Help needed testing 3.2.x Campaign games!

Discuss the future of Warzone 2100 with us.
User avatar
Berserk Cyborg
Code contributor
Code contributor
Posts: 938
Joined: 26 Sep 2016, 19:56

Re: Help needed testing 3.2.x Campaign games!

Post by Berserk Cyborg »

Pushed aaacc29be0c7c99d2003051015e8623788082076. You will have to start Gamma 8 over.
alfred007 wrote: I found one weird issue. I have two saved games, one shortly after the 2nd level missile code is cracked and one a few seconds before the 3rd level missile code is cracked. When I reload the saved game with a few seconds before the 3rd level missile code is cracked the game shows me immediately the video that the 2nd level missile code is cracked, despite the fact that I did this about 15 minutes before.
Ah, yes... I forgot that loading means that all the research a player has is researched again, thus getting captured by eventResearched.

It appears that the LasSat would start to hit the missile silos with about six minutes left on the timer. I also made sure that droids/structures being targeted exist so as to stop an assert when accessing coordinates. And it appears the LasSat is incapable of friendly fire on some droids?

Also your save shows an AI bug. I don't know if it is related to the scout order or what, but it appears they have trouble moving up to ledges and hills to get to something (they try moving towards it, but switch over and attempt to hit it from a lower angle below the hill constantly). Not campaign specific since I noticed it with the bots in the mizaMaze map where there are derricks on the hills.

Bethrezen, it would be best to create tickets for source related bugs so everyone knows of them.
Forgon
Code contributor
Code contributor
Posts: 298
Joined: 07 Dec 2016, 22:23

Re: Help needed testing 3.2.x Campaign games!

Post by Forgon »

Bethrezen wrote:[...] for some reason your base/units are red instead of green [...]
Yes, and this ancient behaviour should be abolished.
Bethrezen wrote:[...] as for the size issue looks like it may be an issue with loading old saves.
The savegame you attached stores a radar size of 32, which did not match my hypothesis. I have been unsuccessful in reproducing the tiny mini map in Beta 1 of this screenshot you provided.

To further investigate this, could you please assure me that the screenshot:
  • was created from the savegame you provided to me
  • was taken with the latest master version
  • was taken before resizing the radar
  • was taken with radar rotation disabled
The maximum size of the minimap has not decreased, which is evident after reading #4689 and this comment.

Rather, only with rotating radar do maps whose larger side is < 128 tiles long now appear smaller, while maps whose larger side is > 128 tiles now appear larger.

Radar size could either be absolute with maps rescaled to fit (as had been the case with rotating radar) or depend on map size. I prefer the latter variant, because it does not decrease the level of detail displayed for larger maps, for which the minimap is more useful than for smaller maps.

While the smallest radar sizes are tiny, they cannot be increased without enlarging all other ones. This is not feasible due to the square-cube law: Currently, the largest maps in use are 250x250 tiles large, which leads to a minimap of 500x500 pixels at the default zoom level of 32. Increasing radar scale by 50% would render them absurdly large with a screen resolution of 1024x768 pixels:
7c-Starscavs (250x250 tiles large, close to the maximum of 256x256) with radar zoom set to 48
7c-Starscavs (250x250 tiles large, close to the maximum of 256x256) with radar zoom set to 48
If you find the largest radar size settings for Sk-Rush or Beta 2 still too small, I can thus offer no solution.
Bethrezen wrote:[...] with regards to the mini map looking blurry here is a side by side to demonstrate what I'm seeing, although it should be noted that this is taken from an older version before the mini map changes. [...]

now as you will see the one where the pixelation occurs the detail is a lot sharper, now ideally what would be good is if the mini map could be rendered in HD so that when you enlarge the mini map it remains sharp and detailed without becoming pixilated, probably the best way to do that is via a vector graphic, as this sort of a thing is exactly what vector graphics are for, question is does the game actually support the use of vector graphics, failing that then maybe doing the mini map as a high resolution bitmap would be a reasonable compromise as that is a lossless format so the detail should remain sharp, but again I'm not sure if this is possible.

Certainly I know that this sort of a thing can be an issue with lossy formats, which is why i tend to use PNGs as they are loss less and retain there detail even when heavily compressed to save size, the one catch is that they simply take a bit longer to load if using a lot of compression but with modern computers being so powerful you barely notice that unless the image is huge.
The cause of blurry minimaps is poor utilisation of OpenGL. I am confident better looks could be achieved, but unwilling to promise a fix before the upcoming release.
User avatar
alfred007
Regular
Regular
Posts: 619
Joined: 31 Jul 2016, 06:25
Location: Stuttgart, Germany

Re: Help needed testing 3.2.x Campaign games!

Post by alfred007 »

I defined new assemblies for Gamma 09 but made obviously a mistake that I can't recognize. When I made a mod with the new labels and the new cam3-4.js file every time I started the transporter to Nexus' base I get instantly the message that the mission failed. In the file attached are the new labels for Gamma 09, the cam3-4.js file with the new assemblies and the logs of my failed tries. I also set truckDefense depending on difficulty.
Attachments
gamma 09.zip
(5.43 KiB) Downloaded 104 times
User avatar
Berserk Cyborg
Code contributor
Code contributor
Posts: 938
Joined: 26 Sep 2016, 19:56

Re: Help needed testing 3.2.x Campaign games!

Post by Berserk Cyborg »

alfred007 wrote:I defined new assemblies for Gamma 09 but made obviously a mistake that I can't recognize
You forgot to add a left bracket for the "position_23" position coordinate. Other than that, it starts fine (it is in my campaignJS branch right now). Best of luck to what could be the most grueling mission ever. Wonder how long the power supply will last for Nexus... maybe we should queue it every five minutes to be 999999 again.
User avatar
alfred007
Regular
Regular
Posts: 619
Joined: 31 Jul 2016, 06:25
Location: Stuttgart, Germany

Re: Help needed testing 3.2.x Campaign games!

Post by alfred007 »

I updated to master aaacc29 and replayed Gamma 08. Though you can still not really do very much it's now very better. Now you have a chance to repair oil derricks or build new sensor turrets. But the game still doesn't end and I know why. It's because at the end there are still Nexus units on the map. I used the biffer baker cheat and destroyed the Nexus units and the game ended. Because of the AI bug you have called they tried to enter a hill from a position where it's impossible. But they were out of sensor reach so my artillery couldn't destroy them. We should change the winning condition that way, that you win after the satellites get destroyed no matter if there are still enemy units on the map or not.
I noticed one funny bug of the LasSat. I let a truck patrolling at the two oil derricks in the west. And the LasSat was not able to destroy the truck. The LasSat fired several times at the truck but even when it looked like the truck get hit the truck survived.
In the file attached are the logs of Gamma 08 and two saved games from the same moments as the saved games in my last post.
Attachments
gamma 08.zip
(183.52 KiB) Downloaded 100 times
Bethrezen
Regular
Regular
Posts: 661
Joined: 25 Sep 2009, 02:05

Re: Help needed testing 3.2.x Campaign games!

Post by Bethrezen »

Forgon wrote:
Bethrezen wrote:[...] as for the size issue looks like it may be an issue with loading old saves.
The savegame you attached stores a radar size of 32, which did not match my hypothesis. I have been unsuccessful in reproducing the tiny mini map in Beta 1 of this screenshot you provided.

To further investigate this, could you please assure me that the screenshot:
  • was created from the savegame you provided to me
  • was taken with the latest master version
  • was taken before resizing the radar
  • was taken with radar rotation disabled
As far as I can remember I updated to the latest master at the time
warzone2100-master-20180114-192518-6e805bd.exe (14-Jan-2018)

Loaded up the alpha beta interlude save which was made on an earlier master I then got my self organised as usually converting my units to tracked lancers and then loaded up the transport and began beta 01 and upon starting beta 01 the mini map was as it was in the screen shot I posted earlier, as I mentioned previously I think this may well of been caused by using an old save normally you can avoid issues like this by starting from a save from the previous level but it could well be that something changed between when i made the save and the master where i noted the issue I don't know.

I'm currently updating to the latest master.
warzone2100-master-20180121-182054-aaacc29.exe (21-Jan-2018)

Then I'm going to start from alpha 1 again and then see if the tiny mini map happens again if not then it was probably just a save issue because as it stand right now I have about a dozen different versions of master with different games at different points in the campaign.
The cause of blurry minimaps is poor utilisation of OpenGL. I am confident better looks could be achieved, but unwilling to promise a fix before the upcoming release.
that's fine because it's not a big problem it's just kind of annoying because it makes it harder to keep an eye on what is going on when the mini map is blurry like that.
User avatar
alfred007
Regular
Regular
Posts: 619
Joined: 31 Jul 2016, 06:25
Location: Stuttgart, Germany

Re: Help needed testing 3.2.x Campaign games!

Post by alfred007 »

Berserk Cyborg wrote:You forgot to add a left bracket for the "position_23" position coordinate. Other than that, it starts fine (it is in my campaignJS branch right now). Best of luck to what could be the most grueling mission ever. Wonder how long the power supply will last for Nexus... maybe we should queue it every five minutes to be 999999 again.
Small cause, big effect. Thank you for your help.

I adjusted some assemblies a little bit, the file is attached.

If we calculate for Nexus 2 units per minute with 10 factories are 20 units per minute. Let's be generous and say an average power of 1,000 means 20,000 needs of power per minute. And when we divide 999999 by 20,000 the power last 50 minutes. And I think a more realistic average would be 500 and then the power last 1:40 hour. To be sure that Nexus doesn't run out of power it would be enough to queue it every 30 minutes to 999999 again.
Attachments
labels_cam3-4.json
(3.88 KiB) Downloaded 107 times
User avatar
Berserk Cyborg
Code contributor
Code contributor
Posts: 938
Joined: 26 Sep 2016, 19:56

Re: Help needed testing 3.2.x Campaign games!

Post by Berserk Cyborg »

Pushed your Gamma 9 updates in be3858f35c470c52c5dc1c71e647620486eb59d3. I simply made the power reset every 5 minutes for Nexus.

You will notice that bases that could of had blips now show them instead of the one in the south-west base. There are still around three not used only because I saw no bases around the unused ones.

This mission had a few unused things so for the sake of history:
1. A time limit of three hours in a comment from the original script.
2. I think at one time this mission was going to have an "absorb" ability like in Gamma 5 as evident by two currently unused videos sequences for this mission. You are told take down the Nexus HQ to stop it. If you fail and are completely absorbed then Nexus laughs at you.
Bethrezen
Regular
Regular
Posts: 661
Joined: 25 Sep 2009, 02:05

Re: Help needed testing 3.2.x Campaign games!

Post by Bethrezen »

Ok so I'm in the process of updating the the general issue list and adding those other 3 issues, and i know that you fixed issue 21.) Players can incorrectly resume a failed mission even though the timer is at zero, however when double checking to make sure this was fully fixed, I noticed that you are still able to do this when in debug mode, so i wanted to ask if that was intentional ?
User avatar
alfred007
Regular
Regular
Posts: 619
Joined: 31 Jul 2016, 06:25
Location: Stuttgart, Germany

Re: Help needed testing 3.2.x Campaign games!

Post by alfred007 »

Finally, the storm on Nexus begins NOW!! :twisted:
User avatar
alfred007
Regular
Regular
Posts: 619
Joined: 31 Jul 2016, 06:25
Location: Stuttgart, Germany

Re: Help needed testing 3.2.x Campaign games!

Post by alfred007 »

With master be3858f the final mission of the campaign is completed. I needed one hour playtime to defeat Nexus. Everything worked fine, only the sound pcv440.ogg (Reinforcements are available) was not played after my transporter left the map.

I think there are now only two major issues left: the rearming of VTOLs and the transition from beta to gamma. After these two are fixed I'm of the opinion that we can release Version 3.3.0.
Attachments
logs gamma 09.zip
(7.08 KiB) Downloaded 107 times
User avatar
Berserk Cyborg
Code contributor
Code contributor
Posts: 938
Joined: 26 Sep 2016, 19:56

Re: Help needed testing 3.2.x Campaign games!

Post by Berserk Cyborg »

Pushed 1a05e388b60e930e6293a5922122c7b7beedfbe8. Adds the reinforcement sound in cam3-4, makes "Return to LZ" translatable whenever those are updated, fixes early LZ message for forced return to LZ and total annihilation, and made the Beta 8 base cleanups bigger to include a few structures that would be missed by base auto-removal.
Bethrezen wrote: however when double checking to make sure this was fully fixed, I noticed that you are still able to do this when in debug mode, so i wanted to ask if that was intentional ?
Kind of sounds like a cheat/debug feature. If it is wanted I do know what to remove to stop that.
alfred007 wrote: With master be3858f the final mission of the campaign is completed...

I think there are now only two major issues left: the rearming of VTOLs and the transition from beta to gamma. After these two are fixed I'm of the opinion that we can release Version 3.3.0.
Must not have been as hard as I thought? Script-wise, don't know what else there is to do, aside from some smaller AI quirks in a couple of missions.

I am on Beta 8 right now, so I will have a save on Beta 10 to upload to #3009 shortly for testing purposes.

Edit:
I have created a repo here you all can create known issues in for easier documentation of bugs and such solely for this testing thread. Probably should have made this long ago.
User avatar
alfred007
Regular
Regular
Posts: 619
Joined: 31 Jul 2016, 06:25
Location: Stuttgart, Germany

Re: Help needed testing 3.2.x Campaign games!

Post by alfred007 »

Berserk Cyborg wrote:Must not have been as hard as I thought? Script-wise, don't know what else there is to do, aside from some smaller AI quirks in a couple of missions.
It was tough, really tough. At the beginning, it was defending the LZ, in the middle small steps forward by destroying one factory after the other, later bigger steps. I had the advantage, that I had a squad with one commander and 20 Rail Gun, Tiger body hover tanks, all heroes. That made it easier to survive the Scourge VTOLs attacks. With that squad in front of my LZ, I repulsed all attacks and brought a lot of VTOLs in. With the VTOLs, I destroyed the two VTOL factories of Nexus. Of course, no VTOL of mine survived that. Then I brought the second squad like the first in and so I had one squad to move west and one to move south. I also used a lot of artillery that I assign to sensor towers. At the beginning, I avoid to move into the middle but moved around. Only when I had the SE factory destroyed I moved through the middle with on squad. The other one fought against the base in the NW. And at the end, both squads get united at the entry of the SW base. Then it was just a matter of time when Nexus get destroyed.

Maybe we can decrease one or two factory throttles, but this decision can wait until the next release.

What can I do now to help you?
User avatar
alfred007
Regular
Regular
Posts: 619
Joined: 31 Jul 2016, 06:25
Location: Stuttgart, Germany

Re: Help needed testing 3.2.x Campaign games!

Post by alfred007 »

After finishing the campaign I checked the Scavenger factory throttles and found that they range from 10 to 40 seconds. I think the Scav throttles should not be longer than 25 seconds. In the attached file are the .js files with changed throttles.
Alpha 02: Because there are three factories producing at the same time I set the longest throttle of 25 seconds.
Alpha 05: Only one Scav factory so I set the throttle down to 15 from 25 seconds
Alpha 06: There are three factories but not all three working at the same time. Throttles down from 30 to 20 seconds
Alpha 08: The south factory get destroyed fast but there are also two NP factories so throttles down to 20 from 35 seconds.
Attachments
Changed alpha throttles.zip
(8.3 KiB) Downloaded 102 times
User avatar
Berserk Cyborg
Code contributor
Code contributor
Posts: 938
Joined: 26 Sep 2016, 19:56

Re: Help needed testing 3.2.x Campaign games!

Post by Berserk Cyborg »

alfred007 wrote: What can I do now to help you?
Not sure, I was going to look into some of the AI misbehavior in Alpha/Gamma like regrouping and the retreat-repair in Gamma. For those I would just put in a mod for now since the current scripts in the repo seem to work without any major issues.

Then there would only be balancing (or even insane difficulty specific behavior for each mission) which is a big project in itself.
alfred007 wrote:After finishing the campaign I checked the Scavenger factory throttles and found that they range from 10 to 40 seconds. I think the Scav throttles should not be longer than 25 seconds. In the attached file are the .js files with changed throttles.
Must be why I never saw any scavenger attackers in Alpha 2? Another way to produce faster, I think, is to flood the AI research with extra manufacturing upgrades so that they produce practically within the limits of the throttle which could be have interesting results.

Edit:
Seeing as you were on the final mission map, did the Nexus heavy plasma launcher pit ever destroy itself? I know at times when I got to it the surrounding walls are missing, presumably because it hit the wall and destroyed itself and other structures.

Edit2:
Spoiler:
Post Reply