Two NM demos of Umbrella

[Donce]

Long time no see eh? :)

Yeah, I can't believe I didn't touch Doom at all since May. Anyway, today I played Epyo's Umbrella. The first map was only possible to complete with Zdoom, which I don't record with, Map02 crashed Doom2.exe at one point but was OK with Boom. Doom2 was happy with the last map, just there were no slopes and textures were missing, but I didn't look back anyway...

In short: I recorded two NM demos, Map02 (with Boom 2.02) and Map03 (with Doom2.exe). Just like always, they aren't of good quality, but I enjoyed recording them so I don't care :-)

[Grazza]

Great to see you back again. I'll watch these immediately...

No signs of rustiness there. :) BTW, the map03 demo only plays back OK with Doom2.exe for me - desyncs early on with Boom, Eternity and Prboom.

[Donce]

Grazza said:

No signs of rustiness there. :)

And no progress as well :p

BTW, the map03 demo only plays back OK with Doom2.exe for me - desyncs early on with Boom, Eternity and Prboom.

The only reason for the early desync that I can think of are the slopes. I haven't been editing for nearly 7 years now, but AFAIK slopes are created by some special linetypes or sectortypes. I know from experience that just merely changing a linedef type may be enough for a desync, even if it was never triggered. So, Boom and its derivatives probably treat those special lines differently from the original exe when calculating monsters' behavior.

[Grazza]

That's probably it, though when Prboom and Eternity attempt to play back this demo, they are in their "emulate Doom2.exe as closely as possible" mode.

Still, this is a pretty unusual situation: "Please make your port emulate Doom2.exe's way of ignoring certain Zdoom features". Perhaps that's not something cph (and co.) and Quasar (and co.) have ever considered a priority!

[Donce]

Hmm, I made an experiment, after which I fell in doubt of the slope 'theory'. I changed linedefs 3 and 8 from type 345 to type 0, and linedefs 5 and 34 from 342 to 0. Then the demo desynced in both Doom2 and Eternity, but in Doom2 it stayed in sync far longer. If it really was for those linetypes unusual for the original Doom, the demos should look the same. So, there's something else.

[entryway]

Donce said:

Hmm, I made an experiment, after which I fell in doubt of the slope 'theory'. I changed linedefs 3 and 8 from type 345 to type 0, and linedefs 5 and 34 from 342 to 0. Then the demo desynced in both Doom2 and Eternity, but in Doom2 it stayed in sync far longer. If it really was for those linetypes unusual for the original Doom, the demos should look the same. So, there's something else.

Try to change all linedefs with type 253 to type 0