TripleA 2.0.16244 issues with an unreleased game


  • Moderators

    This is a vague bug report not reproducible, as it is referring to a map I'm not making available at the moment. However, it can highlight issues that may be found otherwise, and I'm available to test myself a new version, in case anyone want to make some blind fixes (now the game is almost unplayable, because of all the frequent, and sometimes breaking, issues).

    I played 1 round of my new unreleased version of ex "270BC 40%", and these are the problems I experienced:

    • On Carthage attacking Salamantica, the battle window gave me the options to retreat or press, while the battle was already over, the territory being conquered. I clicked on retreat to Corduba, and nothing happened (so, everything worked correctly, except receiving a pointless and ineffective retreat prompt).

    • On RomanRepublic attacking Genua, I get both the "RomanRepublic select casualties" and "Retreat?" prompts open, instead of first getting the casualties selection, then the retreat option, as it should work. On that prompt, I kept clicking on "Remain", without selecting any casualties, that I should have already selected, instead. At the end, all worked apparently fine (maybe only because on the following round the neutrals missed all hits), but it was very confusing.

    • On GreekCityStates attacking Sardis, after I selected 2 psilos as casualties successfully, the select casualties prompt stays open, asking me to select two again, albeit nothing happens (sort of a ghost prompt without effect).

    • Again the same matter, GreekCityStates attacking Corinth, I take 1 psilos as casualty, but the select casualty prompt stays open, still listing 2 psilos as eligible casualties, while I already took out one, and on top of that prompt I get the retrat one. Here it is the screenshot:
      20191019_01.png
      This time I decide to take those ghost casualties, I was wrongly prompted to select again, and the battle window got completely stuck. Here it is the screenshot:
      20191019_02.png
      Now even my keyboard is stuck, and I cannot write anything down, so I reload the system, to get my functionalities back.

    I intended to playtest the game for a couple of rounds, but I had quite enough already after less than half a round, and didn't save it.


  • Admin

    I believe I ran across this once or twice, it's not an easy one to pinpoint. I think you can actually close one of the windows by clicking 'X' even though the buttons are inactive.

    @Cernel would you mind moving this to the bug tracker so we can consolidate it with the other pre-release problems?


  • Moderators

    And I can keep clicking on the "Retrat?" button to get infinite "Retreat?" prompts.
    20191020_01.png


  • Moderators

    And a completely empty retreat choice, in a battle where there were no retreat possibilities for any units (partial sea borne, but under v1 rules):
    20191020_02.png

    I was able to click my way to finish the battle by hitting the cross of the empty prompts several times.

    This is all I have in the Console:

    ott 20, 2019 7:12:48 AM games.strategy.triplea.ui.BattlePanel ensureBattleIsDisplayed
    GRAVE: battle not displayed, looking for:cdf141ff-876c-4160-96b7-b8b8bda93bec showing:null
    ott 20, 2019 7:13:09 AM games.strategy.triplea.ui.BattlePanel ensureBattleIsDisplayed
    GRAVE: battle not displayed, looking for:cdf141ff-876c-4160-96b7-b8b8bda93bec showing:null
    ott 20, 2019 7:13:29 AM games.strategy.triplea.ui.BattlePanel ensureBattleIsDisplayed
    GRAVE: battle not displayed, looking for:cdf141ff-876c-4160-96b7-b8b8bda93bec showing:null
    

  • Moderators

    I really wonder if I'm hitting something very strange, for some reasons, or just nobody else is using the prerelease at all, as I run into issues at almost every step. Maybe there is already a GitHub issue that is tracking these matters? All these things are basic stuff, that works fine in the current stable, and nothing has changes; why all the bugs?


  • Donators

    @Cernel said in TripleA 2.0.16244 issues with an unreleased game:

    ...or just nobody else is using the prerelease at all...


  • Admin

    @Cernel would you mind moving this to the bug tracker so we can consolidate it with the other pre-release problems?

    There are dev-level notes that woudl be good to add to this, would also like to see this get tagged with pre-release problem so we can see them all in one place, secondly would be great to be able to assign and mark as closed when completed. I am not asking we migrate this for no reason. I think it's probably time we move all 'bug-reports' in forums to a closed forum and leave just the one post that states to post bug-reports in the proper place the first time.

    There are a number of questions here as well, how often are you able to reproduce that problem @Cernel ? If you have a known way, please let us know, that is critical information. I took some time to test out retreat and it seems to work fine, are you seeing problems with it in any specific ways or just when the combat ordering gets unstuck and goes out of order?


  • Moderators

    @LaFayette I could open it in GitHub, but, as I said, nobody else has this game I'm playing. So, this was meant to be just a notification, until someone meets these same issues in an available game (I might eventually try to test this myself). Are you sure you want me to open an issue about a game nobody else has?


  • Admin

    While testing on WWII revised I saw a similar problem rarely and it seemed to maybe have gone away after a 'blind fix'. Basically a fix where you can't confirm you are fixing something but you think you see where the problem could be. Usually those are non-fixes and the problem still remains.

    Having a way to reproduce this problem is critical. The location of this bug is in a bad place of the code where it could be many things, there are some very subtle timing issues. I could imagine someone spending a couple weeks tracking just this down. I don't think anyone will have the 80 hours to dedicate to that, as an open source project that time investment could span several months. This is probably going to be the biggest bug in the 2.0 release, if we need a few months to track this down, 2.0 will be majorly delayed or we'll have to tell every user all about this and how they can bypass it until fixed (not ideal).


  • Admin

    @Cernel this is a very important bug, the most significant and dangerous for the 2.0 release. Are you able to consistently reproduce the problem in any manner?


  • Admin

    Bump @Cernel , are you seeing these errors consistently whenever you repeat the same attack sequence, or does it sometimes happen but other times not?


  • Moderators

    @LaFayette All the times, but I still haven't tested it on any actually available games, yet. Also, I'm not sure how to reproduce the only really critical one, getting the window completely stuck. I believe that happens only if you somehow click too many times somewhere.


  • Admin

    @Cernel I'm very interested in reproducing the problem locally to debug further. The XML to game-logic code should not be so brittle where such mechanical errors can occur, in other words, if you can build a map with reasonably valid XML, such problems should not happen. Worst case or at minimum, the fix would be specific to that map, best case, a deeper problem is revealed that would impact any map. If you can, please attach a zip to this thread for more investigation.


  • Moderators

    @LaFayette I'll see if I can reproduce it using some available maps, first.


  • Donators

    @Cernel I have been having a similar problem for about a month. But I never have been able to get a repeatable error. I can save the game, by pulling the window out of the way, but when I restart the save the error does not recur.


  • Admin

    Create a related bug-report in tracker: https://github.com/triplea-game/triplea/issues/5399 (ideal if we can have bugs land there directly so we do not have the conversations split). None-the-less, I remain quite concerned by this bug; it could be very difficult to find, without a known way to reproduce it, we're going to struggle to know if we've found it, let alone have fixed it. The possible surface area for a fix is several thousand lines of code, the main file that is responsible for this MustFightBattle.java is itself 5k lines. Given the high impact and uncertainties, this is an incredibly bad bug to be confronted with.


Log in to reply