World Conquest II

Discussion of all aspects of multiplayer development: unit balancing, map development, server development, and so forth.

Moderator: Forum Moderators

Post Reply
Dummy991
Posts: 2
Joined: November 25th, 2014, 1:05 am

Re: World Conquest II

Post by Dummy991 »

I encountered a bug where when you resume a multiplayer save in WC, the person who hosts is given control of the AI.
User avatar
tekelili
Posts: 1039
Joined: August 19th, 2009, 9:28 pm

Re: World Conquest II

Post by tekelili »

Dummy991: Not my bug. You could have known about it reading 1st post of this thread
Bugs:
Using planing mode with recruits will fail, as planning dont detect recruit replace.
There are not other bugs currently confirmed for last 1.11 developed version...
However there are still some engine bugs in BfW 1.11.18 that affect campaign, the most noticiable is that reload game gives AI sides control to player and have to use "droid x" command for each side. Host will also have some problems in share vision with allies after reload and droid AI sides.
BfW 1.11 also gives often "random" oos. Universal way to solve oos is an inmediate save reload (prefered done by host)
I have already reported this bug, but for some reason they decided call 1.12 stable without fix it :whistle:
Be aware English is not my first language and I could have explained bad myself using wrong or just invented words.
World Conquest II
User avatar
iceiceice
Posts: 1056
Joined: August 23rd, 2013, 2:10 am

Re: World Conquest II

Post by iceiceice »

tekelili: The bug report you made on forums doesn't really count, no one will look at those pages again most likely. Any bug that is important enough to be a blocker goes on the bug tracker. The bug report which I made on the tracker after I reproduced your report is here: https://gna.org/bugs/?22328

thunderstruck never responded unfortunately, but it doesn't mean he didn't look at it at some point. Note that since I could not consistently reproduce it it makes it much harder to work on, since whoever works on it may never know if they fixed it.

I want to comment also regarding the "shared vision":
thunderstruck had some new ideas in his gsoc about how the mp setup should work. One of the changes is that the "mp connect" mechanism always is run, even in cases where you transition the next scenario without going to the setup screen. This is done to make it easier to know what the game configs will look like when we actually try to run a game on them. Because of this, in a mp campaign you almost surely want to have "force_lock_settings=true" -- otherwise the mp connect may use some internal logic to change the configuration of the sides. This is done basically to make things more user-friendly when the setup screen is showed in between levels, you would have to ask thunderstruck for the exact reasons. I found that when I ported RBY, I needed to add "force_lock_settings=true" to all the "2nd" scenarios or I could not get the fog / shroud settings to work right.
User avatar
tekelili
Posts: 1039
Joined: August 19th, 2009, 9:28 pm

Re: World Conquest II

Post by tekelili »

iceiceice: I did also a report on gna #22968

I think is a quite easy to reproduce bug, it happens always in any multiplayer game ;)

Edit: and talking about bugs that affect WC II I also would like #22373 were fixed. Currently some villages are totally covering flag and you can not know owner.

Edit2: Also note both bugs above are easyly reproduced. There is another severe bug I have spotted a lot of times: Some players moves will randomly fire shroud_data oos (harmless) and it looks at some random moment they become harmfull oos. But as it looks in order to report I should figure out how to reproduce it... I didnt care to report it :roll:
Be aware English is not my first language and I could have explained bad myself using wrong or just invented words.
World Conquest II
gfgtdf
Developer
Posts: 1432
Joined: February 10th, 2013, 2:25 pm

Re: World Conquest II

Post by gfgtdf »

"shroud_data oos" is not harmless. The shroud data is used to know how long a units move is (when it will be stopped becaue you have sighed an enemy unit under the shroud). If the movement calcuations missmatch on the clients, the units will be at different positions.
Scenario with Robots SP scenario (1.11/1.12), allows you to build your units with components, PYR No preperation turn 1.12 mp-mod that allows you to select your units immideately after the game begins.
User avatar
iceiceice
Posts: 1056
Joined: August 23rd, 2013, 2:10 am

Re: World Conquest II

Post by iceiceice »

tekelili wrote: I have already reported this bug, but for some reason they decided call 1.12 stable without fix it :whistle:
tekelili: I'm just trying to explain to you one reason why this happened.

People report many bugs in wesnoth ranging from "semicolon is missing" to "program suddenly closed". Sometimes these reports are good and sometimes they are useless, and they report in many different channels -- sometimes in bug tracker, sometimes in technical report forum, sometimes just in comments on release thread "here's some problems I noticed [laundry list]". When 1.12 was released we made a list of what bugs we considered "blockers" which we must address before releasing. Usually only bugs that are reported on the bug tracker might be considered blockers. It is too much work to ask the release manager team and other folks to read every single forum thread looking for bug reports. It's easier to assume that any bug that is severe enough to be a blocker is on the bug tracker, because everyone knows that we can't really keep track of reports that aren't in the tracker.

There was no doubt that mp campaigns were going to be very buggy in 1.12.0 even compared to 1.10. The problem is that the release cycle took way too long -- people 2 and even 3 years ago pushed really big changes to things like, how save files work, among many other things, and these changes broke everything. These changes were committed in gigantic lumps with no documentation or explanation, and these people took no responsibility to fix the problems. When I started testing in january of this year, I could not even transition any mp campaign without immediately crashing. If gfgtdf and I did not work on it, then I doubt that we would have supported mp campaigns at all in 1.12, because no one else took responsibility. In my view, "blocker" means "the program crashed without warning" or "the game and all my save files became corrupted so that I could not continue and all my time was lost". For me, avoiding that was the goal for 1.12.0. It's not even sure that we achieved it, however I'm quite sure that there were no bug reports like this on the tracker at the time that 1.12 was declared stable.

Note also that most mp-related bugs for instance that SlowThinker reported in 1.10 were in fact fixed in 1.12. It's only the new bugs that we didn't find and didn't get reported that didn't get fixed. Edit: And that weren't considered blockers, there are some reports like the name thing that were known but that we didn't figure out and didn't consider important enough to block 1.12.0 release.
User avatar
tekelili
Posts: 1039
Joined: August 19th, 2009, 9:28 pm

Re: World Conquest II

Post by tekelili »

iceiceice: I accept your explanation and thanks for take a moment to write it. I used a vague "some reason" because understand developers can be involved in problems unknown for me; and I one more time: thanks for work done by free :D

I just hope BfW learned from past mistakes, and in future, large pieces of code undocumented dont become admited. I am not a professional coder, but I studied for a year informatic in a university and one thing I learned fast is that a good code is not a code that "works", but one that is also modular, documented, and maintanable.
Be aware English is not my first language and I could have explained bad myself using wrong or just invented words.
World Conquest II
User avatar
iceiceice
Posts: 1056
Joined: August 23rd, 2013, 2:10 am

Re: World Conquest II

Post by iceiceice »

tekelili: Yeah... Part of the problem also is that there are so many different kinds of games, and so many different modes they can be played in now, with or without random generation, and so many different kinds of save files, and the whole business about players leaving and joining, or observers taking a role... it's ultimately just too much stuff to test all the combinations manually, and any combination or the whole thing can break at any point.

And I don't really like to fix bugs if I think the same thing will just be broken again in a year or two... it feels like a waste of time.

So for me at least I think the way to fix is with automatic tests. I have been working to make bots that can automatically host games and join them... I finally figured out what I think is a good design, and it seems to be working now. https://github.com/wesnoth/wesnoth/pull/338

It's still not finished in that they cannot report the results of the tests yet, and also they cannot save the game so they cannot test save reloading yet. But maybe the whole thing will be working soon, and then we can do "bugfix + test" for all the 1.12 mp problems.
User avatar
Slann
Posts: 66
Joined: March 2nd, 2008, 3:47 pm

Re: World Conquest II

Post by Slann »

Does anyone have replays of this campaign? I cannot understand how could this be beaten. I've been playing it with more people and we rarely advance to map 2. I'm not a pro, but i've been playing this for almost 10 years. I'm really interested in see how do you survive through the enemies.
User avatar
tekelili
Posts: 1039
Joined: August 19th, 2009, 9:28 pm

Re: World Conquest II

Post by tekelili »

Slann wrote:Does anyone have replays of this campaign? I cannot understand how could this be beaten. I've been playing it with more people and we rarely advance to map 2. I'm not a pro, but i've been playing this for almost 10 years. I'm really interested in see how do you survive through the enemies.
I dont have anyone right now, but as soon as I have ones of a game reaching last map I will post them for you. It can take some weeks however, I need find time to play and remember dont let them get deleted by autosaves while testing. Anyway check "warning for average players" in 1st post, maybe you are selecting too high difficulty? I always play 3 players at nightmare with no save/reloads for RNG, and have beaten it some times... but seriously... my mates and me are mostly beign pawned, so a 0% victory rate for less skilled players can be expected at that mode ;)
Be aware English is not my first language and I could have explained bad myself using wrong or just invented words.
World Conquest II
User avatar
Slann
Posts: 66
Joined: March 2nd, 2008, 3:47 pm

Re: World Conquest II

Post by Slann »

Right, im not complaining. I knew it was really hard, just i wanted to see how good players can deal with the campaign.
Dummy991
Posts: 2
Joined: November 25th, 2014, 1:05 am

Re: World Conquest II

Post by Dummy991 »

Welp, I live up to my name.

This time I might have actually found something wrong. I noticed on the third map (I'm playing peasant difficulty) that the pairs of recruits aren't rotating. So I'm playing the Guild and I can't recruit skeletons because it is stuck on Wose. Likewise, I can't recruit elvish archers because it is stuck on skeleton archers, etc. I've tried opening the save in the main menu and the local game menu.
User avatar
tekelili
Posts: 1039
Joined: August 19th, 2009, 9:28 pm

Re: World Conquest II

Post by tekelili »

Dummy991: I have experimented similar bug in past (but not seen it since time ago). It always was due to a save/reload in turn 1 that made some kind of oos between savegame and game variables. If you have autosave of last turn at map2 maybe reload it will solve your problem.

If that is not case, then post savegame and I will try find out what happened. Relevant information I need is WC II version used and if you upgraded version in middle of campaign.

Edit: I did a quick answer while waiting my turn and didnt think carefully. Now I doubt is due to a save reload on turn 1 (that usually caused empty recruit list). If pairs are not rotaing means era event is not being fired o trying to use variables with different name, and in both cases most probable cause is campaign version was upgraded between scenarios (not recomended). Anyway, I have played games hosted by other players and campaign worked correctly in all scenarios, so definitively is a bug in the save. I hope can give you a more precise explanation after examine save.
Be aware English is not my first language and I could have explained bad myself using wrong or just invented words.
World Conquest II
User avatar
tekelili
Posts: 1039
Joined: August 19th, 2009, 9:28 pm

Re: World Conquest II

Post by tekelili »

Slann wrote:Does anyone have replays of this campaign? I cannot understand how could this be beaten. I've been playing it with more people and we rarely advance to map 2. I'm not a pro, but i've been playing this for almost 10 years. I'm really interested in see how do you survive through the enemies.
I attach a compressed file with 5 replays where, paso, Honor and me, reached last map in Nightmare this week. Ignore oos in second scenario, is harmless and is due I created replay from last turn autosave. Btw, after this game, I will probably nerf AI in last map a little :augh:
Attachments
WC_II_Demo_replay.rar
(1.44 MiB) Downloaded 306 times
Be aware English is not my first language and I could have explained bad myself using wrong or just invented words.
World Conquest II
User avatar
Slann
Posts: 66
Joined: March 2nd, 2008, 3:47 pm

Re: World Conquest II

Post by Slann »

Thanks, i will watch them now
Post Reply