Lemmings Forums
NeoLemmix => Bugs & Suggestions => Closed => Topic started by: namida on March 18, 2016, 02:54:12 PM
-
So, I think that before we look too much further into the physics, we should have a set of test maps we can use to see in practice what happens. Looking at the code is of course very valuable, but this doesn't really let us see how things work in practice, nor is it optimal at all for visualising a situation and asking how it should work.
As such, I'd like to suggest a project of making a set of test maps. I'm making a graphic set specifically for this, that focuses completely on functionality and not at all on any visuals beyond what is needed to understand how things work (or in some cases, simplify it, such as having pieces designed to help with judging distances).
I'm starting by converting the existing Climber test maps. I'll probably end up doing most of this myself, but if anyone would like to help - whether by actually making levels or just by putting forward suggestions - feel free to do so. It doesn't really matter if some things get duplicated; what's important is that they produce useful information.
The current contents can be found here, as can the graphic set:
https://www.dropbox.com/sh/u3tqe18rcwqarvu/AABlpM1PCn0kSIW5V6JRhkUra
-
Here's a load of situations involving 1 pixel gaps/overhangs that used to cause glitchy behaviour in lemmini, along with a lot of climbers stuck in terrain that would lead to the famous climber glitch in most ports of lemmings.
-
Nice! Although it doesn't look at eg. different heights for most behaviours, this will be a useful map to get a quick look at all behaviours. As such, I've put it in as the "Climber General Test". Nice idea to put the steel copies of other terrain; while glitches relating to steel's presence are unlikely, they'd also be much trickier to notice if they did occur, so this map will be good for that.
I myself made one that tests the hoister situation with both 1px and 2px tall gaps, at various climbing heights.
-
One general idea: If possible, make the test maps in such a way, that the level gets solved exactly when the behavior is the intended one. Then we can put all these test maps into a test level pack and record replays for all test situations.
Right now this is more work, but once itis done, we can just use the mass replay checker to automatically verify that the code produces the correct lemming behavior.
-
Should we also have every single object type in the test graphic set? It would make sense to use this graphic set as a testing ground for all objects and how the lemmings and skills interact with them. It will also be very useful to test the trigger areas of the objects against terrain.
I haven't seen a graphic set that uses every single object type yet, so this one would be the perfect testing ground for all of them.
-
Should we also have every single object type in the test graphic set? It would make sense to use this graphic set as a testing ground for all objects and how the lemmings and skills interact with them. It will also be very useful to test the trigger areas of the objects against terrain.
Absolutely. I just haven't got around to putting them all in yet.
Do note, however, that the graphics should be designed for functionality, not appearance. Except for the pre-placed lemmings (and a flickering one-pixel, which is designed to mark the spot at which a skill should be assigned), the objects so far consist of a text caption describing the type of object, and a purple square / rect / pixel marking the trigger area. This theme should be kept across all objects in the set. In the case of those where the frame count matters (eg. triggered traps), this should also be indicated in some way, but I'll sort that out once I get up to such an object.
-
As Nepster has made lots of physics changes in the experimental release, I wonder if we're still going to make test maps that reflect these changes (and add the objects in the graphic set, of course).
-
At some point, maybe. This idea is one of those that seemed good in theory, but never took off in practice - perhaps due to the large number of physics changes currently being discussed.