Page 5 of 8

Re: New release planning (3.2.4)

Posted: 10 Jan 2018, 01:48
by Per
I think I've found and fixed the rearm bug now. Please test to ensure that it is at least working as well as in 3.1.5. It should at least be easier to debug now, can usually use the trace feature (ctrl+l in cheat mode) to track VTOLs while flying to figure out what's going wrong.

Re: New release planning (3.2.4)

Posted: 10 Jan 2018, 06:48
by Berserk Cyborg
It does better, but there still is something off.

All of the VTOL units that are not going to a free pad show something like:

Code: Select all

[vtolReadyToRearm] rearm pad at 16,19 won't snatch us - we already are snatched by 22,18
The pad at (16, 19) was completely free while (22, 18) is of course already taken.

Re: New release planning (3.2.4)

Posted: 10 Jan 2018, 10:25
by Per
The rearm code is just bollocks, and is in dire need of a rewrite. The question is - is it still worse than 3.1.5? If not, I think we should revisit this after the release.

Re: New release planning (3.2.4)

Posted: 11 Jan 2018, 17:08
by Lord_Kane
Per wrote:The rearm code is just bollocks, and is in dire need of a rewrite. The question is - is it still worse than 3.1.5? If not, I think we should revisit this after the release.
Buildbot is dead.

Re: New release planning (3.2.4)

Posted: 12 Jan 2018, 09:42
by Forgon
Lord_Kane wrote:[...] Buildbot is dead.
Buildbot is now up-to-date again, although its Linux releases are still broken (see ticket #4702).

Re: New release planning (3.2.4)

Posted: 13 Jan 2018, 05:30
by vexed
After a very quick skimming, and since savegames would be broken again, it makes more sense to do a 3.3.0 release than a 3.2.4.

I also think a release should wait until the campaign issues are worked out, along with the other new crop of bugs that appeared like the VTOL issue, and also the balancing, since some changes are vastly different from 3.1.5 (which apparently is still the only version people are playing for MP battles.)

Re: New release planning (3.2.4)

Posted: 13 Jan 2018, 07:52
by Forgon
vexed wrote:After a very quick skimming, and since savegames would be broken again, it makes more sense to do a 3.3.0 release than a 3.2.4.

I also think a release should wait until the campaign issues are worked out, along with the other new crop of bugs that appeared like the VTOL issue, and also the balancing, since some changes are vastly different from 3.1.5 (which apparently is still the only version people are playing for MP battles.)
I agree.

Re: New release planning (3.2.4)

Posted: 13 Jan 2018, 14:33
by Per
What are the remaining balancing issues?

VTOLs - if there are further issues, I'll need a way to reproduce them (the best would be a savegame). To me behaviour looks like 3.1.5 at this point (although the 3.1.5 behaviour isn't great, either).

Re: New release planning (3.2.4)

Posted: 13 Jan 2018, 16:27
by alfred007
vexed wrote:After a very quick skimming, and since savegames would be broken again, it makes more sense to do a 3.3.0 release than a 3.2.4.
That's what I always said
vexed wrote:I also think a release should wait until the campaign issues are worked out, along with the other new crop of bugs that appeared like the VTOL issue, and also the balancing, since some changes are vastly different from 3.1.5 (which apparently is still the only version people are playing for MP battles.)
I think the only issue along with the VTOL rearming problem that should hold us back to release 3.3 is the transition problem beta/gamma. I don't know if maybe my saves are broken, but during several tests, the first transport in gamma mission did not include the units I loaded in but random units you get when you start gamma mission new from the menu. We have had this bug for a long time (#3009).
Balancing should us not hold back to release a new version because all the balancing issues we have are also in 3.1.5. Like overpowered assault guns and lancers.

At the moment Berserk Cyborg and I are working on Gamma 04 to make this level a bit more interesting because in the first tests it was possible to do a rush and win this lever in under 4 minutes. Because of the fact that it will take a lot of work to make this level really interesting, we will move on as soon as we have an acceptable solution. Bethrezen, Berserk Cyborg and I are also doing some work on balancing, at the moment with adjusting the base damage of lancers.
I'm clearly aware that as soon as Berserk Cyborg and I finished testing gamma mission and ticket #3009 is fixed we should release 3.3.
Everyone fell free to call a bug that he would like to be fixed before the release of 3.3.
All the balancing issues we have could better get fixed with more player playing the game. At the moment we are three testers and if you want to have the balancing issues solved before the next release you have to wait until 2019 I suppose.

Re: New release planning (3.2.4)

Posted: 13 Jan 2018, 16:51
by Berserk Cyborg
Per wrote: VTOLs - if there are further issues, I'll need a way to reproduce them (the best would be a savegame).
vtolRearming.zip
There are some pads spread around sk-rush. Sometimes a few go rearm after an attack run, other times all sit at one pad. Tell them to attack and it should show the weird behavior.

I was looking into why donateObject() fails to donate an HQ to Nexus in Gamma 5 and discovered that it thinks Nexus already owns an HQ on map (none should be on the map). It does not donate because of

Code: Select all

if (asStructureStats[statidx].curCount[player] + 1 > asStructureStats[statidx].upgrade[player].limit)
{
	return QScriptValue(false);
}
asStructureStats[statidx].curCount[player] is already 1. I do notice that since you changed structure limiting that the mission displays the assert about "there are too many X on map" when the mission loads. I think it is happening on home maps exclusively right now, though this issue had to have existed long before the structure limit change in the first place.

I would not let campaign balance stall a release. It has mutated and gotten worse along with time and the only way to fix it would be to start over completely.

Re: New release planning (3.2.4)

Posted: 14 Jan 2018, 06:10
by vexed
Per wrote:What are the remaining balancing issues?
Wasn't it changed to 100% accuracy for some weapons?

I am thinking that the MP crowd will be yelling about this, among other things they have yelled about in the past.

So, while it may not be a blocker, I wish we would have more feedback--actually, I'll just make another announcement thread to see if we can have more testers for MP stuff.

Re: New release planning (3.2.4)

Posted: 14 Jan 2018, 13:17
by Per
vexed wrote:Wasn't it changed to 100% accuracy for some weapons?
That was rolled back, along with most other balance changes, before the release of 3.2.

Re: New release planning (3.2.4)

Posted: 15 Jan 2018, 05:59
by vexed
Per wrote:
vexed wrote:Wasn't it changed to 100% accuracy for some weapons?
That was rolled back, along with most other balance changes, before the release of 3.2.
Saw viewtopic.php?f=42&t=13865&p=139616#p139615 and thought it was still the case by that comment, "but I'm still confident that precision works differently than in 3.1.5"

So... that still needs fixing / balancing or whatever you want to call it?

Re: New release planning (3.2.4)/ 3.3.0

Posted: 15 Jan 2018, 09:03
by Prot
I understand that HQ is no longer needed for battle.
First, it removed the dependence to research the machine-gun towers.
For now the design is made without HQ.
At the moment, HQ is only needed for the map-GUI, but not for logic or dependency researches.
Another useless structure as Command Relay Center.
And hello rush with towers and machinegues at the 1-2 minutes of the game.
If the design no longer requires a HQ, maybe to return to HQ the need for machine-gun towers?

Re: New release planning (3.2.4)/ 3.3.0

Posted: 15 Jan 2018, 20:24
by Per
Ah, the great HQ discussion. Now that is a trip down memory lane (and not a pleasant one). I had totally forgotten about all that. Long story short: It went missing somewhere during the CSV -> INI -> JSON conversions and the back and forth HQ changes we did before 3.1 was released. It does not seem to have been terribly missed? There is a ticket, but it seems to have been mistakenly closed: #3437 Since we restored the HQ to its old stupid self before 3.1, we should probably do the same before 3.3. (Although making techs building dependent in this manner is really counter-intuitive and newbie-hostile, since there are no ways to learn of this dependency in-game except by trial and error.)

You still need a HQ to make designs.

As for precision, I'll need something more accurate than "I'm still confident that precision works differently". The only precision related change I can find that still is in 3.2 is the merging of long and short accuracy, and that happened (uncontroversially) in 2012, before the release of 3.1.0, although it was not added to 3.1 branch at the time. Is there an actual problem?