Page 5 of 22

Re: OpenMW 0.46.0

Posted: 04 Nov 2019, 12:37
by Capostrophic
Hm, no, the changelog I constantly upload hasn't been touched by anyone else (it would be nice if it was), although a few ideas were picked up from Atahualpa's old raw draft. Attached for convenience.

Personally I think all OpenMW-CS new features and almost all fixes can be covered in a single video. As for the engine itself, obviously shadows and recastnavigation should be demonstrated, a bunch of modding gimmicks additions too. Some of more significant changes easy to show in a video IMO are:
#4544 + #4805: they make the timing and the feeling of the Morrowind Prison Ship opening much closer to vanilla
#5004, #4961, maybe #5074 and #4787 and very maybe #5155 (you can lock actors too): for humorous moments. 4961 might be difficult to show in a video, but it can still be mentioned.
#4952: light source flickering was adjusted before, but it shouldn't need to be adjusted anymore.
#3006 + #4778: fixes for these make Interiors of Illusion puzzle in Sotha Sil Expanded mod solvable. SSE was mentioned before in a newspost, and this is the most obvious result of the compatibility improvement. You're supposed to hit a fire atronach with a projectile from the spell you get to light up the semitransparent road to your goal.
#4240 + #4783 + #4240: while there's still the problem with clouds during transition, ash storms and blizzards are much more sensible than before.
#4810: Rakis from Tamriel Rebuilt are very cute, and now they can move properly.
#4918 + #4989: both are necessary for the end-game Dagoth Ur to have his shield ability's VFX display properly. #4828 could be mentioned too.
#5122, #5147, #4962, #4540: Easy to show.
Aside from NiSwitchNode which basically shares its place with Glow in the Dahrk-like and Graphic Herbalism-like behavior demonstration NiRollController is a NIF record with the most obvious effect (rotation).
Of course not all of these must 100% be in a video.

And significant distant terrain optimizations should be mentioned.
This allows for further work to support newer NIF features such as... NiRollController, NiSwitchNode, NiPalette, NiTriStrips and the fact that many NIFs from those later games can already be loaded into OpenMW.
Pedantry: these NIF record types aren't from the later format versions, they're still a part of 4.0.0.2 and are used in Morrowind mods. NIFs from later games can't yet be loaded in git revisions of openmw, although some very simple non-animated geometry from Oblivion and Civ4 can be in my branch. Overall it's obviously totally inferior to cc9cii's (not yet public) work. I had some positive response to an idea to explore some basic differences of the NIF format used in Morrowind versus the NIF format used in Oblivion in something intended to be published as a blog post. Maybe I'll get to writing that soon™.

There were cynic observations elsewhere that we mentioned Fallout 3 mesh loading but had no trace of Lua scripting in the recent newspost. Welp, the former is easier to add.

Re: OpenMW 0.46.0

Posted: 05 Nov 2019, 04:40
by raevol
psi29a wrote: 03 Nov 2019, 20:39 Raevol, as far as I'm aware is still release manager... if he is still game.
Yep. And I'll have time to do it this time, unlike last release.
Capostrophic wrote: 04 Nov 2019, 12:37 Hm, no, the changelog I constantly upload hasn't been touched by anyone else (it would be nice if it was), although a few ideas were picked up from Atahualpa's old raw draft. Attached for convenience.
Where is this being stored? Is it this phpbb file attachment? I used to keep it locally and then put in in a gist when I needed outside input. I had an idea a while back to start an openmw-community git repo a while back to store non-engine-code things like this.

It looks like we already have an openmw-docs repo, that might be a place to start?

Re: OpenMW 0.46.0

Posted: 19 Nov 2019, 07:32
by akortunov
I suppose these things are left for 0.46:
1. Get an approval from AnyOldName3 about backporting OpenMW-related fixes from upstream OSG to the OSG-on-steroids.
2. Run a yet another coverity scan.

Excepts of these tasks, the code itself is good enough to start an RC-phase, IMO.

Re: OpenMW 0.46.0

Posted: 19 Nov 2019, 12:31
by psi29a
How is OSG a blocker for 0.46?

I will do another coverity scan.

Re: OpenMW 0.46.0

Posted: 19 Nov 2019, 14:04
by akortunov
psi29a wrote: 19 Nov 2019, 12:31 How is OSG a blocker for 0.46?
It is not a STRICT blocker, but upstream OSG contains several fixes, important for mods compatibility (e.g. improved TGA loading and support for paletted DDS). I suppose it would be very good to have these fixes in the OSG-on-steroids (which is used at least in builds for Windows).

Re: OpenMW 0.46.0

Posted: 19 Nov 2019, 14:45
by psi29a
Okay, let me know which commits those are as well so I can propose them for OSG-3.4 in Debian (Ubuntu).

Coverity scan inbound...

Re: OpenMW 0.46.0

Posted: 19 Nov 2019, 15:38
by akortunov
psi29a wrote: 19 Nov 2019, 14:45 Okay, let me know which commits those are as well so I can propose them for OSG-3.4 in Debian (Ubuntu).
I suppose Capostrophic is a right person to answer such questions since he cherry-picked related commits.

Re: OpenMW 0.46.0

Posted: 19 Nov 2019, 22:37
by Ace (SWE)
I'd respectfully request waiting at least till next week before starting an RC round, even if everything else works out ahead of that.

I'm currently busy doing a certification this week - and weekend, and have no good way of triggering any builds remotely at the moment.

Re: OpenMW 0.46.0

Posted: 19 Nov 2019, 22:39
by testman
Good luck with certification.

Re: OpenMW 0.46.0

Posted: 19 Nov 2019, 22:52
by psi29a
Roger that.