Latest Posts

Topic: Bugs

SirVer

Joined: 2009-02-19, 15:18
Posts: 1445
Ranking
One Elder of Players
Location: Germany - Munich
Posted at: 2016-12-15, 22:02

What a wonderful, delightful bug!! It sounded too weird, but it is an instance of a classical computer science problem: Cache invalidation is hard. I propose a fix in this branch which should hit trunk soon.

Edited: 2016-12-15, 22:03

Top Quote
simplypeachy

Joined: 2009-04-23, 12:42
Posts: 153
Ranking
At home in WL-forums
Location: Hampshire, UK
Posted at: 2016-12-18, 12:47

SirVer wrote: ... which should hit trunk soon.

A rimshot for the joker...only a Widelands player could laugh at that one face-wink.png


WARNING: New-style view packet not found. There may be strange effects regarding unseen areas.
_aD on IRC

Top Quote
WorldSavior
Avatar
Topic Opener
Joined: 2016-10-15, 04:10
Posts: 2091
OS: Linux
Version: Recent tournament version
Ranking
One Elder of Players
Location: Germany
Posted at: 2017-02-06, 02:24

Okay, some more bugs. I'm not registered at launchpad, so I post them here.

2 : Often you don't get an attack message. I'm almost sure that this bug didn't exist in build 18.

3 : Something goes wrong with the random generator which rules battles between soldiers. In the tournament there were games which have been heavily influenced by luck, for example the match of SirVer and me vs. king_of_nowhere and notabilis ( http://www.widelands.org/~sirver/wl/161121_tournament_2016_replays/Round06/ ) or the match between notabilis and waylon531 ( http://www.widelands.org/~sirver/wl/161121_tournament_2016_replays/Round05/ )

(Notabilis, what the hell are you doing? face-tongue.png )

4 : In the match between Tando and waylon531, Tando's headquarter was destroyed because a guardhouse close to it was conquered. I think that it's a bug, because a headquarter should have higher priority, like a fortress. (A guardhouse next to a fortress cannot be conquered, it will be always destroyed automaticly). To watch this situation, just watch a little bit of the time after the replay (after 230 minutes) http://www.widelands.org/~sirver/wl/161121_tournament_2016_replays/Round04/

5 : Crystal mines work like crazy. They ignore all economy settings. Not all the time, but it might be caused in situations when you build something. I think I've read something about a similar bug with granite mines which have been fixed (maybe).

I've discovered many further bugs, but it's not easy to remember all of them at a certain time face-wink.png

Edited: 2017-02-06, 02:28

Wanted to save the world, then I got widetracked

Top Quote
kaputtnik
Avatar
Joined: 2013-02-18, 20:48
Posts: 2433
OS: Archlinux
Version: current master
Ranking
One Elder of Players
Location: Germany
Posted at: 2017-02-06, 07:16

WorldSavior, on which version are you playing at the moment?


Fight simulator for Widelands:
https://wide-fighter.netlify.app/

Top Quote
WorldSavior
Avatar
Topic Opener
Joined: 2016-10-15, 04:10
Posts: 2091
OS: Linux
Version: Recent tournament version
Ranking
One Elder of Players
Location: Germany
Posted at: 2017-02-06, 12:53

Build 19. Are the bugs fixed at the newest inofficial version? face-wink.png

More bugs:

6 : A defeated player is not really defeated if he's building a new warehouse just in that moment. If it will be finished, the player can continue with almost everything and is an observer at the same time.

7 : The Barbarian Trainings camp does a lot of breaks if you don't use war helmets. It seems that the productivity cannot go beyond 75% or something like that.

Other trainingsites might be concerned, too. Ah, yes: The Labyrinth is similar if you don't use shields, or if you use just one shield.

It might be also possible that they just show a wrong productivity in that case, I don't know. But i believe that they just work slower because of unnecessary breaks.

Edited: 2017-02-06, 12:54

Wanted to save the world, then I got widetracked

Top Quote
kaputtnik
Avatar
Joined: 2013-02-18, 20:48
Posts: 2433
OS: Archlinux
Version: current master
Ranking
One Elder of Players
Location: Germany
Posted at: 2017-02-06, 19:46

WorldSavior wrote:

Build 19. Are the bugs fixed at the newest inofficial version? face-wink.png

Likely not all, but e.g. the missing attack sound is maybe fixed with the sound improvement change. I am not sure though. So the bugs you're report must all be tested with the current development version, too see if they are still there. That means additional, time consuming, work for the developers, which could be easily done by players, if they use the current development version face-wink.png


Fight simulator for Widelands:
https://wide-fighter.netlify.app/

Top Quote
king_of_nowhere
Avatar
Joined: 2014-09-15, 18:35
Posts: 1668
Ranking
One Elder of Players
Posted at: 2017-02-06, 21:07

productivity statistics are wrong. they store in memory the latest 20 work cycles, whether they were succesful or failed, and calculate the prooductivity from that. but they fail to keep into account that a failed cycle is shorter than a worked one. So for example if an idle cycle lasts 10 seconds and a working cycle lasts 20 seconds, with 10 worked and 10 skipped cycles the system will give 50% productivity, even though the building actually worked 200 seconds out of the last 300 seconds, so its real productivity is 66%. egardless of exact numbers, there is a systematic underestimation of the real productivity value. I pointed that out, but it was never fixed.

Edited: 2017-02-06, 21:07

Top Quote
Tibor

Joined: 2009-03-23, 23:24
Posts: 1377
Ranking
One Elder of Players
Location: Slovakia
Posted at: 2017-02-06, 22:30

I think the production cycle can have three outcomes each with own duration

  • success
  • fail because of lack of input wared
  • skip because output is not needed.

Current counter is simple - 20 bools and skipping is not even recorded there ( because of bools). So if output was not needed in last 10 minutes and thus nothing was produced, the old statistics is still shown over the building... This was discussed, but there is no consensus about what exactly should statistics show - more specifically: if skipped cycles has to be included there...


Top Quote
GunChleoc
Avatar
Joined: 2013-10-07, 15:56
Posts: 3324
Ranking
One Elder of Players
Location: RenderedRect
Posted at: 2017-02-07, 11:02

WorldSavior wrote:

Okay, some more bugs. I'm not registered at launchpad, so I post them here.

2 : Often you don't get an attack message. I'm almost sure that this bug didn't exist in build 18.

There are multiple messages related to attacks. which one of those do you mean exactly?

3 : Something goes wrong with the random generator which rules battles between soldiers.

Considering the 1.000 or so claims on the Battle for Wesnoth forum that the RNG is wrong that have been shown to be false, we can't trust human gut feeling here. I think we are just a library for the RNG, so it shouldn't be far wrong.

4 : In the match between Tando and waylon531, Tando's headquarter was destroyed because a guardhouse close to it was conquered. I think that it's a bug, because a headquarter should have higher priority, like a fortress. (A guardhouse next to a fortress cannot be conquered, it will be always destroyed automaticly). To watch this situation, just watch a little bit of the time after the replay (after 230 minutes) http://www.widelands.org/~sirver/wl/161121_tournament_2016_replays/Round04/

Might be a bug indeed.

5 : Crystal mines work like crazy. They ignore all economy settings. Not all the time, but it might be caused in situations when you build something. I think I've read something about a similar bug with granite mines which have been fixed (maybe).

If verified, definitely needs fixing.

6 : A defeated player is not really defeated if he's building a new warehouse just in that moment. If it will be finished, the player can continue with almost everything and is an observer at the same time.

The game considers him defeated, so in multiplayer, he should be an observer with no control over everything, and in single player be able to continue the game if he wishes. Not sure how we should handle this edge case.

7 : The Barbarian Trainings camp does a lot of breaks if you don't use war helmets. It seems that the productivity cannot go beyond 75% or something like that.

Other trainingsites might be concerned, too. Ah, yes: The Labyrinth is similar if you don't use shields, or if you use just one shield.

It might be also possible that they just show a wrong productivity in that case, I don't know. But i believe that they just work slower because of unnecessary breaks.

The Atlantean Steel Shield is used for the first program, the Barbarian War Helmet in the last one. Maybe this is just a skew in the statistics?


Busy indexing nil values

Top Quote
WorldSavior
Avatar
Topic Opener
Joined: 2016-10-15, 04:10
Posts: 2091
OS: Linux
Version: Recent tournament version
Ranking
One Elder of Players
Location: Germany
Posted at: 2017-02-08, 21:50

kaputtnik wrote:

WorldSavior wrote:

Build 19. Are the bugs fixed at the newest inofficial version? face-wink.png

Likely not all, but e.g. the missing attack sound is maybe fixed with the sound improvement change. I am not sure though. So the bugs you're report must all be tested with the current development version, too see if they are still there. That means additional, time consuming, work for the developers, which could be easily done by players, if they use the current development version face-wink.png

Okay. Maybe some other players can test if the bugs are already fixed ;-) (I'm not in the mood to do it)

king_of_nowhere wrote:

productivity statistics are wrong. they store in memory the latest 20 work cycles, whether they were succesful or failed, and calculate the prooductivity from that. but they fail to keep into account that a failed cycle is shorter than a worked one. So for example if an idle cycle lasts 10 seconds and a working cycle lasts 20 seconds, with 10 worked and 10 skipped cycles the system will give 50% productivity, even though the building actually worked 200 seconds out of the last 300 seconds, so its real productivity is 66%. egardless of exact numbers, there is a systematic underestimation of the real productivity value. I pointed that out, but it was never fixed.

That's a bug with high importance in my opinion. It's not really new to me... Nice that you've reported it...

But as I've said, I believe that the Barbarian trainings camp does breaks indeed (Bug 7). Just a feeling. (The Empire Trainings camp can show 100% even if you don't use golden armors, as far as I know. So there is a difference)

Tibor wrote:

I think the production cycle can have three outcomes each with own duration

  • success
  • fail because of lack of input wared
  • skip because output is not needed.

Current counter is simple - 20 bools and skipping is not even recorded there ( because of bools). So if output was not needed in last 10 minutes and thus nothing was produced, the old statistics is still shown over the building... This was discussed, but there is no consensus about what exactly should statistics show - more specifically: if skipped cycles has to be included there...

What's a bool? But I guess that it's right what you are thinking about it. And I think that it would make more sense that skipped cycles would be included. Otherwise you could be very confused about your economy. That's what I experienced ;-) And there would be a possibility to implement to show both kinds of productivity, or to chose one by preference...

GunChleoc wrote:

WorldSavior wrote:

Okay, some more bugs. I'm not registered at launchpad, so I post them here.

2 : Often you don't get an attack message. I'm almost sure that this bug didn't exist in build 18.

There are multiple messages related to attacks. which one of those do you mean exactly?

The message "you are under attack" if that certain building is under attack. By the way, the message "you are under attack" is always send if a soldier walks close to a miltary building, for example because of the random walk of lost soldiers face-wink.png

3 : Something goes wrong with the random generator which rules battles between soldiers.

Considering the 1.000 or so claims on the Battle for Wesnoth forum that the RNG is wrong that have been shown to be false, we can't trust human gut feeling here. I think we are just a library for the RNG, so it shouldn't be far wrong.

That is an indication, but not a prove face-wink.png

4 : In the match between Tando and waylon531, Tando's headquarter was destroyed because a guardhouse close to it was conquered. I think that it's a bug, because a headquarter should have higher priority, like a fortress. (A guardhouse next to a fortress cannot be conquered, it will be always destroyed automaticly). To watch this situation, just watch a little bit of the time after the replay (after 230 minutes) http://www.widelands.org/~sirver/wl/161121_tournament_2016_replays/Round04/

Might be a bug indeed.

5 : Crystal mines work like crazy. They ignore all economy settings. Not all the time, but it might be caused in situations when you build something. I think I've read something about a similar bug with granite mines which have been fixed (maybe).

If verified, definitely needs fixing.

I've already played matches were I had tons of crystal and diamonds in the warehouse (around 100 crystals, around 50 diamonds), more than stones (just 20 or something like that). I think that this is new in this build.

6 : A defeated player is not really defeated if he's building a new warehouse just in that moment. If it will be finished, the player can continue with almost everything and is an observer at the same time

The game considers him defeated, so in multiplayer, he should be an observer with no control over everything, and in single player be able to continue the game if he wishes. Not sure how we should handle this edge case.

I don't know what happens in multiplayer mode. I didn't know that there is a difference... It might be possible to consider a player as defeated only if he has no more chances to complete a new warehouse or port...

7 : The Barbarian Trainings camp does a lot of breaks if you don't use war helmets. It seems that the productivity cannot go beyond 75% or something like that.

Other trainingsites might be concerned, too. Ah, yes: The Labyrinth is similar if you don't use shields, or if you use just one shield.

It might be also possible that they just show a wrong productivity in that case, I don't know. But i believe that they just work slower because of unnecessary breaks.

The Atlantean Steel Shield is used for the first program, the Barbarian War Helmet in the last one. Maybe this is just a skew in the statistics?

Maybe it is. But I guess that that might be not the case... As I said, the empire trainings camp looks fine there even though their trainings-system is almost the same...

Edited: 2017-02-08, 21:56

Wanted to save the world, then I got widetracked

Top Quote