I'll be sure to include that detail. However, to answer your question now, it offers 4 different options for how NeoCustLemmix acts when used in test mode with the editor (IIRC, traditional CustLemmix has these options too, except maybe the last one?):

* Display normal preview / postview screens
* Skip preview screen, show postview screen
* Skip both screens; exit test mode when level terminates
* Skip both screens; exit test mode and save a replay when level terminates

So, trying with 4-pass, it doesn't seem to give much of an improvement in filesize on most levels, especially considering how much longer it takes to run (nearly 5 hours since I launched it; one rank is finished and two others are very close, while Mayhem has a fair way to go).

Relative to 1-pass:

Fun: 1.41MB > 1.37MB (2.8% reduction)
Tricky: 1.10MB > 1.07MB (2.7% reduction)
Taxing: 0.98MB > 0.94MB (4.1% reduction)
Mayhem: 1.07MB > 1.00MB (6.5% reduction)

Overall reduction on the EXE was only about 150KB.

However - it's also very possible this is due to subtle differences between the Copycat and Cheapo graphic sets (something as simple as a terrain piece having one pixel that's off just by a slight amount in just one of the RGB values could make the program not detect the piece in the level image); indeed I've noticed some pieces not being detected when I can see very little reason why they wouldn't be, or some levels having leftover pieces in the 3rd or 4th runs where from a quick look at the level map I'd've expected it to pick up everything on the first pass. I'll do some test runs with true Cheapo levels at some point and see how the results look there.

Quote from: NaOH on February 23, 2015, 01:54:29 AM

Edit: out of curiosity, would it be possible to save the smileys used by existing posts as-is and constrain all the changes we decide on strictly affect posts hereafter?

AFAIK, it's not possible for the same sequence of characters to translate to different smilies based on when the post was created.  New smilies will need to get their own sequences of characters, and existing sequences of characters in used by existing smilies should probably be left alone.

Nothing is impossible.
Option A) Modify SMF's code to implement this difference.
Option B) Have a hidden, second code for all of the pre-existing smileys, and write a script to modify old posts to use this code.

Seems to be working well, so I'm running the first full-scale test run. From the first 10 levels of Fun alone, it's cut about 100KB off the total filesize.

Comparisons:

Level 1: 51KB > 49KB (3.9% reduction) [Dirt]
Level 2: 31KB > 17KB (45.2% reduction) [Pillar]
Level 3: 23KB > 8KB (65.2% reduction) [Fire]
Level 4: 108KB > 96KB (11.1% reduction) [Crystal]
Level 5: 71KB > 50KB (29.6% reduction) [Marble]
Level 6: 74KB > 69KB (6.8% reduction) [Dirt]
Level 7: 25KB > 18KB (28.0% reduction) [Pillar]
Level 8: 102KB > 90KB (11.8% reduction) [Crystal]
Level 9: 75KB > 55KB (26.7% reduction) [Marble]
Level 10: 129KB > 126KB (2.3% reduction) [Dirt]

Overall: 684KB > 573KB (16.2% reduction)

As you can see, the straight-edged tilesets benefit far more than the rough-edged ones, but all of them benefit at least a bit.

EDIT: For all of Fun; 1.66MB > 1.41MB (15.1% reduction)
This is including the LEVEL000.DAT file, which has increased in size by only 7KB.

Tricky: 1.46MB > 1.10MB (24.7% reduction)
Taxing: 1.31MB > 0.98MB (25.2% reduction)
Mayhem: 1.60MB > 1.07MB (35.6% reduction)

And in total, for the EXE:
8474KB > 7089KB (16.3% reduction)

While still very large, this is definitely an improvement! I'll test it out a bit myself before I upload anything.


EDIT: Goes without saying that there'd be some issues... xD Time to look into it...

EDIT: Looks like the issue is a very minor one, and solely in the LVL files, not the VGASPECs. It can be fixed with a simple adjustment to them, without the need to reprocess all the levels. I'm also doing some multi-pass test runs to see if this is what was causing the problem with multi-pass terrain building. (It's very plausible that it was, and running extra passes could potentially offer a huge improvement on some levels.)

EDIT: This was indeed the problem, or at least, a two-pass conversion comes out fine on the one level I tested it on (where I was noticing issues before). Time to try 4-pass conversions and a different level.

EDIT: All worked perfectly! There's some oddities in which exact pieces it uses for what purposes (which of course is invisible to the end player), but the resulting level is pixel-perfect to the directly-made-to-VGASPEC one at a much smaller filesize. And since multi-pass is working now, I'm running the conversions again with four passes rather than one; we'll see how much of a difference this makes. (The two levels I was using as primary test runs were Fun 1 and Fun 5; on the former 4-pass makes a noticable difference from 1-pass, while on the latter it's negligable - but 1-pass already does a lot on the latter.)

Okay, well it seems that there's no remaining major issues, so time to tackle some optimization issues. I'm attempting to write code that tries to recognize any terrain pieces in the level image that match ones from the graphic set, and thus blank out those parts of the image and use regular terrain pieces instead. This should considerably reduce the filesize.

For first attempt, it's only going to look for pieces in their normal orientation, but once I get it working, I'll attempt to identify flipped pieces as well. In many cases, particularly with rough-edged sets, this won't completely remove the need for a VGASPEC, but should cut down on how much of the level uses one, the result being most of the VGASPEC will be blank space, and therefore, a much smaller filesize. Of course, the tradeoff is that the conversion will be a lot slower.

EDIT: It's working well. Very slow performing, but functionally perfect. Without yet implementing any handling for flipped pieces, it already cuts the filesize nearly in half for Fun 5 (a straight-edged level); it isn't so effective on Fun 1 (a rough-edged level).


EDIT: I've made some minor optimizations that should speed it up, as well as working on adding support for flipped pieces (any combination of horizontal and vertical flipped states). I've also applied a (rather kludgy) fix that should prevent it repeatedly duplicating pieces with repetitive designs (such as the striped pillars in the Marble (aka "Pink") style), but rather cover them in as few pieces as possible. I'm no expert on image processing so there's only so much I can do, though ultimately, it should be good enough to work, even if it's not as optimal as potentially possible. In terms of speed, I can think of ways to further optimize it, but I'm going to worry about getting the current attempts to work first, then I can look at improving the speed.

Okay, after some further testing, adding support for flipped pieces definitely helps further. On the other hand, I also experimented with running more than one pass - the improvement it offers is negligable, especially compared to the extra time it takes (whereas a single pass isn't overly time consuming). Additionally, multiple passes seems to give minor oddities that don't appear with a single pass. So, I'm limiting the tool to single-pass only (for now anyway).

Im torn on this one i can't really decide
i like this idea but how are you find secret levels becuase well
in the system.dat edition you have an option to put in secret level the question is how are you susposed to know the codes to them?
but however i wanna see how this plays out

Well, generally secret levels are accessed by a secret level trigger (the small 8x8 solid color block objects in each graphic set). You set the S value to the rank and L value to the level number of the secret level. (eg: If Taxing 31 was a secret level in Orig, you'd set the S value to 3 and the L value to 31)

However, you can indeed hide them by code - though obviously, make sure there's some way for people to figure out the code. In this case, the best way to find the code would be to enable accessing secret levels via cheats, then use a cheat code to reach the level. The code will be displayed on the failed screen, as always. You can then turn the option to allow reaching secret levels via cheats off after you've got the code (or leave it on if you prefer).

It isn't possible with Flexi to have a set code for the secret level (in the way that "Trivia Time!" from LPIII Bonus has). That was a hardcoded thing specifically for that pack.

I'm considering adding more ways to unlock secret levels in future versions.

---

On the subject of ones for existing games, the only one I definitely have in mind for an official game is 100% on Tricky 25 "Cascade".

I don't plan to make a "combined" player, so there won't be any cross-game achievements.

One for completing each rank is something I'll definitely include by default (possibly excluding, eg. the 2P ranks in Orig / OhNo).

I'd like to stay away from really pointless ones, like quitting a level via Esc 100 times.

Likewise, the "no upper route" one for Doomsday Lemmings Outbreak 4 is probably a bit complex to include while still allowing flexibility to enough of an extent for people to make their own packs. It'd probably be limited to things like "find / beat a certian level", "complete X levels", "find a secret level", "beat a level under a certain time", "beat a level without a certain skill or with no more than X of a certain skill", etc.

Also - the 1-builder one for Tame 13, I don't know if that's possible on NeoLemmix. IIRC, it relies on the blocker-pushing-into-terrain glitch, which is fixed in NeoLemmix (as is virtually everything considered a "glitch", apart from direct drop which I've intentionally left unfixed).

I'm sure it could be done if we do decide on making any such changes.

What do people think about the idea of adding an achievements system to NeoLemmix?

Obviously, there wouldn't be any online database - it'd just be a fun thing that's kept track of in the save file and players of packs can aim for. Of course, it'd be supported in custom packs, as well as me adding some of my choosing to the Lemmings Plus games and Doomsday Lemmings (and perhaps we can look at some community-chosen ones to add to the NeoLemmix players of the official games too).

These could be things like "Save 100% on level x", "Complete level x without using builders", "Find all the secret levels" etc, as well as the more-generic ones for completing each rank.

Let me know what you think about this feature; it's one I considered in the past (in fact, originally I thought about having it debut in Lemmings Plus Omega, but never got around to making it) but didn't implement yet.

---

Now that this feature is confirmed, I'd like to start gathering suggestions for achievements for the official games. Feel free to make suggestions either in this topic or by PM. Please take into account the limitations of the achievements feature. The primary one is that achievements take the form of challenges on a single level, there can't be any that span multiple levels in any way. It is, however, possible for a single level to have more than one achievement related to it. That aside, the criteria for achievements must be among the following (it is possible to mix-and-match more than one of these in a single achievement):
> Saving a certian number of lemmings
> Completing the level in under a certain time (this can involve seconds and even frames, rather than just whole minutes)
> Limiting the quantity used of a specific skill
> Limiting the total number of skills used
> Keeping the release rate within a specified range

Regarding Copycat Lemmings, I will first contact Essman to see if he'd like to pick them himself, and if not, I'll take suggestions here for that too. EDIT: Essman has said that while he'd like to come up with them himself, he's also happy for some community input. Thus, I'll take suggestions here, but it'll be ultimately Essman's decision which ones do or don't make it into the game.

Remember - for obvious reasons, the challenge must be possible in NeoLemmix. Just because it's possible in DOS / traditional Lemmix doesn't mean it's possible in NeoLemmix; if it doesn't rely on glitches it's probably possible, but this isn't always true, either due to subtle mechanics differences, or differences in the levels themself (for example, the Orig and H94 levels in NeoLemmix are based off the Amiga versions, not the DOS versions; and a lot of levels have further minor modifications to make use of NeoLemmix's capabilities). Some that are very challenging under DOS may also end up being trivial under NeoLemmix, for example one that was suggested was "Complete Taxing 30 with no more than 20 diggers", which would be quite challenging under DOS / traditional Lemmix, but it can be done with only two diggers very easily in NeoLemmix.

Achievements may be rated as Bronze, Silver or Gold, depending on their difficulty. However, let's get a collection of suggestions before we decide on ranking them, or on which ones to actually use. The (virtually-)direct-ports of official games will also get achievements for simply beating their final levels, but these aside, none should be simply "Beat level X" without further requirements.

Original Lemmings
Save 100% on Fun 8 without builders
Save 100% on Fun 10 without floaters, miners or diggers
Save 100% on Fun 27 in under 1 minute
Complete Tricky 11 without builders
Save 99% on Tricky 23
Save 100% on Tricky 25
Complete Mayhem 3 with only 4 skills
Complete Mayhem 13 with no more than 6 skills
Complete Mayhem 25 with no more than 5 builders
Complete Mayhem 26 without blockers
Complete Mayhem 30

Oh No! More Lemmings!
Complete Havoc 20

Xmas Lemmings 91/92

Holiday Lemmings 94
Complete Hail 16
Complete Blitz 16

Covox

Prima

Extra

Copycat