Is TripleA finally dead
-
@beelee said in Is TripleA finally dead:
as all the veteran members use it.
Hopefully they stay on 2.5.
Cheers...
-
Not the first time, we are under this impression during the last years. It's a pain, indeed.
And, @beelee , we have discussed that before, and it is as you say.
I am not sure about sticking to 2.5, however. What do you guys think has been the latest 'working' 2.7-prerelease?
-
@panther
My 2p is use 2.7.15025 as suggested by @wc_sumpton here
https://forums.triplea-game.org/topic/4223/what-version-and-why/13This gives Capital centring and all the speed and benefits of 2.6 & 2.7
https://github.com/triplea-game/triplea/releases/tag/2.7.15025I have been using 2.7.15025 since wc_sumpton post, dated (25 Nov 25) , for AI testing, since it was so close to my prior test version 2.7.15032 it is stable enough for me.
Occasionally I dip in to the latest release, but keep going back to 2.7.15025
Get it here
https://github.com/triplea-game/triplea/releases/tag/2.7.15025 -
@thedog Thank you, however the issue regarding the providing of source code only had been fixed quickly. I am currently using 2.7.15287 and wonder if I could safely use a newer version / if we could recommend a newer version.
-
@panther
From me no, I have also been using 2.7.15256 but that is older than your version, so no help. -
As @TheDog has started, 2.7.15025 is the last version which had the 'new' centering worked.
If the 'broken' centering does not bother you then 2.7.15256 would be my next best guess.
2.7.15274 and 2.7.15275 had fatal error crashes withERROR o.t.game.client.HeadedGameRunner - Wrong thread (not Event Dispatch Thread).
2.7.15276 has the blank screen for AI players. Even though the screen seemed to 'catch up' on human players the AI player screen never seemed quite right. Did the 'fix' for 2.7.15273 create this screen issue? I don't know. As @TheDog has documented any AI game rum on the engine past version 2.7.15256 is problematic. What happened between 2.7.15256 and 2.7.15273 is unknown because the 17 changes were not released.
There have been so many releases since that gap that the root cause to this problem may be buried.
From 2.7.15326 until 2.7.15330 will run 'okey' as long there are no AI players starting the game. These issues, 'bad chance rolls', 'not enough resources', 'can't place', seem to happen only with a new game because a saved game plays just fine. @TheDog has even reported that the game breakingERROR o.t.game.client.HeadedGameRunner - Wrong thread (not Event Dispatch Threadhas returned.
This 'problem' has been ignored, and now 2.7.15333 will not load with more releases burying this.
Major problems need to be fixed. Releases should be halted until then. Otherwise TripleA might as well be buried, the engine has been broken since the beginning of November, maybe even longer. I know I'm tired of seeing and reporting the same issues over and over again. And I think @TheDog might be tired of it also.
And let's not talk about what improvements map makers are asking for, for maps like "North Africa" or "Stalingrad".
It's time to shut down development, kick it to the curb and let it die.Cheers...
-
Ok, so thank you @wc_sumpton for your substantiated recommendation , and thank you, @TheDog , too.
So it seems like 2.7.15025 or 15256 might be the best pick, at least to those who care about a 'reasonable' working AI.
-
@LaFayette @frigoref
Drawing your attention to this thread, above will be a hard read.
TLDR you are losing the goodwill of your testers.Please fix some of the outstanding 2.7 issues.
-
Thanks @thedog for reaching out to us!
Let me first apologize for the current disruption of the user experience. We are trying to improve TripleA and as it is still big and complex those kind of side effect might arise.
In addition, we are all very restricted with the time we can spend and appreciate any kind of support we can get so we can focus on the code. This includes reminders on major issues (including links to them) if they seem to have been forgotten about, prioritizing the issues listed on GitHub or simply summarizing a long thread discussion to the core information required for a developer to act on.
However, a supporting manner would be better than flaming about issues - even though the frustration might be high. TripleA is definitely NOT finally dead! (on the contrary, I am quite sure we can release 2.7 this year)PS: Putting such a statement / thread in a forum might even prevent new players to start playing and I assume no one here would like to see that happening.
-
@wc_sumpton Which issue of yours are you referring to?
-
13915 Which is the same as 13962. This problem has caused one warning message to be removed, 13961 and a PR to remove another warning message 13965.
My first question why are "Warning" messages being removed? They don't cause a game to crash, they provide information that there may be something in the xml causing these problems. PR 13961 should be rolled back.
Issues 13915 and 13962 are describing a loading problem with a new game where the first or all players are AI.
Removing "Warning" messages does not fix this problem.
A saved game always plays correctly. If the first player is not AI then the game plays correctly. If any player is not AI, then the game will correct its display at that players turn, and the game will play normally. @TheDog documented this, and the issue could be traced back to release 2.7.15273.14004 describes an issue with just running TripleA.
I understand about how much time developers have. I understand this is being done on their own free time. What I don't understand is why there have been continual releases when the display still presents bad information with no visible attempt to correct the bad visuals. And now the game won't launch and there are still more releases which cannot even be checked.
In short what happened to the visuals after 2.7.15256 and when can we get a release that launches. Bad visuals and a game that does not launch equal a dead game.
Cheers...
-
There are two pull requests which I request should be rolled back.
https://github.com/triplea-game/triplea/releases/tag/2.7.15352 and https://github.com/triplea-game/triplea/releases/tag/2.7.15317 as both PR state that removing their blocks of code to 'remove crash triggers' or an 'remove illegalState checks'.
Neither blocks causes a map to crash. So why are they being removed.
"1941 Global Command Decision" is causing a lot of problems, one where a trigger would give an AI player some units and deduct the value of these units, but the AI player visuals showed 0PUs. This would throw the 'Not enough resources' error. The AI player really had 139PUs enough to deduct the 26PUs. Why was the game not showing the proper amount of PUs for the AI player? No one can answer this question. So instead of trying to figure out the real problem, the message caused by this problem is remove.
Another problem came when the AI player tried to place units into territories it visually didn't own but had been given ownership of in the xml. Instead of trying to figure out why these territories were not changing properly, the error message is removed.@frigoref said in Is TripleA finally dead:
PS: Putting such a statement / thread in a forum might even prevent new players to start playing and I assume no one here would like to see that happening.
@frigoref said in Is TripleA finally dead:
However, a supporting manner would be better than flaming about issues - even though the frustration might be high. TripleA is definitely NOT finally dead! (on the contrary, I am quite sure we can release 2.7 this year)
As long as error messages are removed instead of problems fixed, then yes 2.7 should be able to launch. DOA
Cheers...
-
@wc_sumpton
I do not understand why you request the roll back of those PR, but you better discuss this with @LaFayette as he created them. Even if they are not solving the problem you have, they could still be legit.
Could you please provide a reproducible example for the problem(s) you are facing / refer to the GitHub issue in which this is described? Otherwise, it will be hard to answer the question "Why was the game not showing the proper amount of PUs for the AI player?".
PS: Please refrain from flaming further. I understand you are frustrated, but it is not helping to get things resolved again. We are all working here in good faith and the best intentions, there is not need for this. -
@wc_sumpton what is your desired goal here? Are you simply venting, or is there something you want to be done differently? I have trouble seeing how this thread is going to be productive, particularly when you have several misunderstandings.
-
@wc_sumpton & @beelee , cut it out with the insults. To fully respond would require quite an essay, which I'm not going to do. Suffice it to say, a number of things you said are wrong, I don't think maliciously so, just a misunderstanding of unrelated things feeling related.
Why was the game not showing the proper amount of PUs for the AI player? No one can answer this question. So instead of trying to figure out the real problem, the message caused by this problem is remove.
I'll respond to this. The reason seems to be because the AI-only trigger that grants PUs to the AI player is bugged and is not working. Why that is so, is not known. If it were, it would likely be fixed. I spent 4 hours on a Saturday looking at that issue instead of doing anything else in my life. It's heartening to now be told that clearly IDGF.
there were many errors, but all these errors could be traced back to the failure of the xml not loading correctly.
@wc_sumpton please clearly communicate your evidence for this. I suspect the XML was loaded correctly, but the conditions for firing the bugged trigger were not evaluated correctly. I believe it would be much more helpful to present why you think that was the case, in a clear and simple manner, compared to spending effort venting.
-
@lafayette I'm new to all this. Like really really new. So please excuse the question but: what do y'all need to get these games and the TripleA engine (those are two different things, right?) to work better?
Is the answer time and programmers? Or just programmers or just time?
Or management ?
Does it come down to Money? If the answers are: yes, no, maybe, and only on weekends, how could I and others help?And what is the scope of the problem? How much is the games themselves not working (like the game I am currently playing seems to use the computer as a sort of digital version of the board game really and not a fully functional computer-game at all with the players not even needing functionality.
Then there seem to me many actual games that contain various numbers of bugs.
Or are you guys talking about the TripleA "engine". and is that what one of you guys said you would proudly be releasing a version 2.7 of?
What might help here it to expand the discussion between current parties so that new parties might feel like they could come in and help.
For example, I don't really know but I get the feeling that some expert gamers are complaining to a part-time volunteer group of programmers in these recent pages.
But wouldn't it be nice to being having a different argument with the same goal? Such as how to justify the $2M we need to raise to blah blah blah and do do do.
In other words: It feels to me like all might be better if people here correctly articulate the bigger problem for us new people and people here to just "play games" and whoever and whatever the other side is here: maybe volunteer programmers.
Then talk about what to do to FULLY and CORRECTLY IDENTIFY the problems, then discuss HOW to get the resources needed to solve those problems. I Assume just hire them.
I'm a PhD in C.S. that has done much game development but nowadays would rather focus on finding ways to get money to funnel into the problem while others (who are getting the money) do the actual solution implementation.
I could see myself doing some programming if the correct structure for that is in place, but I get all hot and excited when I think about buying my game to improve via a programmer to fix the 12 bugs I am seeing in the game I love to play. So I then start thinking about money-raising ventures and even Crowd-Sourcing... coupled with some products often works very well by the way.
So what would it take for a random new guy like me to get the big picture on what is going on here and see if I, or many others, might be able to just brute force solve the issue with cash.
-
@johnnycat
The above is mainly about the TripleA 2.7 engine released as a stable.
TripleA operates on just a few hundred dollars for server use, each year.Depending on your time available, testing this version
https://github.com/triplea-game/triplea/releases/tag/2.7.15364
would be worthwhile as serious time/effort by the Devs has gone into fixing the issues above.There are about a dozen people, who regularly support/test/develop and we always need more.
The real meat of what happens to the TripleA engine happens on GitHub
see here
https://github.com/triplea-game/triplea/issues?q=is%3Aissue state%3Aopen label%3A2.7 label%3A"Major"and new Devs could pick from here
https://github.com/triplea-game/triplea/issues?q=is%3Aissue state%3Aopen label%3A2.7 label%3A"Good First Issue"These two lists are a small snapshot of what the Devs are/could be working on.
As the Devs are volunteers, that is skilled coders who do it for free, every time they post on this forum is less time they are developing, so please bear that in mind when they post. -
I have been off these forms for a few days, maybe cooler heads will prevail. It just seems to me that development is going one way, towards the release of 2.7, full steam ahead, regardless of what tester might say may be broken.
@beelee asked if there was a way to have maps still center on a player's capital when technologies and/or politics are turn off in the options settings. My suggestion was to have a "endTurnNoPU" added as the first step. I understood, at the this was not a clean answer, though it would be nice to have some extra settings for "endTurnNoPU" or "endTurn" to create this ability. This was not discussed, instead, testing and triggers were discussed to mimic/hack a clean "endTurnNoPU" for the centering purpose. @Cernel did not like this hack and raised an issue on GitHub. This issue was tag as a problem, and after further discussions @frigoref made a correction. A month later the correction was broken. It has been pointed out numerous times yet still reminds busted.
Developers are not listening when testers find problems.
Another issue this one regarding "Not Enough Resources. My suggestions for this issue were to change the message type since this error should not be a game breaker and have a discussion about how to convert the problem by the rules. The message was not changed, no discussion about the root problem. The issue was just closed.
Now we come to the problem that @TheDog is having with "1941 Global Command Decision". One that I was saying the whole time that there was a loss problem with an "All AI-Player" game. And your answer was, and I quote:@lafayette said in Is TripleA finally dead:
I'll respond to this. The reason seems to be because the AI-only trigger that grants PUs to the AI player is bugged and is not working. Why that is so, is not known. If it were, it would likely be fixed. I spent 4 hours on a Saturday looking at that issue instead of doing anything else in my life. It's heartening to now be told that clearly IDGF.
Do you know how many hours I've spent on this problem? Did you ever consider changing the message type, as I suggested even in your PR? Something I have be suggesting for a while.
Changing the message to a "Log.warn" would have given this:
The warning does not mess up game execution and is posted in the TripleA error log file. Also, because execution is not hurt, the other, removed/re-added her, error does not fire and this issue does not continue to show. A little time and effort, but, as you clearly stated YDGF!
I'm not stopping here, let's talk about good ideas bad implementation.
Giving SBR capabilities to non-isAir units. Good idea. What happens to these units afterwards? The implementation is so bad that map makers use hacks to circumvent to idea.
Giving units AA capabilities. Good idea. But there has been no discussion on how those rules have been added, and their effects to the rules regarding AA attacks/first strike attacks vs normal attacks.
I'm going to stop here. When things are broken, like the centering problem, and not quickly corrected. When suggestions and ideas are ignored that may help. I'm the one that feels like he's PITW.
Thank you for listening. I feel like I've tried to help only to be ignored. Good luck on getting 2.7 out. Me, I'm just going to sign off for a while.
Cheers...
-
Responding to @wc_sumpton
It just seems to me that development is going one way, towards the release of 2.7, full steam ahead
Yes, why should we do anything else? All efforts become available in future releases & future releases are blocked right now. The best thing to do is get 2.7 out so we can have peoples work see the light of day. I don't see anything unreasonable here.
regardless of what tester might say may be broken.
There is a triage process. The goal is to be sure that 2.7 is no worse than 2.5. It does matter what the issue is. It's a bad idea when you try to boil the ocean.
FWIW, 2.7 will be the last release on our "old release model". After 2.7, every update is going to be a release. Development is a lot more difficult bouncing between 2.5 and 2.7, so being on the newer version will make things just a lot easier. Any patch fixes are going to be a lot faster, they'll be released immediately.
The idea is to not boil the ocean... Get 2.7 to at least be no worse than 2.5 and working. Then from there we can do the patch fixes in a rapid manner, releasing them as soon as they are available. Again, the goal is not to fix everything single thing right now otherwise we'll get nothing released.
There is a cost in our current configuration as well. By popular demand we have 2.7 bots running, 2.7 costs about $15/month right now. Having two sets of deployed applications, managing the legacy 2.5 stuff is all time consuming. 2.5 is really hard to manage right now, it's holding us back on some really important security patches.. The database version that 2.5 uses is so old, it's hard to even find - let alone get running...
@beelee asked if there was a way to have maps still center on a player's capital when technologies and/or politics are turn off in the options settings. ... This issue was tag as a problem, and after further discussions @frigoref made a correction. A month later the correction was broken. It has been pointed out numerous times yet still reminds busted.
That issue has been on my radar. It's sounds like a regression and something worth trying to fix. One thing that does not help is lumping multiple problems in long issue statements.
I've been personally focused on working on the server side components. We need the server side stuff to be done, and nobody else is willing/wanting/able to work on that. My thought is to finish that out, after which if the issue of not centering has not yet been fixed by someone else, then I'll look into that issue then. It's simply a matter of priorities, it's a priority, but behind the server side work for me (which fully blocks 2.7, the server stuff must be done)
Developers are not listening when testers find problems.
I don't like that perception. But, it's your perception and I can't tell you that you must feel otherwise. I feel like the issue is more, I'm just not responding to everything right away. I spend a lot of time responding to many threads, sometimes so much so that I'll spend literally 3 hours just responding to everyone, get nothing actually done, then a few weeks later need to respond about why things are still not done, still broken, etc..
Overall, really trying to bucket things into 3 lists:
- must-fix or else 2.7 cannot be released
- fix as soon as 2.7 is released
- important fixes, things that have been lingering and are higher up on the list.
Whether something lands in list #2 or #3 does not mean that fix is never going to happen. For example, if I could do one fix every day - if we need 5 fixes and then 2.7 would be launched at the end of the week. Though, if we could do the two must-fix issues, then 2.7 is launched on Tuesday, then bug fixes will land and be released every following day. It's the same amount of time overall, but in the latter case 2.7 is launched, a bunch of other things become easier, we start saving money, we start the user migration, and we still will get the same fixes on the same schedule. Now, in reality, it is sometimes a week or two before I get a block of 4 hours where I"m able and willing to do TripleA coding. So it's more one fix per week or month. In which case a launch that is 2 months out vs 5 months is a big difference.
So, the cut-throat prioritization of 2.7 is just triage. It does not mean we're moving things to a "never fix" list. Time is finite and zero-sum, just a question of what to do first and then what next.
Another issue this one regarding "Not Enough Resources. My suggestions for this issue were to change the message type since this error should not be a game breaker and have a discussion about how to convert the problem by the rules. The message was not changed, no discussion about the root problem
Except, the error was breaking to the game. The game continued, but in an abnormal state. You may recall there were a number of other issues that cropped up, those were seemingly related. Just because the game continues, does not mean it is healthy. The error pop-up was actually being triggered thousands of times per minute, just those other errors were masked as we only show the same pop-up just once. Simply changing the messaging type would not resolve that the error aborted a bunch of code execution that needed to happen.
The message was not changed, no discussion about the root problem
There was quite a bit of discussion around the root problem. There were two:
(1) An AI only trigger was not firing, which meant that we were subtracting PUs from zero
(2) The code threw a hard error whenever PUs would go negative.The problem #2 is a design flaw. It means you can't have a trigger that simply says "subtract 1000 PUs each turn". The fact any map worked with a trigger like that was happenstance. Initially all countries have plenty of PUs, but you could easily run into an error late game when a country has 4 PUs, just a couple territories, and we try to subtract 5. So, #2 is a design flaw..
(1) is still an open issue and is being considered a 2.7 release blocker... I don't know what more you would want other than for that to be considered a top priority and a regression. Though.. it does look like the 2.7 AI only triggers are really a new feature, so I'm not sure if we really should delay 2.7 for something that did not even exist in 2.5.
spent 4 hours on a Saturday looking at that issue instead of doing anything else in my life. It's heartening to now be told that clearly IDGF.Do you know how many hours I've spent on this problem?
I don't. But I am not accusing you of not caring. I was accused of not giving a fuck and being on a power trip. I stated the 4 hours to indicate how much effort this all takes. Which is also why I've tried to make it so we can be more efficient in how we handle and communicate about issues. So, that means, not having issues where there is 3000 words of text to read before you can even start programming. The goal is one bug per issue, one clear definition of what that bug is, how to recreate it (spelled out very precisely so there is no question what the problem is, it can be obvious to one person when looking at a problem, but not obvious to another when they read), and all in all issues that can be split out and delegated and not require the core maintainers to do everything.
Did you ever consider changing the message type, as I suggested even in your PR?
Yes, I think I have responded 3 or 4 times now why changing the message type is not a fix. The fix was to remove the error entirely. The game engine was going out of its way to crash the current thread when PUs would go negative. The game engine did not account for a trigger that subtracts a fixed amount of PUs every turn. That situation should just be normal game operation..
The warning does not mess up game execution and is posted in the TripleA error log file. Also, because execution is not hurt, the other, removed/re-added her, error does not fire and this issue does not continue to show. A little time and effort, but, as you clearly stated YDGF!
Okay, if we were to change the explicit exception into a warning message, why should a user care? What is the user to do to fix this? How can a map maker fix this problem? Why should one country see this issue every turn? The warning message says, "if this problem happens frequently and you cannot fix it, please report it". Since a person cannot fix it, and it would happen frequently, then we are telling the user to report the problem. That does not seem like a fix. The 'warning' messages are meant to be more things that are map problems rather than engine problems.
Bottom line, why should this even be a "problem?" I know the background around such a check was actually to try and detect bugs that did things like subtract PUs twice. Those checks to try and detect bugs were themselves buggy, as it did not account for triggers that subtract PUs.
I'm going to stop here. When things are broken, like the centering problem, and not quickly corrected. When suggestions and ideas are ignored that may help. I'm the one that feels like he's PITW.
So.. in the last few months I had to do some overtime work and have been worried about my job. If I do a 60 hour work week, and then work a Saturday, and simply don't respond to TripleA for two weeks - it's not an "ignored" due to maliciousness, just more lack of time. I try to pick and choose where I respond and other things inevitably fall through the cracks. I spent a lot of December getting the servers into shape, about 90 hours in December, and you're seriously yelling at me about "map centering" not getting a quick fix (from me)? I think we're all trying pretty hard, and it's all labor intensive. Let's cut each other some slack.
-
@johnnycat you raise good questions:
Is the answer time and programmers? Or just programmers or just time?
Or management ?
Does it come down to Money? If the answers are: yes, no, maybe, and only on weekends, how could I and others help?Time and programmers is the biggest need. Every bug fix and issue probably takes 2 to 16 hours to fix. This is the biggest gap. Further, TripleA is a larger code base that is very brittle. Making that codebase be easier to work with would help.
Money is useful for keeping the servers running and paying the monthly bills. As much as I would love to be payed for TripleA work, paying any type of reasonable developer wage becomes way too much very quickly.
Additional management would be helpful. Getting issues sorted, cleanly stated, boiled down to a bare minimum would all be helpful.
Hello! It looks like you're interested in this conversation, but you don't have an account yet.
Getting fed up of having to scroll through the same posts each visit? When you register for an account, you'll always come back to exactly where you were before, and choose to be notified of new replies (either via email, or push notification). You'll also be able to save bookmarks and upvote posts to show your appreciation to other community members.
With your input, this post could be even better 💗
Register Login