New master snapshot 20111204 has been released!

[aubergine]

I've been away from the game for a while so coming back after a few updates have been released is awesome - great work WZ team!

Regarding map display on game set-up screen, it is a bit manic as mentioned earlier in this topic - rather than having a 500ms delay before showing a map would it be better to cross-fade between maps over a 125ms period? This would still make it feel responsive but remove the "manic flickering". Possibly have a 75ms delay before starting a cross-fade (or some sort of jQuery-like "hover intent"). I think it's the instant transition between one map and the next that makes it harsh on the eye.

Transports seem buggy - I was only able to get about 3 vehicles on a transport (thought they carried 8?) and then when unloading units from transport the game crashed (forgot to get a stack trace - sorry!).

I also had the flying trucks problem like the screenshot posted earlier. Easily solved by selecting them and telling them to move somewhere so not a show stopper.

Regarding AI, I went in to cheat mode and repositioned all the AI rally points so they were near each other - this caused AI units to be much more bunched up and more formidable, especially with VTOLS. Maybe worth coding that in to the AI?

Is there an auto-save feature?

Love the gates - when did those arrive?!

Is there an option to switch between hardware and software mouse pointers?

A huge number of graphics glitches have been fixed, and I'm delighted to see VTOL functions (circle, repairs on pads, etc) fixed too! Already noticed loads of other enhancements all over the game - amazing work!

Any chance of getting heavy laser and stormbringer hardpoints?

Overall the game is looking much more awesome compared to the last version I played. Kudos!

[b1m2n3]

• Change: Readd the team colour fallback for OpenGL 1.4+ (2f816f5bf4)

Thank you for caring a bit for the low-end casual gamers around here. Not everybody has the latest gaming capabilities. Even my newest machine only has intel embedded graphics, and so will my next: I care primarily for battery life, not gaming power (and seriously, Intel is the only choice when you want battery life these days). So it looks like I will be sticking to OpenGL 1.4 for the next year or two at least. (Even the latest Intel Atom CPU afaik only has OpenGL 1.4 support in its GPU).

[b1m2n3]

P.S. without having tried master yet (no OpenGL 2), some observations with 2.3.8:
The game seems to spend a lot of time in pathfinding. On large maps, with AIs and lots of units, units end up waiting several seconds for their path.
Some suggestions for this:

a) when the routing system is overworked, units could start moving in the general direction to at least get them going. Priority for path finding could then be put on short distance orders, that may complete quickly anyway.
b) approximative routing: when the system is overloaded, units could be assigned an approximative route first, while waiting for the routing queue to be processed
c) units closer to control centers and -units could receive a bonus from that (this would make the relay center more useful!)
d) AFAICT the system is currently using a classic A*-search (or bidirectional A*) algorithm. While this is a classic solution, it is designed for single-shot pathfinding. In this context, where we have a large quantity of units to ask for routes, we should precompute and cache some results. This could be done by downsampling the map to a smaller size (e.g. 16x16 chunks), precomputing a full routing map for this, then using this approximative route as a starting point and only refine it for the actual query (checking for buildings blocking the initial path etc.).

d) could be an interesting science project for computer science students at a bachelor level.

Right now I have the impressions that units just wait until their route request was processed. Assuming that we have a 5 seconds delay, this currently means that units wait 5 seconds, then follow the optimal path.
Instead (a-c above) I'd propose something along these lines:

On route request:
1. Unqueue existing requests
2. unless the routing system is idle, queue into "pre-routing"
3. queue into "full-routing" with a priority based on: distance-to-target * distance-from-nearest-(command-center,commander,relay-center)

Routing thread:
4. process pre-routings (e.g. line-of-sight, precomputed route map)
5. process full routings by priority

[zany]

when is that new release going to be put out?

[raymond]

when is that new release going to be put out?

Yes, there are a bunch of bugfixes, so please release a newer master snapshot.

[Navvarr]

I found what seems to be a bug- in the later missions -I'd say about level 27.- The level where you start at the top of the map and have to fight your way to the lower right in order to bring some 'possibly' absorbed other team gamma or alpha.

Anyway, when I get to this team who are holed up in the bottom right of the map- I can't gain control of them to lead them back to my transport and hence end the mission.

Result is that the level doesn't end.

Perhaps someone else could look into this to check it is indeed a bug.

Thanks

[NoQ]

Navvarr: could you provide a savegame?

[cybersphinx]

I found what seems to be a bug- in the later missions -I'd say about level 27.- The level where you start at the top of the map and have to fight your way to the lower right in order to bring some 'possibly' absorbed other team gamma or alpha.

Anyway, when I get to this team who are holed up in the bottom right of the map- I can't gain control of them to lead them back to my transport and hence end the mission.

Result is that the level doesn't end.

Perhaps someone else could look into this to check it is indeed a bug.

Sounds like #1855, which has recently been fixed.

[aubergine]

In this build the large fortress defences don't seem to work very well against land units - they just sort of sit there and stare at the enemy troops and only occasionally do they fire back at the enemy.