1.41 Non steam VS steam comparison

[The GoldFish]

OK, so the steam version game out *ages* ago.

When it did, the first thing I looked at were the Data files.

The only noteworthy thing that yielded was this thread, which evidently no one was concerned with.

I pretty much stopped there, because, well, we were beta testing the 1.41 patch and I kind of assumed that the only difference between it and the steam release was, the steam check code pasted it. When the armour bugs came up, I threw a few dozen armour at an army of red darwinians, it didn't crash, and decided the ball wasn't in my court.

Recently, though, it has become apparent this is not completely the case.

Issue 1)

In both versions, in only some scenarios, a deployed armour gunturret may become an AItarget style waypoint. Some modders have experienced this, and it is my belief that this can also happen in (at least) Pattern Buffer. It is however virtually impossible to make this behavoir be exhibited in either Biosphere or Temple. Either way, here is a classic TGF mini mod that will exhibit crashing apparently everytime on the Steam version of Darwinia, but not once so far on the non-steam version.

Turret Crasher mod <-- anyone who was in IRC the other day, this trumps the video I recorded to demostrate it (not *totally* sure why I didn't do this then >_>)

edit - Here is a picture of exactly how it looked when it crashed on steam for you

On the nonsteam version of 1.41, if it IS treated as an AItarget, Darwinians will move to the gunturret, and will never leave it to go to another AItarget.

I hypothosize that if one is treated as an AItarget or not it is related to the number of adjacent AI targets - Temple and Biosphere have many, close together, AI targets which all interlink, Pattern Buffer has longer, more defined pathways (as do many mods where the 'cause' is either random, or to try deploying an armour near a spawn point (which usually has only one AI target leading away from it) (edit - this hypothosis seems somewhat unlikely following a few tests I have performed, so, I don't really know how or why this does/doesn't happen)

Issue 2)

For this, please refer to the two avi video files (encoded with the Xvid codec) in this archive; AI Pathing differences

These are both recordings of 1.41 on the nonsteam (norm10cutcomp.avi) and steam (steam10cutcomp.avi) versions of biosphere.

You will notice that, on the steam video, that Darwinians frequently migrate between all 6 spawn points as time progresses. On the non-steam version, this is not the case. From around halfway through the video, ie, once all AItargets have Darwinians at them, all Darwinians move to the centre spawn point. For some verification of this, simply note that at no time in the video, does a group of darwinians ever leave the centre spawn point. In the Steam video, this happens several times. For greater verification, remove the spawn population lock and note what happens between the two versions.

To those of us familiar with the intricacies of Darwinian AI, it would be noted that the Steam AI pathing very much reflects the 1.22- AI pathing, where as the nonsteam version keeps the pathing changes introduced in version 1.30 nonsteam. This basically means that, the "original campaign" and any/all Darwinian using mods play differently between the nonsteam and steam version of Darwinia. I had to go back and update Stricken (for once) because of the pathing changes in 1.3. I possibly now need to do this again to make a steam friendly version. The point of this, however, is not 'grumble grumble more work for TGF', it's more a case of, 'the 1.41 we were given to beta test was *not* similar to the 1.41 released via steam'

From what I am aware of, both the Mac and Linux version Darwinian pathing reflects the current non-steam version of 1.41. They similarly do they feature the Armour crash, however I would expect them to behave in the 'darwinians flock to and never leave the armour' style.

It should also be noted that the AI pathing changes have never featured on any Change log.

This isn't, as such, a discussion, or really much of a report; It is more of a statement. I would like to to add 'sorry I didn't bother to check this properly earlier' - I was under the impression that any issues like this had already been looked at.

Any mac/linux users who, naturally, would be reading the windows issues forum, please feel free to add your experiences with 1.3/1.41 pathing and if Turret Crasher crashes for you (On steam, it will crash well before all the darwinians reach the armour, so if they all get there and just kind of hang about, it didn't crash)