Mike! I want to ask you...

[DragonsLover]

But it could be interesting to correct all of these bugs in the Lemmings games. Instead to have a lot of hacking versions of Lemmings, why not getting the PERFECT version of Lemmings?

[guest]

I would consider the version of Lemmings that came in the original disks as a "perfect" version of Lemmings.  There's no need to hack anything for that.

[Mike]

I was speaking to Russell today, and Lemmigs (as usual) came up. The version of the source I have appears to be "Oh No! More Lemmings", and he seemed to rememebr that there was a fudge needed in that one to make levels work.That means the 3 being added was only in Lemmings, and not Oh No.

So, my source is correct for Oh No, and wrong for Lemmings. Lemmings had a #3 in there, Oh no had a #0. (or so he thinks...) My Amiga source is from L1.

Dont have the ST source.. But being 68K, Id imagine it was just the Amiga one for all the processing....

[Shvegait]

I used LemEdit to import "We all fall down" and "Steel Works" into Oh No! More Lemmings.

In both cases, the levels behaved exactly as you would have expected. In "We all fall down", you need to dig several (3) pixels down before lemmings safely drop. In "Steel Works", exactly one step is required to save the falling lemmings.

Then I imported these two levels into a CustLemm pack. The result?

The drop in both cases is non-fatal. But by how much? The drop in "We all fall down" is a 66-pixel drop. I added builders to the level, and with a one pixel increase in height, the drop becomes fatal.

So, clearly CustLemm is the version with the buggy max fall distance.

Safe Fall Distances (DOS):
Lemmings: 63
ONML: 63
CustLemm: 66

Now if only we could fix CustLemm...!

[guest]

It's probably within my ability to fix custlemm, but the question is, should we?  I don't really want to break someone's level due to the change.

Though hopefully with such a small difference, it shouldn't affect too many levels if any.

[Shvegait]

I think we should, if it wouldn't be too much of a problem for you to do. Besides, if a level is broken because of it, you can always play on the old CustLemm. Of course, you might not know if a level is broken because of it if it becomes impossible... Either way, it would be nice to at least have the option of playing with the correct safe-fall distance.

[DragonsLover]

Should we? I want to answer YES!

For the broken levels, just modify them to make them work.

And by the way, why not changing the safe-fall distance of Lemmings CD version too, to be exactly like the floppy version?

I'm sure you'll answer: Why not playing the floppy version instead? I won't because the first level song always reappear when there's a failure, and not in the CD version.

[Shvegait]

When I was younger, I used to think that the repeating first level song was kind of a punishment for losing the level, and not a bug It did make beating a level and hearing the different songs a good reward Of course, it could get annoying...

[Timballisto]

From what Mike said though I got this:

Lemmings: 60
ONML: 63

[Shvegait]

Timballisto, the tests (for DOS), say otherwise.

With a max safe-fall distance of 60, you would need 4 steps on Steel Works, and you'd need to fully dig to have a safe drop (That is, you would need to use every single digger given to you, since only diggers can get as low as 60 pixels (I think). The lowest a walker can get is 61 pixels.)

[Timballisto]

Okay then.  Just saying...

[guest]

And by the way, why not changing the safe-fall distance of Lemmings CD version too, to be exactly like the floppy version?

I'm sure you'll answer: Why not playing the floppy version instead? I won't because the first level song always reappear when there's a failure, and not in the CD version.

Well, I don't think I have the CD version, so someone'll need to give me a copy of it for me to even attempt doing anything with it.

[DragonsLover]

The CD version is the SAME thing than lemmings on floppy, excepted that the Lemmings musics are ok and the safe-fall distance isn't good. I have this version and it is possible to get it on some abandonware websites.

(I can't send it for now, I'm not at home!)

[Mindless]

If you hack CustLemm, make sure you hack the version number (or something else) so that people are able to check which version they have.

[guest]

I was speaking to Russell today, and Lemmigs (as usual) came up. The version of the source I have appears to be "Oh No! More Lemmings", and he seemed to rememeber that there was a fudge needed in that one to make levels work.That means the 3 being added was only in Lemmings, and not Oh No.

So, my source is correct for Oh No, and wrong for Lemmings. Lemmings had a #3 in there, Oh no had a #0. (or so he thinks...) My Amiga source is from L1.

Dont have the ST source.. But being 68K, Id imagine it was just the Amiga one for all the processing....

Ok, here's a twist to this ongoing saga  :

So, upon popular request, I went ahead and hacked CustLemm tonight.  I was expecting that the change was in the line that compares fallcount to maxfallcount, with the constant value of maxfallcount set to a higher number.

Imagine to my surprise, that particular line remains the same in CustLemm!

So I looked elsewhere and quickly discovered the source of the difference.  And guess what?  The difference is in the initialization of fallcount in [what I make out to be] the "MakeFaller" macro!  (ie. it's the stuff Mike's talking about with the #3 vs. #0 thing.)

What's really interesting though about this is that the same line actually appears a total of 7 times in the code, so there would've been 7 places for you to change if you hack the binary directly.  But in all 7 places it appears in the context of the "MakeFaller" macro, which means that, if you have the source code, only one change (namely, to the macro itself) would've been needed.

If someone is hacking the EXE directly and wanted to change the safe falling distance, it would've been far easier for him to change the value of maxfallcount (just one place in the EXE), rather than having to change 7 different places where fallcount is initialized.  To be sure, search and replace isn't really that difficult, but still.

So it seems to me that, rather than someone's hacking the EXE, it looks like the result of the higher safe-fall distance in Custlemm was more likely a change from the original source code.  So this points to 2 possibilities:

1) Someone obtained the source code to Lemmings/ONML, and did the modification to the "MakeFaller" macro there, and then compile to generate the EXE

2) There is already an EXE out there that has the changed falling distance, and the person who made Custlemm made changes on the EXE, but never actually touched the falling distance stuff himself

Coupled with what Mike said above, I'm leaning towards #2, and so it really does sound like that there is an "official" version of Lemmings or ONML out there that has the FallCount initialization changed, resulting in the higher safe falling distance!

I checked my copy of ONML and it is unchanged with regards to the fallcount initialization (ie. it uses 3 just like Lemmings), which would explain why Shvegait failed to find a difference.  But of course, my copy of ONML was downloaded from abandonia.com, so I don't know how authentic it is.  It's certainly not the exact original of ONML since there is no copy protection in my copy.