixed.


Climbers climb at one pixel per frame (I compensate the doubled resolution with a doubled refresh rate) just like Walkers walk one pixel per frame. I think this should be ok...


Oops, I must admit I considered this animation to be a "climber to faller" animation  :-[.  But you're right. Will be fixed.


Phew, well, this doesn't sound like a major issue to me. Does this affect gameplay?


Hm, if lemmings for windows doesn't have this animation, I'd have to draw it myself, which will most probably look worse than skipping it. But I'll put it on the list (which is already terribly large).


With miners, this is probably a problem of pixel alignment as (almost) everything in Lemmini is pixel accurate vs. 2 or 4 pixel accuracy in the "original" versions. Builder vs. builder is strange since builders only start at even positions and the step has an even width. You don't have a screenshot of this, do you?


True, but what would be the correct behaviour? Build the first brick anyway and then stop?


True. I was too lazy too find out all comments - there are much more missing. But I'll add them later.



Scrolling is limited to screen updates. So I could only increase the step size.
There's a little trick however: if you scroll with keyboard (cursor keys) and mouse in parallel, the speed is doubled.


Yeah, I might rework this. It's so slow at the moment to allow changing it by one step without adding a more debounced delay.


Now that was easy.


Phew, the hell exit should look exactly like in the Windows version and I can't follow you with the brick style exit. The exit area is not black for me. The exit consists of two parts: a static part "styles\brick\bricko_0.gif" and an animated part "styles\brick\bricko_9.gif". Looks ok for me!?


It was displayed on the button, but there's not enough space on my buttons for this since the images are larger. Also the current font wouldn't work on the buttons.


Could you substantiate this? The animation is "misc/lemm_13.gif". As far as I remember, it cycles once from first to last frame, than applies the dig mask, than starts with frame 0 again.

Maybe you got me wrong there. I stumbled across a lot of problems regarding timing reaching deep into the Java virtual machine and even the Windows core. Because of bugs or strange implemenation on both sides, I had all kind of issues from semi-freezes to a speed up of the system clock by the factor of 2 when running Lemmini.
That's why I listen closely if someone complains about wrong speed.

First of all I'd like to congratulate you. This is a very impressive piece of work, and one that I'm sure took plenty of time and patience. So kudos :)

Just a few things though...

-The music starts straight away instead of beginning right after the trapdoor opens.
-The RR increases/decreases at a much slower rate than usual.
-I tried to build towards a wall sloping at 45 degrees above me, but I simply got the chink sound. The lemming would bang his head, but he should at least make one step before turning away.
-http://www.rocketsoft.gm-school.uni.cc/uploads/Lemini%20mistake.JPG If you look at that, the digger seemed to finish off with just a little gap instead of the usual size, leaving that bit of terrain which the lemmings respond to.
-Just something of preference really, but I think you should use the classic toolbar instead of the Cheapo one.
-Speed up is a bit jerky, and occasionally makes the music stutter.
-If you make a lemming in a small, huddled group a blocker, all the others will instantly react and change their direction. As ridiculous as this sounds (And I'm sorry for being so picky), the lemmings should really ignore it.

Apart from those quirks I'm very impressed, and I look forward to the complete release  :)

You can always build one brick. The most important consequence is that when the lemming would hit his head after one brick, you can use the builder skill to make him turn round. A second consquence is that you can use many builders to get higher and higher until you get right up into the ceiling (or through, if it's thin enough). Indeed, this is the key to the normal solution to Mayhem 2: The Boiler Room.

The hell exit does not have horns in the Win95 graphics.  If someone designed a hires horned exit, 0xdeadbeef might add it to his patches.

Other in game keys:

• S - SuperLemming mode?
• F - Fast Forward
• P - Pause

Edit:

My lemm_13.gif: Animated:

At last someone noticed my ingenious patching framework  ;D

And about the keys: "superlemming" mode is just available in cheat/debug mode (enabled by "c"). In cheat mode, there are some more keys, but I wouldn't advice to press them since they can mess the level data.

I forgot to say: Everytime I say 'pixel' I mean a low-res pixel, i.e. you'd need to calculate with 2 px.


I dunno exactly, but comparing with Winlemm or DOS, it is different. It seems to me that, when digging the first layer away, the animation is shown as usually, and when digging the second layer, the animation is shown horizontally mirrored.


Well, the horns also don't show up in WinLemm high-res, however in low-res.
An a little high-res-like looking option is rendering it with '2XSAI Kreed's Engine", here's an image of the Genesis lemmings exit: http://207.58.177.175/~geoo89/lemmings/Hell_exit.PNG; however it doesn't match well to the other style (colors etc.)
The Bricks exit, I just noticed, the problem doesn't appear in all levels: e.g. in Tame 1 it doesn't appear, however it appears in most of the other levels, e.g. Havoc 5, Crazy 2.; again, here's an image: http://207.58.177.175/~geoo89/lemmings/Brick_exit.PNG


Not drastically, at least I don't have problems with it; the player just needs to change the RR a little later the way it is now.


No problem, here's a little image showing the situation: http://207.58.177.175/~geoo89/lemmings/slope_builder.PNG


I somehow managed to have a lemming survive the nuke, I don't know how. But it happened twice, both times in the level "It's all a matter of timing".

Then I noticed some things in Wild 16: Five alive:
- The bomber, who also breaks through the thinned brick in the WinLemm version, doesn't break through; in the solution on the Lemmings solution page it is actually not needed, but iirc I used it when I solved the level and it also seems intended.
- The miner doesn't turn around when hitting steel
- The miner doesn't take out its mask completely when getting on steel, however I don't know whether it is intended to work that way, it firstly also works that way in Cheapo and is secondly more exact

Small error...fixed.

And don't forget about this.

Hm, can't see any mirroring there, but I'll check.


Well, I think it's possible to add the horns, but it's some work to since it has to be done for every animation frame of "styles/fire/fireo_6.gif" and also the height of each frame mustn't change.


Ok I see. Interesting. I have an idea what this could be, but this needs some investigation.



Ok, I'll see if I can fix somehow. Most probably I'll only have to move the mask a little.


Did you press "c" and activated cheat/debug mode by chance?


Hm, dunno about these two. The explosion mask is a little smoother, but not smaller AFAIK. The miner will turn around if he hits steel, only if he can also continue to walk, he will walk into the same direction.


Miners, diggers and bashers don't erase steel areas. Originally, my bombers wouldn't delete steel pixels as well but I changed that for the sake of compatibility. Dunno if it's needed for the miners etc. though.

Ok, I updated to 0.2. If you're using the Webstart link, you should get the new version automatically.
Unfortunately this forces new a resource extraction since I changed the builder mask.
You will become used to it I guess ;)

(Hopefully) Fixed 0.1 -> 0.2
#  Briefing screen of level one shows "Level 1"
#  Debriefing screen shows code of following level
#  Reworked JNLP to get more RAM (128MB) to avoid potential out of memory issues
#  Fast Forward with Enter
#  Animations + Explosions stop in pause mode
#  No builder restrictions for first brick
#  Reduced climber speed by two
#  Moved step mask a little bit so it should work on slopes (please test)
#  A climber reaching a plateau does the animation "climber to walker" now
   instead of using the animation for conversion from climber to faller
#  Release of first lemming doesn't depend on release rate
#  Digger won't stop until a gap of 12 pixels width below him is free.
   Well it should be 14, but this somehow doesn't work.

Now I should get some sleep. I hope to fix some more issues tomorrow, err today.  :-/

In fact, in standard versions a lemming doing anything else who happens to be a climber, floater or athlete is displayed as "climber", "floater" or "athlete" and not the other skill. I don't know if you want to change this behaviour to match; personally I find it rather annoying.

Oh well, I just need to start somewhere. Maybe I need some kind of professional bug tracking :(


I could explain, but you wouldn't like to know. Basically, I have to work around a bug in the JVM, Windows or both.
Believe me: having 100% CPU load beats the alternatives by far.
"Restart level" is so slow because the level is completely reloaded (including style) and redrawn. It's pretty fast on my machine, but I noticed it on other machines as well. Is on my optimization list.


Ah ok, I messed that last last time I was working on the "break on steel" conditions. Will be fixed in next release.


Will be fixed in next release.


Will be fixed in next release.



Hm, I actually like the pause mode the way it is. But I'll consider the automatic unpause.


Sure I can, but it's a bit more work behind that.


I put it on the list.


Does this influence gameplay in a nagative sense?


Ok, will be fixed.


Yes, it looks like in WinLemm. Unfortunately, not all the Hires Graphics look good. Indeed most of them were just upscaled and look horrible. I fixed some things here and there, but it would took me too much time to fix all tiles. Also the size of the patches would dramatically increase.


Indeed blockers only affect walkers at the moment.
In the next release, blockers should also affect builders, bashers and miners. However that'a a little difficult to check.
Also maybe the blocker mask has to be changed to affect builders more effectively.


If necessary, I thought of patching the critical levels instead by introducing a game time based on seconds instead of  minutes.
I'd need a list of levels with second based times for this, though.


Graphics modifications have very low priority for me at the moment. However anybody is invited to to this. The resources are there.


Put it on the list. Indeed I have to rework the whole "lemming under cursor issue", so I might add this then.

For me only levels 17 - 20 are missing. But this is since they're also missing for WinLemm.
I don't know for level 15 and 16 however.

Two more things to add:
- any terrain removing action leaves a black space in the mini map, I think it should be the background green
- lemmings can get past a blocker when they're walking on a little higher level than the blocker and the blocker is placed at a certain point: Situation in Taxing 6

OK, I've had a bit of a play on version 2. Much improved, and SO MUCH FUN!

A couple of points I notice:

-- "All lemmings accounted for" should have a full stop at the end. Yeah, I'm picky. Sue me.

-- I played quite a few levels without crashing, and then got the crash when I tried to increase the RR again. Tame 8 this time.

-- I don't agree with ccexplore that the 2 and 8 in the "number of skills" font are hard to distinguish; 8 and 9, on the other hand.....

-- The toolbar should say that number of lemmings currently under the cursor after "WALKER" or whatever. And the IN percentage shouldn't have a leading zero if it's under 10%.

-- Steel detection seems to be too wide; on 7 Crazy you now need pixel perfect accuracy to break the ceiling and avoid the steel. You should be able to bash even if a few pixels of the bottom of the steel are within the basher's field.

-- I love, love, LOVE the slower screen scroll movement. Don't change it.

Yes, of course it is  :P  As you see, I had second thoughts and deleted that comment (namely, as soon as I noticed all the other ONML musics are different from what I'm used to too, I realised they must just be different on the version he's using).

In fairness to you and other people who've complained about this maybe I should explain why I'm so enthused. Firstly I'm not sure whether it's Cheapo or other version I've played, but I've definitely been annoyed in the past at versions that scroll too fast. Secondly I love the feeling of exploring the level, taking in its design and not just zooming past everything before you get a chance to notice it. Thirdly you do have the mini-map for when you need to jump straight to a particular place, and most of the time when your builder is heading off-screen and you need to catch up with him, you will only want to scroll a short distance. The slower scroll makes it much easier to do that and stop exactly where you want.

I will check, but at first look it doesn't look to bad for me.


The dots are already extended to the right and lower side. I added rounding, but this will not "fix" this as it's a matter of loss of precision that has to be expected.


True, only walkers, bashers and miners were allowed to exit. I now added builders and diggers.
Other interactive objects are not limited.


For exists see #17, they should drown on water however. I just drowned one. but it's not easys to do as the water mask is a little to small.


Yeah I know. It's an optimzation issue. To fix minimal visual issues like this I had to copy the draw background with bitmask transparency.
I think the performance loss on low end systems would be substantial. Anyway, this is low prio IMHO.


Ok. Will be fixed (I hope)


I'm not even sure if I understand the problem. Should not affect gameplay though, right?


Very likely. I'll assign every second frame.


Hell, you're right! Should be fixed in the next release.
 

True, will be fixed.


Couldn't reproduce this with my current version. I would think this is the same issue as your #0 "The lemmings don't fall into the gaps".
If so, it will be fixed in the next release.

Sounds good. Let's hope you like the next version even more.


Will be fixed in the next release. I will rework all the "debriefing" screens later. Lots of messages are missing.


Damn, I hoped it was a memory issue. I'm a little puzzles what this could be. Maybe the sound?
Did you try to start the JAR from the command line. Exceptions of the runtime environment (out of mem etc.) can't be 100% catched by the program, but there should be a error text printed on the command line.


If the more important issues are fixed, I will have a look if I can improve the font or find another one. This was coming with Lemmings for Windows (though it's obviously not used).


The Amiga version definetely has a leading zero. E.g. it display "00%" at the start of each level.
The "under cursor" thing will be completely reworked including priority etc., but this might take some time.


I'm not sure what you mean here. Screenshot? "Pixel perfect accuracy" is something that nearly impossible with the high resolution and all. Steel detection will be a little bit tweaked in 0.3.


There are more important issues anyway. Maybe I'll make it customizable...
Or what about speeding it up when pressing space, shift or whatever?

Does to me; I noticed this problem too.


Needless to say, I want everything to be faithful to the Mac version  :P


First you need to know the solution to Crazy 7. One lemming climbs/floats to the left, builds up to the wall holding the lemmings, bashes right the way across. On the actual game, you have to be a bit precise in where you place the bridge -- you need to strike the wall about one lemming's height below where the crowd are contained, so that you're not too low that you don't release them, and not too high that the steel stops you bashing at all -- but because of what I said, you can still bash if a few pixels of steel overlap the bash tunnel, you don't need pixel precision. On your current version, you do, which is bad news if as you say the high resolution makes the required precision impossible!

Speeding it up when you press shift is a good idea.

I think this is most noticeable when building to the right.  I'm less sure about building to the left


Well, as long as it works most of the time I'm ok with it.  The main thing was that right now you could see the dots floating quite often.


To avoid repeating variations of this at later times, I think detection should be added for lemmings in pretty much every state except possibly those related to dying.  That's pretty much how it works in the real game for example.  (edit:  Falling lemmings should probably also be excluded from exit detection actually)


It still looks to me though that when the digger drowns, he's floating at a lower elevation than a lemming that drowns after falling into water, which doesn't seem right to me.I think the problem was actually due to the digger's vertical position being lower than it should when switching from one action to another.  Since you fixed that problem, it has appeared to fix the drowning issue I striked out.


Basically one should only see these 2 cases:
 1) Lemming walks towards cliff, reaches countdown 0 before falling, stops walking, shudders, explodes.
 2) Lemming walks towards cliff, falls before reaching countdown 0, explodes immediately w/o shuddering.

I'm reporting that I'm seeing a case of mixing up 1 & 2
 3) Lemming walks towards cliff and falls, but nevertheless still shudders first (while also falling at the same time) before finally exploding.

I guess this arguably doesn't affect gameplay too much since you need good timing to repro.  Still, it can be noticeable when nuking.  Anyhow, I'm ok with leaqving this as low priority.

Also, maybe this bug is just another manifestation of #0.  Let's hope. ;) [edit: looks like it is, so I think it's fixed.]

http://www.rocketsoft.gm-school.uni.cc/uploads/Lemming%20jip.JPG

I know it's not very clear in that picture, but if you look the lemmings don't seem to respond to the bridge. It seems that starting a builder lemming on a downwards slope will have them start a pixel to the left of the surface.

EDIT: Also: http://www.rocketsoft.gm-school.uni.cc/uploads/Lemming%20Oh%20no!.JPG

The digger lemming there is actually stuck, he's going through his animatipn but he won't budge :-/

Before I get my well deserved sleep, here's another version.

FIXED 0.2 -> 0.3
#  Walkers now turn to fallers immediately if the fall distance is far enough (solves problems in e.g. Mayhem - All or Nothing
#  Fun 9 renamed from "I have a cunning plan" to "As long as you try your best"
#  Digger now stop (again) if they hit steel.
#  When converting a digger to another skill, the new type is no longer starting 4 pixels too low
#  A lemmings assigned to digger skill, will immediately apply the diging mask
#  Blockers affect builders now (also bashers and miners if someone managed to create such a situation)
#  Assigning a skill (successfully) to a lemming unpauses the game if it was paused
#  Builders can exit a level
#  Improved error output
#  Blocker status is not cleared when the blocker is about to explode, but only when it has exploded
#  Fixed position at which lemmings enter the level (a little higher and 2 pixels to the right)
#  Nuke assigned to new level only every 2nd frame to compensate for higher frame rate
#  No more lemmings should survive a nuke
#  Explosions clipped to level area
#  Small congratulations text (preliminary) instead of crash when last level of a difficulty level was won.
#  Newer versions of MOD files (?)
#  Changed order of music files played. Most probably still not 100% correct.

Just some information about the builder mask: Building to the right, the leftmost pixel of the step is exactly where the lemming is. Building to the left, the rightmost pixel is one pixel to the right of where the lemming is.

EDIT: Still haven't tried it, no winlemm files.

Excellent! The music order does appear to be right now (though I haven't tried every level.....)

But, I now get the crash happening sometimes just when I'm playing, there doesn't seem to be a particular event that triggers it.

If there's no message box displayed it's most certainly an issue of the runtime itself. so again: please start the game for the command line ("java -jar lemmini.jar") and tell me what it prints out there if it's crashing...

@cce: This appears to be modeled after winlemm, which I know nothing about. uhh.... just ignore me until I get winlemm.........

If I understood any of that, I would......

Ok, let's try it again:

1) Download the jar from my website or directly via this link
   http://home.arcor.de/0xdeadbeef/lemmini.jar
and save it somewhere on your machine.

2) Open a command line (command.com or cmd.exe). E.g. by using the start menu, then Execute/Run, then type "cmd.exe" or "command.com" (without the quotes, then press enter).

3) Change to the directory where you saved the JAR file.
E.g. if you saved it on "d:\temp", type the following lines in the command line (press Enter after each line):
   D:
   cd \temp

4) Start the JAR from the command line by typing
  java -jar lemmini.jar

5) Play until the game crashes

6) Look for error messages on the command line
If there are any, make a screenshot or cut out the text and post it here.


And some information about your system would be helpful (RAM size, CPU type/frequency, VGA card, desktop color depth).


About all the other topics: sorry folks, I was long in the office today, debugging some code for a peripheral processor that I messed up while optimizing it for size. So I'm kinda not in the mood now to continue debgging. Also I'll be away for the weekend in two hours. So there will be no updated before Sunday or Monday. We shall see.

Thanks.

I'm afraid I didn't succeed in carrying out step #6 -- ctrl-alt-del wasn't working (or it was, but the program wouldn't "End Task") so i couldn't get out to see what was in the command line and had to restart.

Two small things I noticed while playing:

* ccexplore complained that the safe fall distance is too *high* on Mayhem 1 and they survive the initial fall when they shouldn't. But also, it's too *low* on Tame 4 -- there are several places you can see this, notably falling left off the platform just over the exit. On every version I've played they should survive that.

* I see the Tame 1 exit has been improved to get rid of the unsightly black rectangle -- but Tame 5 still has it!

Yes. Pretty much the same thing that happened to DragonsLover (page 1 of the thread). Since he also has Windows 98, maybe it's something to do with that?

An interesting point -- I did in fact once have the crash problem during a nuke. On the other hand I have several times changed the RR or nuked without triggering it, and I have also had it once when I wasn't doing anything that caused sound effects to play at all.

Hm, I changed quite a few things over the weekend. Hope I didn't mess too many things.
For the guys with crashes: please try to deactivate music and sound.
I'm also still waiting for some system info (RAM/CPU, DirectX installed etc.).


Fixes/Changes 0.3 -> 0.4
#  Fixed ArrayIndexOutofBoundsException when selecting A Blocker that is about to explode
#  When a blocker that is about to explode ("oh no") falls, the blocker mask is erased
#  Blocker skill can't be assigned when overlapping a blocker any more
   However it's (probably) still possible to create a gapless "line of blockers"
   Was this possible in the original?
#  Erased trailing spaces from Fun 9 level name "As long as you try your best"
   that caused the string being displayed too far to the left
#  Changed the way fast forward (and Superlemming mode) works. Should work much better and with
   less CPU load on mid-range machines. Could cause frame skip on very low end machines though.
#  Pressing shift doubles scrolling speed
#  Dragging right mouse moves screen (well, not perfect, but a start)
#  When a Blocker reflects a builder, only the direction is changed, but the builder keeps building
   Also a reflected miner keeps mining
#  Fixed masks for flamethrower (fire) and steam (snow)
#  Shruggers can explode now (hard to check though)
#  When a Blocker is trapped, the blocking mask is always removed now (well, I hope so)
#  Exit mask visible on some Brick exits was resolved
#  Additional skills like climbing, floating etc. can't be assigned to lemmings that are drowning,
   leaving the exit, splatting, trapped or exploding.
#  Hidden exit in Havoc 19 works now
#  Exit in Havoc 15 is working as it should
#  Explosions speed up with FF and stop when paused.
#  Menu option to disable sound
#  Completely reworked lemming coordinate system. This should fix this like shruggers being offset
   and make it easier to become more compatible. However it is very likely that I broke some things
#  Digger animation corrected (doubled animation length)
#  Fixed (WINLEMM) level errors in "Fun - Live and Lem" and "Mayhem - Steel Mines of Kessel"
#  Fixed typos in orig.ini resultin in wrong stats for "Tricky - Lemming Drops"
#  Messed around with fall distance and entry height

Would it be possible to make shift double the speed at which the RR changes too?

EDIT: or an even better idea, increase the speed at which it changes with a normal click, but make shift-click change it by just one at a time, making it easier to get a precise number......

I'll try out the new version very soon.

When on the level intro screen, the level starts upon mouse button press, could you have it wait until the button is released?

Due to your new changes indeed some new bugs appeared:
- builders hitting their head don't turn around anymore, they just keep on building
- when a digger has finished digging or is assigned a skill, he'll walk/execute that skill to the right, no matter what direction he went before
- the bug that lemmings don't walk up bridges starting on slopes (http://207.58.177.175/~geoo89/lemmings/slope_builder.PNG) is back again, however it only seems to happen in special cases: I could only reproduce it in Tame 5 on a miner's slope building to the left, and even then it happened not everytime

I don't know whether you forgot or justs did not do, but also objects should be speeded up in FF.

The we all fall down levels are solvable now and Mayhem 1 works correctly, but 'Take care, Sweetie' (Wild 16) is still impossible: the second fall on the left should actually be survivable, but it isn't. Since everything seems to work fine for original lemmings, perhaps you could do it the same way it is done in the DOS versions: setting the max-safe fall distance of ONML 3 (low-res)-px higher than the distance for original lemmings?

Small update only (yet no resource extraction this time)...

Fixed/Changes 0.4 -> 0.41
#  FF speeds up object animations as well
#  Some memory optimizations - Memory footprint should be distinctly reduced
#  Builders stop building when reaching ceiling again (bug introduced in 0.3)
#  Diggers "remember" direction (bug introduced in 0.3)
#  Fixed error in fall distance calculation for walker converted into fallers -> fixes lethal fall in WILD 16
   (bug introduced in 0.2 while fixing ccexplore's bug #0)

About "safe distances": I'll most probably put that in the ini file of the levelpack (orig.ini and ohno.ini). So people can also define their own safe distances in their own levelpacks. Anyway, at the moment, a lemming will splat if it fell at least 123 hires pixels. Following ccexplore and others, it should survive even a fall of 126  hires pixels (63 lores pixels). I had problems with this higher value though - but I begin to think this was related to the wrong fall distance calculation for walkers turning into fallers.
Will have to check again, but most probably I can go up to 126 being safe now...

I've talked to ccexplore about this before.

With 0.41 (or 0.4.1), at this point:


I get this error:


Every time.

Ok, again a new bug introduced by a bug fix (giving the digger back his direction) :-[
Will be fixed this evening - I can't sign the Jar from work.

Just a small update, yet with resource extraction for the builder issue.
Hope I didn't mess any things this time...

Any news regarding crashes on '98?

Fixed 0.41 -> 0.42
#  Fixed Builder mask (again) to work better on slopes (resource extraction)
#  Fixed ArrayOutOfBoundException thrown when accessing second direction for lemmings
  with only one direction (bug introduced in 0.41 when fixing digger direction
  "memory")

When a climber climbed out at the top of the game field (Havoc 7: 'Creature discomforts', lemming of the left entrance is made climber), I got the following error message: http://207.58.177.175/~geoo89/lemmings/Error_1.PNG

Also, climbers (output) look as they'd be climbing a little in the wall, and at least for climbers facing left, I noticed when they hit the top, there's one frame they get walker (you can assign any skill at that point) before they fall.

And also, builders hitting their head aren't checked correctly still: they can still build a few (three or four) steps too far when their head is already in the terrain, I suppose it's just the checked pixel that is the wrong one.

Fixes/Changes 0.42 -> 0.43
#  Builders shouldn't be able to leave the screen any more to the top
   causing ArrayOutOfBoundExceptions
#  Added basic priority handling for "lemmings under cursor"
   Please tell me if it's ok this way.
#  Reworked entry timing: Lemmings can only enter level if trapdoor is fully opened
#  Music starts only after trapdoor opened
#  Floater, climber or both skills (athlete) are displayed as "(f)", "(c)" and "(a)" behind
   the main skill. Just a test to hear some opinions on this.
#  Reworked Cursor: active Lemming gets its own border, so it's clearer which lemming is
   active. Mixture of Winlemm and Miga look. Also the Cursor is the default Cursor now,
   so it moves smoother

Just a couple of notes:
- Lemmings don't get the updraft that they should when opening thier umbrellas.
- For the sake of future customizablilty, please use separate graphics for left and right animations (e.g. walkers, fallers, etc.).

Again... (you probably just missed it earlier)
Actually, by version number standards, 0.41 is newer than 0.5 since it would be the 41st minor revision. Bug fixes usually go under the third revision section (e.g. 0.4.1)

@MisterDeadbeef:

Can you supply me with some technical information about graphic en level files?
I'm starting my own lemming-clone programming project (it will be freeware too).
Thanks,
Eric

For info on SPR and PAL formats, have a look in this thread.
http://forum.lemmingswelt.de/cgi-bin/yabb2/YaBB.pl?num=1132849781
I posted about everything I know there.

For the LVL format, just google for "lemmings level format".

BTW: do you aim for smaller/different platforms (PDA, cellular phone, GBA) or why would you start your own remake?

Ok thanks.
I start my remake because of hobby. I'm programming databases now and this is much more fun.
I'm using delphi so it's a pure win-application (or eventually Linux)

So you use winlemm as a base. So I must have this.
What about the DOS files? (main.dat, vgaxx.dat, groundxx.dat, levelxxx.dat)

I got an error message caused by the new addition to the text saying what lemming is selected. http://207.58.177.175/~geoo89/lemmings/Error_2.PNG
It was outputting "Walker(f)-12" and was giong to switch to "Builder" which was probably too long.
About the general system, I don't know too much about the original one, but it seems ok, as far as I checked.

Having the normal cursor AND the border for the lemming is a bit...strange (to me). I don't like it actually. Waiting for other opinions...

OK, I did a "silent" update on this one. Should not happen again...


The Windows version does the same thing, but border and cursor are monochrome and less bold. Maybe it would look better with a little thinner graphics. Anyway, I must admit I like the idea of split Lemmings frame and cursor for a number of reasons.

Search the forum. Some guys here (ccexplore and Mindless I guess) investigated the DOS files. I was only interested in the Hires graphics from Winlemm.

What is the best free disassembler that anyone here knows of?

Did you try PvDASM?
http://pvdasm.reverse-engineering.net/index.php?Section=1

I must admit however that I used the trial of PE Explorer the last time I felt the need for a disassembler.

Then you seem to have a different version of WinLemm since in my version either the cursor (when no lemming is selected) or the border appears, but never both.

I actually like the idea that the selected lemming is pointed out. It's just that having additionally to the cursor the border around the lemming makes it unintelligible and somewhat strange.
Just to give a different aspect, in Cheapo an arrow from above is pointing on the lemming that would be selected.

Fixes/Changes 0.43 -> 0.44
#  Reworked whole redrawing scheme. Redraw will take a little longer
   but "hidden" exits don't need special treatment any more. Also graphical
   glitches like removed pixels in water should be gone now.
#  Checked/Fixed lemmings leaving playfield and according Exceptions
#  Fixed Diggers "hanging" on steel in special situations


Well, I'd bet I screwed something when changing the redraw since this was a major one.
I'd be also interested if the performance was degraded distinctly from a prior version
to 0.44.

That looks good for a PE, but it can't disassemble lemmings. (which is what I'm trying to do, I should have been clearer)

Then why don't you use a trial of one of the commercial disassemblers (PE Explorer, BDASM, IDA Pro or whatever)?
At least some of them should still support 16bit executables.

The trial of IDA Pro wasn't good. I shall use google.....

EDIT: All disassemblers seem to be:
free, good, can disassemble dos exe   -   choose 2

Are any of the ones listed on Wikipedia of any use?

To get this back on topic, the usual release after midnight (boy I need some sleep)

Fixes/Changes 0.44 -> 0.45
#  Reworked MiniMap internals. Also "hidden" exits are no longer shown neither in the Minimap nor in the level preview (briefing).
#  Reworked Floaters: there's a little "jump" up when the parachutes opens
#  New debouncing for release rate icons/keys: first press is debounced, then the RR increases/decreases faster
#  Reworked nuking. Now really no lemming can survive a nuke. Also types like exiting or drowning lemmings should get nuked.
#  Messed around with "free pixels above" again to improve builder behaviour.

I FINALLY got around to playing this.  Excellent, excellent, EXCELLENT job.  I like it much better than WinLemm.

Nice to hear that some folks actually like it  :D

Anyway, here is version 0.50  (and no, I won't change my chaotic version numbers  ;D)
The game is pretty much complete now feature wise. I will probably rework the menus a little and may add replay if anybody wants it, but I think it's ok to call it a beta now, since it's pretty close to what I imagined to be the first "real" release.
I also think I fixed all of the bugs noticed by ccexplore and other testers (apart from graphical thinks like horns, colors etc.).
Yet I am convinced that there are still lots of bugs around. So give it a try.

Fixes/Changes 0.44 -> 0.50
#  Changed: Player handling is in (Menu "Player"). This also means that only the first
   level of each difficulty level is available by default (see below).
#  Changed: new Menu item "Enter Level Code". The Amiga level codes should work.
   To enable all Levels for testing, use levelcode "0xdeadbeef". This also enables debug/cheat mode
   (press "C"). In cheat mode, create a lemming at the cursor position with "space" or draw/erase
   bricks by pressing left/mouse button after changing to draw mode with "D".
#  Changed some pixels in the number font. So 2, 8 and 9 should be a little better to keep apart.
#  Fixed entry order for levels with 3 entries (was shifted by one). However, WinLemm doesn't
   have special treatment for 3 entries at all.
#  Reworked dirt_39 since a one pixel offset would lead to a fall in Mayhem - 1 being not
   lethal.
#  Increased safe fall distance to 126 and moved this as configuration entry to the levelpack inis
   (so "oh no" and orginal lemmings could have different safe fall distances). To keep Mayhem 1 fall
   from trapdoor lethal, I changed the entry (foot) position to 20 hires pixels under the upper
   trapdoor coordinates. I may add that in Winlemm, the fall from the trapdoor is NOT lethal,
   but it is in the Amiga version.
#  Added more debriefing screens. Should be complete now???
#  Fixed: Scolling via MiniMap or mouseDragging could cause screen flicker.
#  Fixed: on the rightmost border of the level, dragging the mouse left didn't work
#  Reworked homepage a little: some minimal instructions can be found there now

Debug cheating is way too fun.  ;D

I think the debug features are really helpful for testing borderline cases - that's why I posted how to enable them.
Then again, I must admit I caught myself trying to solve levels by only painting and erasing in cheat/draw mode  ;)

Spend the whole morning investigating why the java process would get more and more memory while the heap inside my program was constantly low. Indeed I found one clear memory leak (music), one resource leak (sound) and one not so obvious resource leak (painting levels). The music issue is my fault, whereas the other two issues are a little obscure since the JDK documentation doesn't really tell that the resources HAVE to be freed explicitly.

Anyway, while the memory used by 0.50 could easily reach 300MB and more (when playing lots of levels and playing lots of sounds - especially easy when using the RR buttons), the footprint of 0.51 should always stay below 64MB.

Considering the sound related crashes on Win98, this could be solved now. Give it a try.

Fixes/Changes 0.50 -> 0.51
#  Fixed (kind of) resource leak when playing sounds (Clip.close() via LineListener).
#  Fixed memory/resource leak when loading a new music
#  Fixed (kind of) resource leak when painting level (flush() is used on all Images/BufferedImages)

Well, the RAM usage stayed steadily at around 64MB.  That's good, but I found a bit of a bug.  Upon finishing a level, the music starts to play like the computer's frozen.  It'll play about 1-2 seconds worth, then go back to the beginning of that 1-2 second fragment and play it again, and again, and again.  This only happens on the "Level Complete" screen.  As soon as I click away from it, the music fixes itself.

I also noted that Lemmini runs smoother than any of the official ports that I've played (PC/Amiga/Mac/Win95/you name it).  I like that.

EDIT:  Another thing that bothers me is that the noise the builders make on their last 3 bricks is supposed to be the clicking noise, not that sound made when a miner hits steel.

While this didn't happen at my system, I changed the music stop code a little for 0.52. Could you try it again?


This might be due to the fact that the screen update speed is doubled to compensate for the higher resolution.



From what I can tell, this is the same sound as in the windows version, so most probably I just don't have another one. Are you sure the sound should be different?

About the builder sound, I was thinking of the Amiga version, and it's like that in the Mac version too.  The Windows version wasn't right.  I checked the WinLemm data files, the CORRECT sound from the Amiga/Mac version is TING.WAV.

Will check the 0.52 ASAP.

I'm afraid my computer at home isn't working at the moment and has had to go in for repair, so I haven't had a chance to test out the new version yet.

Some smaller updates...

Fixes/Changes 0.53 -> 0.54
#  Fixed: stopper mask was not completely removed in case of explosion etc., which made setting a new
   stopper at the same position hard to impossible (walkers were usually not affected though).
#  Fixed: bug in LineListener. Maybe connected to "freezes" happening from time to time since
   I introduced the LineListener. Most probably not, but who knows.
#  Fixed: Lemmini doesn't work on Macs because version number contains characters (0-beta)

Fixes/Changes 0.52 -> 0.53
#  Slowed down RR changes a little
#  Fixed empty masks when using miners/bashers/builders to the left.

OK, I've had a go with the new version.

No crashes when changing the RR as yet (though I'll keep trying and see if anything happens).

I did however get a different sort of crash when using the pause function. The music stopped, and when I opened a different window and went back to Lemmini the screen was grey and I couldn't get it to respond except by using ctrl-alt-del > End Task.

A couple of suggestions, not related to the crash thingy:

* "Squasher" is a very silly name for the third Lemmings style. Why not use one of its standard names, "pink" or "marble"?
* How about adding option-click on the release rate buttons to change it straight to the minimum or maximum?

Ah, that's something.


Hm, interesting. Is this reproducable? How long did you wait before shutting Lemmini down? How much free RAM do you have?
If you can reproduce it, please also try to disable sound and see if it still happens.


The name squasher stems from the LVL format description by "rt". I must admit I don't really cared since this is an "internal" name anyway. Is there any info available what's the "official" name?


I'd bet ccexplore would complain this would make some levels too easy.


BTW: It really runs pretty well under Linux (just tested with Knoppix)

Fixes/Changes 0.54 -> 0.55
#  Changed drawing in minimap (alpha value always 0xff) to avoid different background color under Linux

I'll keep trying.... when I have time. I'm afraid time is really difficult for me at the moment. Essays, you see.  :(


I don't think any of the styles have "official" names. Except on Cheapo, of course, where that one is called "pink". I've also seen it called "tile" or "marble".


It can't affect difficulty, since you could always achieve the same thing by pausing and changing the release rate the slow way. I'm not proposing it be possible to lower the release rate below the initial value for that level, of course, just that you can get back to the minimum value in one click. Just a little convenience to the player, you see. (Or a big convenience, given that wanting to change the RR to the maximum or minimum is much more common than wanting to change it to a precise in-between value.)

I'm afraid I got the RR-change crash glitch again.

How do I find out how much free RAM I have? I seem to have 256MB but I think that's total RAM.

I'll try with disabled sound next and see what happens.

EDIT: tried, and it seems to be crash-free. The other bug I mentioned (with pausing) also doesn't happen. And I checked the fatal fall height in Tame 4; it seems to be OK now.

A few more points, though:

* In normal (non-fast-forward) mode the screen seems to flicker a lot; it makes it painful to look at and also seems to affect lemming selection -- it doesn't work if I click while the screen is flickering.

* When I nuked and then right-clicked to return to the main menu, the text in the middle of the intro screen didn't display correctly.

* Your "WALKER(c)" and "WALKER(f)" displays to show that a lemming is a climber or a floater might look neater with capital C and F to match the other capital letters. Maybe a space between WALKER and the bracket, too.

* And when there are a number of lemmings under the cursor, the display WALKER-2 looks like it's talking about a negative number  :P  Every version I've played just says "WALKER 2" with a space, no dash.

* RR change seems to be faster on Lemmings levels than ONML ones. Could be just me though.

* It's irritating to lose the player (and hence which levels I've solved) information when I download a new version. Could this be remedied somehow? Also, how about a command to delete players (if I get tired of one, or just if I typoed)?

EDIT 2: I can't seem to reproduce the pause crash glitch even with sound enabled. Could have been because I had a lot of other programs open at the time, perhaps?

Phew, that's not too much. I stripped down Lemmini to use only 64MB of heap, but when playing sounds, the Java runtime environment seems to allocate additional RAM which I can't really control. I would guess that this is the source of your problems and I don't really know what to do against it.
I didn't use Win98 for many years. Doesn't the Task Manager show the available/free RAM?


I could limit the whole sound output to just some output lines. I guess this would reduce the chance of crashes since to RAM outside the JRE would be allocated when playing a sound. However this  is a major rework and I'd rather do an editor (no promise) than that. I'm also not sure if this would really fix the problem. But maybe I'll do it anyway.



This is a complete mystery to me. The whole screen is drawn to an offscreen buffer, than the offscreen buffer is copied to the frame buffer in one go. That's the maximum I can do to avoid flicker. I'd suspect this could be something in the JRE as well.
What colour depth does your desktop have? I'd guess Lemmini wil run best for truecolor modes. If you don't use truecolor, could you give this a try?


Phew, the screen is blanked and completely redrawn. This really shouldn't possibly happen. Can you reproduce this or make screenshot? Anyway, I think this can really only happen "outside" Lemmini. E.g. memory corruption or bugs in the JRE or whatever.


In my (fading) memories, the Amiga version had dashes. But probably I'm wrong.
@all: Any other opinions on lowercase f/c/a and dashes???


Lemmini doesn't really know if a level is original or ONML - it just sees a level. I could only image that if your CPU load is 100% that maybe some levels with more objects are slower and the the whole redraw is slowed down which would also affect the release rate. Did you check your CPU load when running Lemmini (e.g. Task Manager). Is it close to 100%?


Hm, is it really overwritten? I must admit I didn't notice it since most of the time I play my local version of course.
I will move the player directory to the same directory as the resources.


Hm, how stable is your system anyway?
Just to get an impression - could you try Paradroidz? This is a Java remake of the C64 classic Paradroid in 3D:
http://www.jpct.net/paradroidz/

Fixes/Changes 0.55 -> 0.56
#  Speed up cleaning of bgImage when changing levels (not noticeable though)
#  Moved players directory ro resource directory and player definitions from main ini to players.ini
   in the resource directory. Also allowed capital letter for players ini files.
#  Skill attributes a,c,f now displayed in capital letters A,C,F
#  Space is displayed instead of dashed for number of lemmings beneath the cursor.

Um...... where will I find Task Manager? (Sorry to be such a hopeless n00b...... I've always hated Windows and the way it seems to go out of its way to be as user-hostile as possible. But now I'm stuck using this computer because it's the one that has the internet connection. Damnable, really.)


I don't want you to do a major rework just to make it OK with antiquated systems!


Yes, I do use truecolor.


I'll have a go.

No, it doesn't, it only lists tasks.  Process Explorer might tho.

Wow excellent work with this, ive just downloaded it so i'll give it a go and get back to you if I find any problems. Hopefully I wont post any which have been posted before  [smiley=cool.gif]

Screenshot:

EDIT: screenshot taken down as the problem seems to be solved in the new version

Fixes/Changes 0.56 -> 0.57
#  Added a sychronization monitor to paint()/redraw() to (maybe?) avoid flicker on low end machines
#  Added a sychronization monitor to the textscreen class to (maybe?) avoid redraw issues on low end machines
#  Double-Clicking the release rate icons will jump to maximum/minimum value.

Excellent! The flickering is completely gone, and I note that the player information retrieval works perfectly too.

And did you notice you can change the release rate to min/max with just one double click  8-)

Fixes/Changes 0.57 -> 0.58
#  Limited number of sounds to played in parallel to 6

It didn't take much noticing, since you drew attention to it in your post  ;)  But thank you.

Right, now that I can actually play without it crashing all the time, I'll keep testing it and see what I notice!

Just a couple of points thus far:

* The colon is missing in the title of 1 Crazy, "Quote: That's a good level".
* When you nuke and return to the main menu, the window titlebar still shows the level title. I don't think it should.
* Sometimes a click doesn't succeed in assigning a skill; it seems a bit hit-and-miss. It's annoying, because it means I have to pause an awful lot on the higher-precision levels.
* Stretching a bridge is hard or impossible. By that I mean that on other versions, when a builder shrugs and becomes a walker, he can always take two steps before falling off his own bridge, so you can start the next bridge a little bit along. I can't get this to work in Lemmini.

Would be nice to have no crashes any more on Win98. I'll keep my fingers crossed.


This is somewhat interesting. Actually, the colon is also missing on the Amiga version of ONML. However it's definitely in the Win95 LVL file. Then again, thereis no colon in the Amiga font. I wonder if it was just forgotten? I could easily add the colon to the font of course and most probably will do.


Agreed


I'll increase the minimum distance for a lemming to get considered to be beneath the cursor.


While stretching a bridge is actuall quite easy IMHO, as you have the whole "shrugging" phase to assign the builder skill again, I'll look if I can/should shift the lemmings or step position a little. I moved the step to the opposite direction lately to fix the "builder on slopes" issue.

Yes, assigning another builder to a shrugger is easy. I'm taking about stretching the bridge by letting him take one or two steps AFTER transitioning back into a walker. So that the bridge reaches further horizontally than it would if you assigned the builder immediately. It's possible and not too difficult on all other versions of Lemmings.

Ok, I see. Most probably I'll just have to move the shrugger a little more to the left (if building to the right), so if converted to a walker, he'll start further to the left.

Fixes/Changes 0.58 -> 0.59
#  Increased distance at which a lemming is considered "beneath the cursor"
#  Changed window title back to "Lemmini" when jumping back to the intro screen.
#  Use of own synchronization monitor for paint()/redraw() instead of using image as monitor
  (else first paint() could throw a NullPointerException resulting the menu being corrupted).

Whew! Good work for finding all that out and thanks for going to all the trouble of explaining it! I do hope 0xdeadbeef can manage to make some use of all that information.

However, I have a query. You say that climber/floater status doesn't affect priority. But what about the "floater method" for getting out of a situation where the lemmings are crowded in a pit and it's imperative to build out in the right direction? (11 Taxing, 26 Taxing etc.) I can think of three possibilities -- (1) you missed something; (2) it only works on certain versions; (3) it doesn't work at all and the times it's succeeded for me have actually just been luck.  :-?

Well, I'll have to take your word for it..... that's disappointing. I really thought I had something there....


OH! When I read this bit I was thinking "wtf, the floater isn't any more likely to be a lemming that comes out late", but of course that's not true, because when you select the floater priority will be given to a lemming that comes out late. (Which I think might mean that the opposite of your second is true -- it works best when the area the crowd are confined to is as small as possible. But it's hard to calculate exactly.)

So the method does work but not with certainty as I originally thought. Still worth knowing, I think (or hope). Thank you!

Probably not, but any discrepancies that actually make the level harder to solve should definitely be got rid of. Like the bug I noticed to begin with and still don't know if it's fixed or not, that if you had two lemmings under the cursor, one a builder, sometimes it would look at that lemming and think "already a builder so can't assign builder skill" and not even consider the other.


Because I didn't know what I know now about skill selection mechanics!  ;)  In fact, I think I can use this to provide a method that really does give certainty:

(1) Make last lemming a floater as it leaves entrance
(2) Hold cursor just touching edge of crowd confinement area and wait for floater to appear
(3) Calculate when floater will be in the centre facing the right direction

Similar to what you were saying, but you see that it doesn't depend on "tracking" the lemming from the time it leaves the entrance to when you're ready to release the crowd.

Damn, ccexplore's vacation is already over  :-/  ;D

Better start sleeping in the office! ;)

Whether that's a good or a bad thing depends on which he has more time for -- finding more faults with Lemmini or getting back to work on my MIDIs!  :D

About the passwords. There are indeed many valid passwords for a level. Maybe it would help if a we collect a list of valid passwords for a level and try a little "reverse engineering". Try to see what's the same about these strings...

I think the logical solution would be to follow Cheapo -- the lemming being killed isn't actually a lemming any more, it's part of the trap animation. So cursor should say "WALKER 1" if there's one other lemming under the cursor and be blank if he's the only lemming under it.

Oh -- you're right that I confused the two types of traps, but I was still at least half right. Turns out that with constant traps, Cheapo doesn't decrease the number of lemmings IN until the death animation is over, but just as I said, as soon as the animation starts the lemming no longer counts when counting how many are under the cursor.

I actually disagree for a number of reasons. Firstly, the levelcodes are just a bonus. The way Lemmini is supposed to be played is by clearing one level after the other and in contrary to the Amiga version, Lemmini automatically saves the progress. That being said, the level codes are merely a gimmick to make it feel a little bit more like the Amiga.
Secondly, the codes given by Lemmini work for Lemmini. You couldn't expect more. I must admit, I orginally thought about writing my own algorithm with my very own level codes. I gave this up mainly for having more flexibility.
Thirdly, codes based on the original algorithm wouldn't necessarily work for a custom level pack. I mean even if I could read the info "game 3" from it, what would than mean for a custom level pack? Whereas the way it is now, the creator of the level pack can create his/her own codes and they will work fine. They can even have some meaning to hint which level they belong to.


The player management is a little buggy, I agree. Will be fixed in the next release I hope.


Will be fixed


Not sure if I completely got this, but if you mean that a builder (or any other turns) when you assign a blocker directly next to it, this is probably true. Indeed, when creating a blocker, the blocker mask is copied into the mask array. For a lemming which is inside the blocker mask at this moment it "feels" as if he bumped into the mask, since for the last step it wasn't there and now it is.
Dunno, maybe I should have two different blocker masks: one to the left of the blocker and one to the right. Then every lemming could check if it the blocker is left or right to it. I'll think about it.


Indeed, the whole steel digging/bashing/mining issue is a little trickier than it seems. My hope was I could find a "generic" solution without too much special treatment. However, no matter where I set the threshhold for number of steel pixels in the mask before stop mining/bashing/digging, the are always corner cases where it looks shitty.


Yeah, some states don't have names as the one which indicates a lemming being killed by a trap (but not part of the trap animation). Internally I call these guys "trapped". I think I should exclude all nameless lemming from the "under the cursor" routine since they can't be assigned a skill anyway.

No, I don't want to steal from Cheapo. Indeed, I painted a minimalistic jumper animation and put the jumper skill in in the last half hour. So it will be in 0.60. Looks a bit silly though.

Hm, if you keep picking on this, I'll go back to my previous solution. However I'm sure someone will complain ;)


That being said, I notice that the builder actually gets stuck in Mayhem 2 as well. So there is something to fix anyway.

No, there really is a jumper animation. (And 0xdeadbeef: you wouldn't be stealing from Cheapo, since the Cheapo lemmings animations are themselves stolen from the original game......  :P )

Well, just for the record, I'm with ccexplore on this one.  ;)

About the steel-destruction-by-explosion: I would leave it as it as in the original game by default.
A few of this detail-things could be made optional: instead of making a choice leave the decision to the end-user.
This item is definitely an "option" to me :).
That's because it is a bug.
But i've played levels where lemmings just did NOT die falling from a certain distance because of the bug.
I do not remember which level but a standard one.
Also the miner has a "feature" of mining through steel sometimes (also in the original :)).
Is this known?

Nightly update - let's see what I've broken this time :-?

Fixes/Changes 0.59 -> 0.60
#  Replaced menu "New Player" with "Manage Players" which also allows deleting players
#  Fixed wrong background color of Minimap in 16bit color mode.
#  Current player's progress is stored when changing player or entering the "manage players" menu
#  Introduced "Jumper". Looks a bit silly though.
#  Re-enabled old (exact) explosion masking: only destructible parts are erased.
#  Keys 1..8 can be used to select skills
#  Changed inner workings of blocker mask: it's split into left and right part. A lemming moving right
   will not be influenced by a right side blocker mask and a lemming moving left won't be influenced
   by a left side blocker mask. This should fix lemmings turning around if a blocker was assigned
   next to them though they already passed him.
#  Lemmings with no name (e.g. drowning, killed by a trap, splatting etc.) will not be considered
   for cursor selection. Thus also no "empty" names can appear any more.
#  Fixed name of Fun 25 ("Lemmings Lemmings everywhere")

@ccexplore: I have no example of deep tunnels through steel :). The "mining" a referred to was the "lack of masking" you described.

I don't think there's anything wrong with configurable lemmings, especcially because it is a remake. It's just a some checkbox for highly advanced users: "Apply Steel Explosions Native Bug To Stay Compatible".
Well it was just an idea.

Hm, interesting. I have no idea how I possibly could have messed the sound effects in 0.60 as I didn't change a single line there. And the jumper was appended to the end resource wise...
Sure this didn't happen in 0.59? Anyway, I will have a look at it when I come home from work.

Obvious problem: that would cause havoc with user-created levels. Either you'd have to always check that your level was compatible with every setting (I say every, not both, because I guess there are other bugs that could also be turned on or off), or else the player would have an unfair disadvantage if he was playing a level with the wrong settings and didn't realise different ones were meant to be used. If this is meant to be distributed more widely than just this board, that's intolerable.

Oh, and please don't say "just an idea". Sorry, it's a bit of a pet peeve, but I hate that phrase. Explaining why I disagree with your idea -- even complete disagreement as in this case -- is not a personal attack and you don't need to get defensive about it.

If the option was implemented, it would be best to leave the option up to the level designer, not the user.

Well..... better, yes, but not best. There would still be two severe problems: we'd have to decide which option to implement for the standard Lemmings and ONML levels, and whichever we chose would almost certainly be used as standard by level designers (how many levels does this really affect?). Worse, users from outside this board who don't know about all the bug discussion would be thrown by levels using non-standard game rules -- it would be radically unfair.

The best (and in my opinion the only) option is to do as ccexplore suggested, fix the bug as though it had never existed.

I have, and I get the same as you, the lemmings exiting make the "creak" sound effect.

Incidentally, the best thing about the instant release rate change facility is that now I can play with sound on without fear of crashing! Just so long as I don't nuke.....

Fixes/changes 0.60/0.61
#  Fixed stencil after extending it from short to int (again). Partly short access led
   to wrong sounds being played, wrong trap behaviour and so on.
#  Eliminated trailing spaces on Fun 25 ("Lemmings Lemmings everywhere")

And about the explosion mask issue: it would be no problem to make this a level pack configuration like the safe fall distance. But to be honest, I don't like the idea.

I must admit I don't see the point. You get a levelcode and if you enter this levelcode, you will get to the according level, though most people wil never use this anyway. As a bonus, the levelcode also works for the Amiga version.
To come up with and type in 200+ levelcodes just because someone might be disappointed if he entered an Amiga levelcode which doesn't work on Lemmini, seems a lot of effort for a little benefit ... especially since the situation is not improved at all: still the Amiga levelcodes wouldn't work.
And again: in a "closed" game which only supports a given set of levels, algorithmic level codes are a fine thing. However lemmini is supposed to be extended by custom level packs which are treated exactly like Lemmings and ONML. IMHO the best way to do this is to have fixed level codes which are part of the levelpack configuration.

If you thought that would be too much work, I could easily write a program to generate 216 random ten-letter codes. There would be a benefit: the player who wasn't part of this board and hadn't followed this discussion wouldn't expect his Amiga codes to work after seeing that the first couple work perfectly and guessing that the rest will too.

I really don't see the big deal about the level codes anyway.  You'd have to use a some type of hash to avoid having multiple levels in different level sets having the same password.  And if the password is hard-coded into the level files, what will happen if two designers choose the same password?

Edit: I can ccexplore's point about having seeded passwords, but IMHO it's a lot of work that won't be extremely useful.  Perhaps when the game is nearly finished this could be added, but it doesn't seem important right now.

Edit, again: If the player is really that desperate to get to the level that he was on, he can hack the player file.  :P

The situation is even worde for algorithmic levelcodes, so I don't see the problem here. If two levelpacks use the same password, then  the levelpack has to be known. For algorithmic levelcodes, the levelpack has ALWAYS to be known to have the seed. Also it's obvious that different seeds might create the same password. The situation is FAR easier to control with fixed codes.


The implementation itself won't be much work. The trouble created from it will however.

Just a question.... why the heck are we whining about PASSWORDS!?  It's not like it's a huge deal, is it?

...hehe... sorry, it seemed to me that some people were complaining that Lemmini didn't have its own passwords or something.

And a Q:  I hadn't played Lemmini for a while, so the Java Webstart upgraded me from 0.51 to 0.61.  This caused me to lose my saved progress, so I had to start all over.  Is is possible to prevent this frustration?

That's a flaw that I already pointed out and has been fixed in between those two versions. So now you've got 0.61 it won't happen again.

The usual nightly update...

Fixes/changes 0.61 -> 0.62
#  Forced resource extraction (needed for the fun 25 title change fixed in 0.61)
#  Fixed drawing order: animated objects that don't have the "no overdraw" flag set
   will not be drawn behind the terrain any more.
   Should fix ccexplore's #58-#60
#  Drawing order also fixed in minimap (all objects were drawn behind the terrain)
#  Walkers, Jumpers and Builders can't leave the level to the top any more
#  Increased number of pixels that have to be free above a builder to continue building.
#  Reworked jumper animation. Looks a little less silly in my eyes. Yet maybe not in yours.

Just one small point..... version numbers don't carry, so what you've called 0.60, 0.61 and 0.62 should really be 0.5.10, 0.5.11 and 0.5.12. I mention this because with all these little updates, and with ccexplore finding so many bugs, you might run yourself out of available version numbers very soon.....

Does that mean you'll be getting on with the MIDIs now?  ;)

As for me, I should have my essay finished by tomorrow -- then a break at last, and I will certainly use at least some of the time to keep playing Lemmini and see what happens!

[EDIT: ooh, my 1111th post! Nice!]

Not bug but wish:
I always played DOS lemmings and used to its function keys.
F1 = RR-, F2 = RR+, F3 = Climber... F10 = Digger, F11 = Pause, F12 = Nuke
Yet in WinLemmings the keys are different (I can't get used to them)
So I badly need keyboard input. And there would be at least one happy user if these were configurable.

Fixes/changes 0.62 -> 0.63
#  Jumpers can now be assigned at least "additional" skills like Floater and Climber
#  Fixed a bug that would let Jumpers survive a nuke. Now also splatters, drowners etc.
   can be nuked - I thought they could, but there was a bug obscured by the fact that
   these types die anyway.
#  Keys F1-F8 can be used in addition to 1..8 for skill assignment.

I noticed it myself. I should really get some sleep, but at least the weekend is near.
Will be in the next release I guess.

BTW: are you sure about the Jumper behaviour? In Mary Poppins's land it's now pretty hard to set a blocker on that steep ramp. It's not in the windows version since it has no jumpers. But also in the Amiga version, the stoppers can be easily placed on the ramp.
Lemmini now definitely behaves differently from the Amiga version due to the jumpers and their limited skill assignment.

OK, getting on with the new version. Four things, two of which are ones I've brought up before but haven't been dealt with yet.

(1) Stretching bridges. I got it to work once so it's possible, but it's very very hard, much more so than on the real game. This really needs to be amended; it's an immensely useful trick. (And the player used to other versions of Lemmings will get really exasperated with levels like Crazy 4 where it's so hard to break the habit and deliberately not try to stretch.....)

(2) Bashing steel. At the moment you can't bash when any steel is in the basher's way. This makes 7 Crazy well-nigh impossible. On the real game you can still bash when there are just a couple of pixels of steel at the top of the basher's field.

(3) I know you dealt with the level graphics flickering, but the cursor still flickers  :(  It's most annoying.

(4) Not sure about this one, but I notice that when you build into a wall and make the builder bash, if after bashing there is a gap of just one hi-res pixel, the basher will continue into the wall and not fall through the gap. It's such ages since I played the Mac version, but I don't remember this being possible.

I did not really test the stuff but the following things I noticed while playing:

1) I agree with A. Stretching bridges is difficult but not very difficult.
I only have comparison with DOS in which I think it was quite easy to do. Maybe it is just the resolution (?)
2) In my opinion too many "cursors". Difficult too focus on the lemming you want to select. I better like one cursor changing. I really can't see the lemming under the rectangle and +.

I tried to fix it, but it didn't work out well. The builder animation clearly defines that the builder stands on the corner of the step.
I can't move the lemming back since this would be visible, I can't move the step back since then steps on slopes wouldn't work and I can't enlarge the step, since this would create much broader steps in contradiction to the original.
Really dunno how to solve this.


As I said earlier in this thread, the approach to steel detection is generic at the moment, but obviously it doesn't work well. I will rework this one way or the other, but I'm not really sure yet how.


If you mean the main cursor: I don't even draw it (I just replace the system cursor with the lemming cursor for the lemmini window) and it doesn't flicker on any system I tried (including an 800MHz Athlon XP with 256MB RAM and a very slow video card). If the cursor flickers under Win98, it's abug in the Java runtime environment and I don't have a chance to fix this.
I could only go back to a software drawn cursor but it's sluggish compared to the system cursor.



Though I don't really get the problem described, I agree that lemmings can walk through one hires pixel walls and won't fall through one hires pixel gaps. It's somehting I built in long ago because at that time I thought that one pixel gaps/walls could only be artifacts. I think I will remove this to avoid confusion.

It's not a huge deal and it doesn't affect gameplay; if you can't fix it then don't worry about it. No point mucking up the game to make it compatible with my antiquated machine  :P

Btw, I'd like to add a voice in support of what Eric said about the cursor -- having both the box and the cross is also annoying.

Lemmini Retro

I'm not sure what the point is, but this replaces the hi-res styles with the low-res ones.  This does cause problems with jumpers, I think.

Nice, then I'm not the only one complaining. ;) I'd still suggest the arrow (like in Cheapo) as an alternative option.
As for the bridge stretching, it also seems more difficult to me, I dunno however. There seems to be still only one frame to make a bridge overlap of 2px, but due to the doubled frame rate, it is twice as hard as for the DOS version to hit the right frame? Additionally, for the DOS version it is possible to stretch a bridge with only one px overlap when building to the right. However, this is only possible for building to the right, so I don't think that it would be neccessary for Lemmini.

The last four Tame levels are still missing, just in case you'd like to add them, I uploaded them: http://207.58.177.175/~geoo89/lemmings/Tame4l.zip

Some old bugs I'd like to remind you of, and a few new ones I found:
- miners and bashers still don't turn around when hitting steel (they should do even if they'd be able to continue walking)
- it is possible to build through a 1 lo-res pixel thick wall, other walkers however cannot pass such a wall
- bashers can (as Ahribar mentioned for the hi-res pixel) bash past an 1 lo-res pixel gap below them
- ccexplore once mentioned that the palette for the objects in the special graphics levels might be changed
- when a digger breaks through, he immediately starts to fall while still doing the digger animation, at this point it's impossible to assign a skill to him. This behaviour causes that the digger can continue digging when he fell low enough, i.e. he can pass a 2 lo-res px gap without stopping digging. The other thing is, for a digger who's on thin air i.e. has no terrain below him, in the DOS version he does another complete digger animation before falling, and during this period you can still assign any skill to him.
- the horns for the fire exit are still missing. If anyone wants, I could try to design those.

Chances are low for this as I like it the way it is and so do my "personal" beta testers.
Additionally, making it an option is not as easy as it sounds.
The main cursor will however be a little smaller ("thinner") in the next version.


No, since the walker also walks only one hires pixel per frame.



I will have a look


It's not necessary to repeat stuff like this, especially if it was repeated before. Anyway I commented on all of these topics before and usually ignore repeated postings. This does not mean however, that I forgot.
BTW: The one pixel hires wall/gap check was not really bug by the way, more like a feature (though maybe misguided). Nevertheless I already changed it yesterday in my development build, though I fear that 1 hires pixel accuracy might add more problems than it solves. Then again, maybe I'm wrong.


I'm not sure if you're talking about Lemmini or DOS lemmings in the first two sentences.
In Lemmini, a digger can fall 1-7 hires pixels without stopping digging at the moment. It will fall (up to) 3 hires pixels per frame as a faller, but is still a digger. After falling down 8 hires pixels, it will be converted into a faller.
Also, as a falling digger is still a digger, you can assign any skill to it that you could it it was not falling (makes no difference).

You mean that also for this case when the lemming is standing on the edge there are two frames he'll build the the same place? I had thought this might be a special case and one of the frames is already the faller frame...well, then the seemingly increased difficulty is probably just illusion.

The first two sentences are referring to Lemmini.
The behaviour is quite a bit strange to me, but if it is intended to be a feature, well, it also has indeed advantages.
Although a skills can be assigned to the digger after the 'fall', it cannot be done during the fall, that's what I meant with my sentence before.

Ok, to be more precise: if a builder built right, built the last step, converted to a shrugger and then to a walker, the walker (in his 1st animation frame) will stand just on the edge of the step, meaning his rightmost "toe" pixel is still on the step. The "foot position" (used for collision detection with the mask) of the walker however is three pixels left of this. Since a walker moves 1 hires pixel per frame, he will be on the step for at least 4 frames. In Lores, it would be only two pixels and thus only two frames, which however would last twice as long. I even tried to increase the step by 4 more pixels. Even then stretching a bridge was more or less a matter of luck.


Well as I said, a digger is still a digger, even if it's falling. So also the skill assignment ist the same. However, some skills can't be assigned mid-air, especially digger and miner (but you couldn't assign a digger a digger anyway). It's a little hard to test, but at least there is no code inside Lemmini that would hinder a falling digger to be assigned climber, floater, bomber or whatever.
If this behaviour is correct, is another question of course. The idea behind is, that a digger will not move down, but he will actually dig a hole and fall in. So the mask determines how deep he falls.

Ok, to be more precise: if a builder built right, built the last step, converted to a shrugger and then to a walker, the walker (in his 1st animation frame) will stand just on the edge of the step, meaning his rightmost "toe" pixel is still on the step. The "foot position" (used for collision detection with the mask) of the walker however is three pixels left of this. Since a walker moves 1 hires pixel per frame, he will be on the step for at least 4 frames. In Lores, it would be only two pixels and thus only two frames, which however would last twice as long. I even tried to increase the step by 4 more pixels. Even then stretching a bridge was more or less a matter of luck.
I might add that 4 hires frames or two lores frames are just 132ms, so this is not a very long period.


Well as I said, a digger is still a digger, even if it's falling. So also the skill assignment ist the same. However, some skills can't be assigned mid-air, especially digger and miner (but you couldn't assign a digger a digger anyway). It's a little hard to test, but at least there is no code inside Lemmini that would hinder a falling digger to be assigned climber, floater, bomber or whatever.
If this behaviour is correct, is another question of course. The idea behind is, that a digger will not move down, but he will actually dig a hole and fall in. So the mask determines how deep he falls.

Could you clarify this? The digger in lemmini will convert into a faller as soon as there is "nothig more to dig" - which means in the line below him, there are at least 12 (or something) holes deep enough to convert him into a faller (>= 8 hires pixesl deep).
My "dig a hole and fall in" algorithm just kicks in if the hole is not deep enough - either because it is the one dug by the digger or if there is a small horizontal "channel" due to level design.
I'm not sure about how this was implemented in the DOS/Amiga versions, but I guess, they just applied the mask and moved the lemming down. This would mean that a gap only on pixel in height would stop the digger. Is this the way it should be?

Since I'd move down 2 hires pixels at once, this means that only 1 hires pixel gaps can be detected that are by chance on the same as the digger foot's Y coordinate. If the gap is 1 pixel off, the digger will jump over it. Else the digger would have to go down pixel by pixel and each time check for a gap.

Lots of changes this time with lots of chances to have messed something. Give it a try.

Fixes/changes 0.63 -> 0.64
#  Reworked skill assignment. Assignment request is buffered and applied at the beginning
   and if that fails at the end of the lemmings movement update. This should make skill assignment
   generally better to apply.
#  Reworked main loop. Usually missing animation updates are caught up if possible to maintain
   game time. However, under special cirumstances (high system load for a second or so) this led to a
   kind of forced "fast forward" mode until the game time was synchronous with the real time again. To avoid this,
   any time difference larger than 5 frames (165ms) will now be skipped instead of being caught up.
#  Reworked function key assignment: F1 to decrease RR, F2 to increase RR, F3 - F10 for the skill assignments,
   F11 to pause and F12 to nuke.
#  Removed check that considered one pixel walls/gaps as artifacts. So Lemmings are reflected by one hires pixel
   walls and fall through one pixel gaps now.
#  Made main cursor a little thinner (there are issues though, since obviously, the cursor is scaled up to some system default)
#  Put "load level" feature back in to load a single level file (INI or LVL). It's still a little unstable though (quick hack),
   so use it on your own risk ;)
#  Added the 4 "Tame" levels 17..20 missing in the windows version
#  Reworked break conditions for miners, diggers and bashers when hitting an indestructible pixel (e.g. steel).
   Now they all use an additional mask which should improve the behaviour.
#  When a digger, miner or basher hits steel, the "chink" sound is played.
#  Changed the way the digger works. Now he will move 2 pixels down at each animation cycle instead of falling down
   digging. Now a digger can be stopped by a vertical gap just one hires pixel in height. However, as he moves down 2 hires
   pixels at once, he will "jump" over such a line with a probability of 50%. Only if the gap is 2 hires pixels
   high, he will be stopped for sure.  
#  Added the dreaded horns to the fire exit. May need some rework, but for the start, it's ok.
#  Renamed "squasher" to "marble"


BTW: since I changed resouces quite a little, it might be a good idea to remove your resource directories from time to time to force an extraction and have a "clean" resource directory again. Just don't remove players.ini and the players directory if you want to keep your player stats.

Lemini seems quite cool. Keep it up Oxdeadbeef. Some good stuff there. Can you end up putting your own levels into the game?

Excellent!

* The skill selection seems much, much smoother now.

* The thinner cursor looks a lot nicer, as well as making the flickering much less annoying -- indeed, I hardly noticed it.

* For some reason the cursor sometimes displays walker lemmings as FALLER even when there are no fallers around.

You can already load and play own LVL files. Additionally, it's possible to create whole own levelpacks and styles. Then again, there's no documentation on this yet. However, if you look closely at the ini files in the resource directory, you'll see that this is no rocket science. The only part missing right now is a full featured LVL->INI converter (Lemmini loads levels from INI files which are more or less ASCII versions of LVL files). If there's demand, I could quickly hack a converter together. Anyway, there's a trick right now: if you load a LVL file, Lemmini will actually write a file "test.ini" in you resource directory and then load this file as level.
If you copy this file and rename it, you have your LVL converted to an INI file.

Ah, at least something ;)


Maybe I'll go back to software cursor anyway. A drawback of the system cursor (apart from flickering under win98) is, that its size is changed to some system default. That makes e.g. the yellow dot in the middle larger than it should be. Then again, the system cursor moves much smoother as it's not bound to the 33ms screen update of the rest of the game. Tricky...


Actually, this happens ALWAYS, not only sometimes  :-[
I simply forgot to swap the strings as I cleaned lemming.ini. Will be fixed in the next update.

While I'm at it: what are your expectations regarding action replay? How do you like the action replay in WinLemm? Should action replay in lemmini be similar or different?

Key points:
- Should the cursor/playfield move around or do you want to control the cursor/scrolling?
- Should this be just a passive replay to show how you did this level or do you want to take over the action like in WinLemm?
- Is storing/loading stored action replays a needed feature?

Any other ideas/comments about action replay?

Yeah, action replay should be identical to that of Winlemm: A replay of all your moves, but should the player come in and make a change not done before commence the end of the replay.

Oh, and storing/loading action replays would be pretty nifty, but by no means necessary.

Sure. There are even fun levels which at least are usually finished by nuking.


Sure. I never actually planned to record plause/FF anyway since the replay needs to be animation frame based to be sure that it's 100% accurate.


Hm, for the very first go, I think the change will not be marked, just like it WinLemm. For the future, it might be nice to have some gfx effect on any Lemming that's assigned a skill (like changing colour or having a box around it or whatever). You can see which skill by either checking the lemming with the cursor or by looking which icon is pressed and which counter decreased.

Just a thought: maybe saved ARs should move the cursor around but shouldn't be interruptable?  and unsaved should be the opposite?
It would limit the user and be more work to program, so I'm not sure that it's really a good idea.

I don't really see why this is so important. If the player sees a lemming climb or float, he'll know he's meant to make it a climber or floater at some point prior to that. There isn't a specific point when it's essential to assign the skill, unlike with the other skills.

Tried to run under Linux (Ubuntu Dapper). Case-sensitive file/directories are somewhat ignored, so none of the levels will convert from LVL to INI if your export directory contains a capital letter. (interestingly, the graphics still convert without a hitch)

After picking a path that's entirely lowercase, the conversion process works fine for a while, but keels over when trying to patch the first level ("CRC of source differs from that in patch header"). One guess is that during the LVL2INI conversion process, the file I/O methods in UNIX Java translate "\n" into just a linefeed character (i.e. the civilized way), whereas under Windows Java it may instead give a carriage-return + linefeed (2 characters, a nasty relic from DOS and the norm for Windows text editing).

If this were the case it would certainly throw off the binary diff. Of course this is all guesswork, I haven't got a working Windows machine to litmus-test the theory on. Just for fun, here's my extracted copy of lvl1000.ini:
# LVL extracted by Lemmini # lvl1000.lvl
releaseRate = 1
numLemmings = 80
numToRecue = 80
timeLimit = 3
numClimbers = 0
numFloaters = 0
numBombers = 0
numBlockers = 0
numBuilders = 12
numBashers = 0
numMiners = 0
numDiggers = 0
xPos = 1216
style = brick

# Objects
# id, xpos, ypos, paint mode (), upside down (0,1)
# paint modes: 8=VIS_ON_TERRAIN, 4=NO_OVERWRITE, 0=FULL (only one value possible)
object_0 = 1, 1344, 128, 4, 0
object_1 = 1, 1408, 64, 4, 0
object_2 = 1, 1472, 0, 4, 0
object_3 = 0, 1376, 232, 4, 0
object_4 = 9, 1376, 264, 4, 0

# Terrain
# id, xpos, ypos, modifier
# modifier: 8=NO_OVERWRITE, 4=UPSIDE_DOWN, 2=REMOVE (combining allowed, 0=FULL)
terrain_0 = 24, 1376, 198, 0
terrain_1 = 24, 1440, 198, 0
terrain_2 = 24, 1436, 128, 0
terrain_3 = 24, 1500, 128, 0
terrain_4 = 24, 1500, 64, 0
terrain_5 = 24, 1564, 64, 0
terrain_6 = 31, 1508, 262, 0
terrain_7 = 31, 1568, 198, 0
terrain_8 = 31, 1632, 124, 0
terrain_9 = 30, 1508, 278, 0
terrain_10 = 30, 1508, 310, 0
terrain_11 = 30, 1568, 214, 0
terrain_12 = 30, 1568, 246, 0
terrain_13 = 30, 1568, 278, 0
terrain_14 = 30, 1568, 310, 0
terrain_15 = 30, 1632, 140, 0
terrain_16 = 30, 1632, 172, 0
terrain_17 = 30, 1632, 204, 0
terrain_18 = 30, 1632, 236, 0
terrain_19 = 30, 1632, 268, 0
terrain_20 = 30, 1632, 300, 0
terrain_21 = 30, 1696, 288, 0
terrain_22 = 30, 1696, 256, 0
terrain_23 = 30, 1696, 224, 0
terrain_24 = 30, 1696, 192, 0
terrain_25 = 30, 1696, 160, 0
terrain_26 = 30, 1696, 128, 0
terrain_27 = 30, 1696, 96, 0
terrain_28 = 30, 1696, 64, 0
terrain_29 = 30, 1696, 32, 0
terrain_30 = 30, 1696, 0, 0
terrain_31 = 2, 1432, 302, 0
terrain_32 = 2, 1368, 302, 0
terrain_33 = 2, 1304, 302, 0
terrain_34 = 19, 1714, 0, 4
terrain_35 = 19, 1778, 0, 4
terrain_36 = 20, 1834, -8, 4
terrain_37 = 20, 1834, 8, 4
terrain_38 = 20, 1850, -8, 4
terrain_39 = 19, 1714, 288, 8
terrain_40 = 19, 1778, 288, 8
terrain_41 = 20, 1834, 302, 8
terrain_42 = 20, 1834, 286, 8
terrain_43 = 20, 1850, 302, 8
terrain_44 = 19, 1240, 288, 8
terrain_45 = 19, 1208, 288, 8
terrain_46 = 19, 1208, 0, 12
terrain_47 = 19, 1272, 0, 12
terrain_48 = 19, 1336, 0, 12
terrain_49 = 19, 1368, 0, 4
terrain_50 = 20, 1424, 8, 4
terrain_51 = 20, 1440, -8, 4
terrain_52 = 20, 1424, -8, 4
terrain_53 = 20, 1296, 286, 0
terrain_54 = 20, 1312, 302, 0
terrain_55 = 20, 1296, 302, 0
terrain_56 = 42, 1216, 0, 0
terrain_57 = 42, 1248, 0, 0
terrain_58 = 35, 1234, 32, 0
terrain_59 = 42, 1250, 160, 0
terrain_60 = 42, 1218, 160, 0
terrain_61 = 42, 1232, 192, 0
terrain_62 = 42, 1752, 288, 0
terrain_63 = 42, 1784, 288, 0
terrain_64 = 35, 1768, 160, 0
terrain_65 = 42, 1752, 128, 0
terrain_66 = 42, 1784, 128, 0
terrain_67 = 42, 1766, 96, 0
terrain_68 = 20, 1200, 286, 0
terrain_69 = 20, 1200, 302, 0
terrain_70 = 20, 1200, 8, 12
terrain_71 = 20, 1198, -6, 12

#Steel
# id, xpos, ypos, width, height

#Name
name =  It's all a matter of timing

Hm, not too long ago I booted a Knoppix CD and played lemmini without any problem. Then again, probably I chose/created a lower case path back then. So probably I convert a whole path to lower case where I should only convert the file name...
The linefeed theory however is interesting but wrong, since all text files are created the "unix" way even on my windows machine.
I will try again with Knoppix, if the current version still works there.

Flipper sounds like a dolphin (or a part of one  ::) )
And I don't think it would be a good iedea to change a Walker's name to Jumper each time he jumps up. This would cause the name to flicker like hell in certain levels.

But back to Linux:
I had a look at the extraction code and found two placed where I converted the whole path to lower case where I wanted to convert only the file name. I tried this version with Knoppix and it worked with a resource (target) directory containing capital letters.
The source (WINLEMM) directory always contained capital letters, but as it's an NTFS partition, I'm not sure if capital letters are preserved.
Anyway, give the new version a try. It also fixes two other errors, one of them found while testing under Linux.

Fixes/changes 0.64 -> 0.65
#  Swapped Walker/Faller strings. Was overlooked when cleaning up lemming.ini for 0.64
#  Lemmini wouldn't work on Linux if the resource or game directory used capital letters. Should be fixed now.
#  Width of yellow frame in MiniMap was derived from Screen resolution, not from window's width. Fixed.

I think you got me wrong here: I mounted my NTFS-Windows Partition as drive under Linux (Knoppix) and I'm not sure if Linux is case sensitive for NTFS-Partitions, cause as you said, NTFS itself is not (really) case sensitive. Since I used the WINLEMM directory from my (Windows) NTFS partition as extraction source under Linux, I'm not completely sure if that proves that also the extraction source paths with uppercase and lowercase characters work under Linux.

FAT32 partitions mounted under Linux have ambiguous case recognition (i.e. 'cat /mnt/win-c/Games/ReadMe.txt' and 'cat /mnt/win-c/games/rEaDmE.TXT' will find the same file), presumably NTFS support does the same thing.

I'll give the new build a try later on.

I had an idea for solving the controversy over level codes. Instead of going to the trouble of generating new codes for every level, how about just removing the first (or the last, as you prefer) letter of each one?

One feature I definitely hope Lemmini will support is minutes-and-seconds time limits for user-created levels; they open up so many new possibilities!  ;)

When I'll have finished the replay stuff (which may take some time, since I was lazy lately), I'll come back to the level code. I might even consider auto generated codes again. No promises, but let's just wait and see.



Indeed, there was no support yet in 0.65, though it was long prepared and on my todo list. But it's in 0.66 now.


Regarding ccexplore's bugs: should be all fixed in 0.66 I hope.

Fixes/changes 0.65 -> 0.66
#  Support for second exact level time. Keyword is "timeLimitSeconds" instead of "timeLimit". E.g. "timeLimitSeconds = 356" for 5 minutes and 56 seconds
#  Fixed gap in pillar of LVL0025.ini (Taxing 4)
#  Fixed one pixel shift in last animation frame of entries of ROCK and PILLAR


BTW: I moved the Lemmini page to http://lemmini.de. Since I updated the JNLP on the original site and also set a redirector to the site, I hope this should not create any problems.

The JAR link is broken. It's easy enough to guess what the link should be, though.

Oops, I will have a look when I'm at home. This should have been a relative link anyway...
The JNLP should work however.

Just some random weirdness, but I noticed 0xdeadbeef on Slashdot.  Odd.  :P

Just another guy who thought this would be a funny name ;)
But if you'll put it this way, there's even an article in wikipedia about "me": http://en.wikipedia.org/wiki/0xdeadbeef

Talking about strange web findings, I recently found this: http://slashnet.org/channels/bullshit/
It's a little offending to see my site as title in a channel named "bullshit".
And I'd really like to know what "Join the cult! It's all bullshit! (FREE PUNCH AND PIE)" would mean in this context  :-/


Anyway, Replay is more or less finished. I'd like to do some more testing and maybe need to rework some "visual" stuff.
But I guess there'll be a new version out by tomorrow evening.

Here we are:

Fixes/changes 0.66 -> 0.67
#  Added replay functionality
#  Introduced text buttons instead of left/right clicking in debriefing screen
#  Set Focus explicitly on graphics pane. Maybe helps with flicker on low end machines?
#  Fixed some (more) bugs in LVL1016 (doubled entry, wrong water object)
  Who made this level anyway? It containes at least three bugs ;)
#  Split loading and painting levels so when restart/replay is used, the level is not loaded completely
  but (more or less) only redrawn
#  Reset Basher to Walker conditions to what it was before 0.64


Please test the replay functionality. Should be pretty much complete, but I'm sure you'll find something.

When 100 Lemmings are out, the time gets knocked off the screen.


After successfully completing a loaded level.  I've only tried loading Kindness That Can Kill, though.

I fixed the crash with a "silent" update (same version number).

About the font issue: it's already on my list since quite some time. Problem is that I created the font a little too broad. There are also some levels which have too long names for the briefing screen. Will be fixed soon, but rescaling the font is a stupid work and up to now I never could decide to do it.

Fixes/changes 0.67 -> 0.68
#  Lemmini crashed when a single level loaded with "Load Level" was won
   (indeed fixed as silent update in 0.67)
#  Reduced font width from 20 to 18 pixels. Looks more like the Amiga
   and fixes display issues when text is too long.
#  Added colon (":") to font (though it was missing on the Amiga as well)
#  Fixed wrong briefing display after changing levels (introduced in 0.67)
#  Key "S" does a full level screen shot ("level.png") in resource directory

Not much going on here lately. Are you already tired of it  :o ?
Some feedback on replay and the algorithmic level codes would be nice...

Fixes/changes 0.68 -> 0.69
#  Replaced fixed levelcodes with codes calculated by the original algorithm
#  removed levels/levels.ini. Now subfolders of levels will be assumed to be
   levelpacks (alphabetical order)
   Note: in this context, also the level folder names have changed. So it's
   time for deleting the resource directory to clean things up.
#  Extended Extractor/Patcher to support files of same name in different
   folders (hope this doesn't destoy Linux compatibility)
#  Added fade in/out effects like on the Amiga
#  Lemmings walking down a steep slope (at least 45° or more) accumulated
   "fallen" pixels which were never reset. This led to splashing if they
   converted to fallers after the slope even if the fall distance was
   minimal -> fixed
#  Levels (*.ini) loaded via "Load level" are checked for start string "# LVL"

I am not tired of it:)
Lemmini is already better than the WinLemmings original I think!
I am very happy with the function keys.

Found another bug I think.
This level I created with my editor. When you build in front of the trap, the stairs are not visible.
(Between dotted lines = ini file)

--------------------------------------
# LVL ericbrick.ini
# Edited with Alpha version of Lemmings Level Editor
releaseRate = 1
numLemmings = 50
numToRecue = 48
timeLimit = 4
numClimbers = 1
numFloaters = 1
numBombers = 1
numBlockers = 0
numBuilders = 8
numBashers = 2
numMiners = 1
numDiggers = 1
xPos = 420
style = brick

# Objects
# id, xpos, ypos, paint mode (), upside down (0,1)
# paint modes: 8=VIS_ON_TERRAIN, 4=NO_OVERWRITE, 0=FULL (only one value possible)
object_0 = 9, 964, 192, 4, 0
object_1 = 0, 964, 160, 4, 0
object_2 = 7, 786, 165, 0, 0
object_3 = 1, 496, 85, 0, 0

# Terrain
# id, xpos, ypos, modifier
# modifier: 8=NO_OVERWRITE, 4=UPSIDE_DOWN, 2=REMOVE (combining allowed, 0=FULL)
terrain_0 = 5, 946, 229, 0
terrain_1 = 5, 1010, 229, 0
terrain_2 = 5, 1074, 229, 0
terrain_3 = 5, 1138, 229, 0
terrain_4 = 5, 1202, 229, 0
terrain_5 = 5, 818, 229, 0
terrain_6 = 10, 434, 225, 0
terrain_7 = 35, 750, 101, 0
terrain_8 = 35, 750, -27, 0
terrain_9 = 5, 1116, 48, 0
terrain_10 = 5, 1180, 48, 0
terrain_11 = 5, 1244, 48, 0
terrain_12 = 5, 562, 229, 0
terrain_13 = 5, 1266, 229, 0
terrain_14 = 5, 498, 229, 0
terrain_15 = 5, 626, 229, 0
terrain_16 = 5, 690, 229, 0
terrain_17 = 5, 754, 229, 0
terrain_18 = 7, 746, 48, 8
terrain_19 = 7, 810, 48, 0
terrain_20 = 7, 874, 48, 0
terrain_21 = 7, 938, 48, 0
terrain_22 = 7, 1002, 48, 0
terrain_23 = 7, 1066, 48, 0
terrain_24 = 7, 691, 48, 8

#Steel
# id, xpos, ypos, width, height

#Name
name = When
-------------------------------------

Not sure if it's a bug. When I have the "no overwrite" flag set for the trap, then it's ok.

Indeed. However, I can't 100% tell how WinLemm would behave here. For Lemmini, the stairs are terrain, so if the NO_OVERWRITE is not set for an object like traps, exits and so on, it will overwrite terrain (and stairs).

Ok. As soon as I can save LVL files (sigh...) I can test it with WinLemm.

In Lemmini's level format, shouldn't

numToRecue

be

numToRescue

?

Very sharp :)
I just copied the "mistake" into the code of my level-editor.
Indeed in Lemmini ini files we see "numToRecue"

Ah, damned typos  :-[

I will kick out a new release for this later today.

Then again I wonder if anybody want to comment on new features as replay :-?

Ok, should be fixed now in 0.70. Give it a try.

Replay looks ok on the first sight. (I never used it in WinLemmings, so I am not an expert)

I was replaying a saved replay, which was not a standard level. I got this message:

"The resource is missing.
Please restart to extract all resources."

On the next restart lemmini is indeed extracting again all the resources.

Looks ok to me at first sight (I am surely not a replay-expert).

However: when opening a saved replay of a custom (not standard) level I get this message:

"The resource is missing.
Please restart to extract all resources"

And on restart Lemmini extracts all files again.

Maybe some extra information should be stored in the rpl-file

EricLang when is that level editor going to be finished?

Have patience... I will start a new thread with an alpha version within a few weeks.

Hm, most probably I'll disable the possibility to save replay for levels started via "load level". For custom level packs, it should be ok though.

That should be good. Will this editor show extended graphics while editing a level?

2 bugs:

• Upside-down objects are not drawn
• Upside-down trapdoors deploy Lemmings

Actually there is more.
1) Some objects upside down (entrances) are not drawn.
2) Other objects upside down (water, exit) are not drawn upside down but just normal.

--------------------
# LVL ericbrick.ini
# Edited with Alpha version of Lemmings Level Editor
releaseRate = 1
numLemmings = 50
numToRescue = 48
timeLimit = 4
numClimbers = 1
numFloaters = 1
numBombers = 1
numBlockers = 0
numBuilders = 8
numBashers = 2
numMiners = 1
numDiggers = 1
xPos = 420
style = brick

# Objects
# id, xpos, ypos, paint mode (), upside down (0,1)
# paint modes: 8=VIS_ON_TERRAIN, 4=NO_OVERWRITE, 0=FULL (only one value possible)
object_0 = 9, 964, 192, 4, 1
object_1 = 0, 964, 160, 4, 1
object_2 = 7, 786, 165, 4, 1
object_3 = 1, 496, 85, 0, 1

# Terrain
# id, xpos, ypos, modifier
# modifier: 8=NO_OVERWRITE, 4=UPSIDE_DOWN, 2=REMOVE (combining allowed, 0=FULL)
terrain_0 = 5, 946, 229, 0
terrain_1 = 5, 1010, 229, 0
terrain_2 = 5, 1074, 229, 0
terrain_3 = 5, 1138, 229, 0
terrain_4 = 5, 1202, 229, 0
terrain_5 = 5, 818, 229, 0
terrain_6 = 10, 434, 225, 0
terrain_7 = 35, 750, 101, 0
terrain_8 = 35, 750, -27, 0
terrain_9 = 5, 1116, 48, 0
terrain_10 = 5, 1180, 48, 0
terrain_11 = 5, 1244, 48, 0
terrain_12 = 5, 562, 229, 0
terrain_13 = 5, 1266, 229, 0
terrain_14 = 5, 498, 229, 0
terrain_15 = 5, 626, 229, 0
terrain_16 = 5, 690, 229, 0
terrain_17 = 5, 754, 229, 0
terrain_18 = 7, 746, 48, 8
terrain_19 = 7, 810, 48, 0
terrain_20 = 7, 874, 48, 0
terrain_21 = 7, 938, 48, 0
terrain_22 = 7, 1002, 48, 0
terrain_23 = 7, 1066, 48, 0
terrain_24 = 7, 691, 48, 8

#Steel
# id, xpos, ypos, width, height

#Name
name = When
----------------

the resource extractor doesnt do anything for me.... help help  :'(

Fixes/changes 0.70 -> 0.71
#  Upside down objects are drawn correctly. Also upside down objects are inactive.
#  Fixed bug introduced in 0.69 which prevented level progress from being saved in the player stats
#  "Save Replay" is not possible any more for single levels started with "load level"
#  If a level doesn't contain a valid entry, Lemmini won't crash any more.

@ccexplore
I sure hope so, though I'm in a "not so much lemmings" mode as well at the moment ;)

@Happy_Feet
What exactly is the problem? Do you have Lemmings for Windows at all? Did choose the WINLEMM directory as source directory for the extraction? What happened then? Any error messages? Your OS would also be helpful...

i have windows xp professionel and the latest version of java. of course i have the windows version of lemmings, im not an idiot  :-? . yes of course i set it to extract from the WINLEMM directory. it just gets to the Extracting Resources window and stops... so i have to click cancel... la de da da da nothing :'( :( so yeh, thanks!

There is nothing printed in the extraction dialog? No message boxes as well? If so, I would guess either Lemmini or the JVM throws an Exception that is either unexpected and not properly caught or uncatchable (e.g. out of memory).

You could try to download the JAR from my site and start it from the command line. Start cmd.exe, change to the directory, in which you copied the JAR, then type "java -jar lemmini.jar". Now (try to) extract as usual and have a look at the command line. Are there any errors/exceptions printed?

I'm not sure if this should be considered a bug or just a side-effect of hi-res.

Edit: Shouldn't "no" be capitalized in Oh No! More Lemmings?

To be honest, I messed around quite a little with this damned builder break conditions and either the digger will continue to dig through pretty large gaps or he'll stop digging too early. I put it on my list, though I'm not sure completely how the perfect algorithm would look like.


Most probably true. Will be fixed.

Last update for some days, as I will be away until monday evening.
Happy Easter to all of you  :D

Fixes/changes 0.71 -> 0.72
#  All level files start with "# LVL" now (well, I hope so)
#  General exceptions are caught during extraction. Also tried to catch hard errors (java.lang.Error)
   during extraction and in main game loop.
#  Renamed "Oh no! More Lemmings" to "Oh No! More Lemmings" (capital "N")
#  Extended basher break condition by one pixel to compensate for hires problems (one pixel left after
   mask is applied, but basher stops).

I noticed in the previous version that at the beginning of a level the lemmings go faster for some reason. The falling is faster and then for about 1 second the walk faster and then the tempo becomes normal. Did anyone notice?

2 requests:
custom lemming graphics support
special graphics lvl's cannot be loaded

I like the new code dialog, btw.

Custom lemmings and "special graphics" styles have to be added as Lemmini styles as I can't extract them from WinLemm.
Also I must admit that I'm not so keen on writing another converted for the DOS format. Well, it would be possible of course, but I don't really think it's worth it.

Anyway, creating styles is possible already and everybody is invited to do so. For this, all the graphics have to be converted to or (re-)created in hires of course, which will be the main work to do. For a first fo, simple (linear) upscaling should do (though I wouldn't guarantee that some lemming types might misbehave). Graphics have to be in GIF format, animated object are stored one frame above the other.  Writing the styles ini is then straightforward even without editor support. I'm also willing to support any styles creator with that if anything is unclear.

Loading "LVL" files that were designed for other styles than that supported by WinLemm will always be a problem since there is no definition what the style value inside the level file means for non standard levels (even if the according Lemmini style would exist). Note that loading LVL files is a bonus anyway to allow people to play their custom levels in Lemmini without conversion.
The usual way however should be to load levels in ini format. Indeed as I wrote somewhere in this thread, Lemmini converts the LVL to ini anyway, so you can can grab this temporary file and e.g. change the style to an existing Lemmini style. This way, also loading levels with custom styles will work.

i changed my OS to ubuntu dapper drake 6.06 beta. installed the latest java runtime engine and switched to it as well. ran java -jar lemmini.jar. when extracting, following dialogue appears in terminal O.o

Exception in thread "Thread-2" java.lang.NullPointerException
       at extract.Extract.extractLevels(Extract.java:318)
       at extract.Extract.run(Extract.java:60)
       at java.lang.Thread.run(Unknown Source)

any ideas?

This should mean that there are no files in the directory used as source of the level extraction.
E.g. "\LEVEL\OHNO". Another possibility would be that your source directory names aren't capital letters. Or maybe there's still a glitch path handling if the source path contains uppercase and lowercase characters. I could try some testing with Knoppix next week.
What are your full source and destination paths you enter in the extraction dialog?

I just tried on my Mac and I get the same error discussed earlier with the patch failing.  Works fine under Windows 2000 in Virtual PC.  I'm running OS X 10.4.6 with the latest version of Java.  I even tried puttting the resources folder at /winlemm/ and the extract to folder at /lemmini/ and it still didn't work.  I'd be happy to help you with further testing if you'd like (sounds like you don't have a Mac).

Hm, yeah, I sometimes tried under Knoppix, but my source folder was on an NTFS partition. Chances are that Linux treats mounted NTFS volumes case insensitive. I will check again if I somehow messed the capital letters in the source path.
Anyway: Can you please post the exact error message and maybe the last lines printed out in the text window?

Ok, I just tried again with Knoppix. This time I copied everything (source, jar, dest) into the Knoppix RAM disk. Also I used a mix of upper and lower case characters for source and destinations paths. Worked flawlessly.

My guess it that the path or file names of the source directory are wrong. In the WINLEMM directory if the original CD, every path and file name is uppercase. If you use the files from the CD or copied them from there, this should be the case and the extraction should work. If, for any reason, the paths/files inside WINLEMM are lowercase, then rename them to uppercase.

Anyway, I added some error checking, so when source paths are missing, the correct error should appear now. Give it a try.

I downloaded a new copy from your website and tried it again.  This time I extracted straight from the CD's WINLEMM folder without copying to the hard drive first.  The Finder as well as an ls from the terminal both show that all the folder and file names in there are uppercase as you would expect.  Lemmini still fails with the same error.  I'm using /lemmini/ for the destination folder.  Here's a screenshot of the error: http://www.lemmingsuniverse.net/lemmini.png

In case it's important, the complete path to the WINLEMM folder on the CD is /Volumes/PAINTBALL/WINLEMM (I have Windows Lemmings as part of Lemmings Paintball).

Well, maybe that's exactly the problem...
I wasn't aware that there are other distributions of WINLEMM but chances are that some of the files were changed fot this re-release on the Paintball CD.
Since you are a legal owner of WINLEMM: could you try the data files from the Lemmings File Portal (link in Mindless' sig). Maybe you could make a diff of both directories to see which files differ or send me your files so I can have a look. My mail address is on my lemmini page.

And btw: why is there no link to yet to Lemmini on your site  ;) ?

That's strange...I already wrote this once, but it didn't show up.

I downloaded the Windows Lemmings data files from the Lemmings File Portal, and renamed the folder they came in to WINLEMM.  Ran the extractor and got the exact same error as before.  Running diff showed that there were no differences between files.  The only thing it reported were some missing files in the stuff from TLFP, namely the exe and dlls (to be expected).

Anyway, I don't think it's the files on the Lemmings Paintball CD that are the problem, because I ran the extractor in Windows 2000 in Virtual PC (on the same machine) and it worked fine.  I even was able to copy the extracted files over to OS X, modify lemmini.ini, and Lemmini worked fine.  I guess the good news is that other than this problem with the extractor, everything else seems to work fine under OS X.

Is there anything I can do to run Lemmini and have it give you more output that would be helpful?  I don't know Java, but I'm reasonably good with C and Objective-C.  I'm not completely inexperienced when it comes to development/debugging, so I'm certainly willing to help with whatever I can.



Done :).

(I linked to lemmini.de.  If you'd prefer that I link to the redirector you posted here, let me know and I'll change it.)

Yes, sorry, you're right. If it works in Win2k emulation, the files must be ok.


I uploaded another version which prints out both checksums (expected, real). Though I thought I cared about endian issues, one of the ideas I came up with would be that on a PPC Mac, the CRC32 might create different results due to big endian format. Or do you already have an Intel Mac?
Give it a try and post (or mail me) both CRCs printed in the error message.
If the checksums are just not endian flipped, please send me the GIF file the extractor complains about (bricko_0.gif), so I can investigate why it's different.
My second idea would be that the GIF extractor I use creates different files on WinXP/Intel and MacOSX/PPC. If so, I could try to switch the extractor or use PNG instead.


No, Lemmini.de is fine. Thanks :)

Ok, I downloaded the new version and ran it.  Same error.  Here's the output:

Patch files
styles/brick/bricko_0.gif
Patching of file /lemmini/styles/brick/bricko_0.gif failed.
CRC of source (0x60d9a1ef) differs from that in patch header (0xc429a510)

I'm doing all of this on my main machine which is a 1.33 GHz PowerBook G4.  I've also got a dual G5, but I haven't tried Lemmini on that machine.  I don't have an Intel Mac yet, but I'm considering buying one pretty soon.

I'm having real trouble with this forum loading (as I think are other people).  Do you mind moving the discussion to email?  My email address is on my website, and I'm assuming the address on your website is OK.

Ok, I think know what the problem is. Indeed, the Mac JVM behaves a little differently when it comes to handling images with indexed colors. I'm not quite sure if this is a bug or a feature but I think I know a way to work around this.
Then again, for this I need to patch lots of my reference resources, so I might need some days to come up with a new version.

Glad to hear that you think you've figured it out.  Let me know if you need me to do any more testing for you.

What's the current fall-speed of fallers and floaters in Lemmini?

Damn, this forum is so slow :(

Anyway, Fallers fall at 3 pixels per frame, floaters at 2 pixels per frame.
Pixels here are hires pixel with 1 lores pixel = 2 hires pixels.
To compensate for the higher resolution, Lemmini runs at 30ms/frame instead of 60ms/frame.
I did some (rough) visual comparison with the PAL Amiga version which looked ok for me.

Regarding Mac issues: the problem is the palette index. Problem is that when I extract the pictures to GIFs, some pictures need the palette at offset 0 and others need it at offset 0x80. So I just copy the palette to both offsets which I supposed to be a "quick'n'dirty" solution at the time. However it would have been more clever to patch the indices in the image data to always use the "low" palette.
The problem is that on the Mac, for some reason, when creating the GIF, "automagically" only the low indices are used, while on Win/Linux, the indices are not altered (so some are low, some are high).
I first thought that a API function called "PixelGrabber" would behave differently by detecting double entries or whatever. This theory proved wrong unfortunately, though it seemed to be the most logical explanation. Indeed I wrote my own cheap Pixelgrabber-Code, but the result was the same.
Now obviously the BufferedImage-Class itself is implemented differently on the Mac and I can't simply recreate it or modify its behavior.
Solution is to adapt the Linux/Win-Versions to the behavior on the Mac: I will now patch the indices of the pixel data to always use the low data - so the outcome should be the same as now on the Mac. Unfortunately I patched quite a lot of pictures which I use as reference for creating the patches - which I use to create the wanted resources from the extracted resources. This means I have to patch all of my changed "reference" pictures to low indexed palettes which would be too much work to do manually. So I have to write myself a tool for this, which will delay the fix a little.
Still I hope this issue will be fixed next week. The upcoming weekend is a "long weekend" in Germany, since Monday is a holiday. Let's just hope I manage to stop playing around with my new GPS-PocketPC (great toy) and write that bloody palette patcher ;)

Ok, thanks to the long weekend and Andrew's (LU Admin's) testing efforts, Lemmini should finally work on the Mac. Well, gameplay obviously worked before, but not the extraction.

Fixes/changes 0.73 -> 0.73
# Fixed bug that hindered Lemmini from loading custom styles from the "styles" folder
# Changed GIF extraction routine to work around bugs in the Mac JVM: in indexed color mode
  the MAC JVM doesn't use the correct indices when creating a BufferedImage from a pixel buffer
  and an IndexColorModel. Firstly, it always selects the lowest possible index for colors that appear
  multiple times in the palette. Secondly, sometimes wrong indices with similar colors are used.

Instead of saying "You finished level <insert rating name here>" could the congrats message say something like "You finished all the <insert rating name here> levels."?

Should be no issue. I think I'll change that in the next update...