Jump to content


Lobby Moderators
  • Posts

  • Joined

  • Days Won


Everything posted by Norse_Harold

  1. Indeed. I've installed kush-extreme version 0.26.2 via the in-game interface. Thanks for developing this, and thanks to the WFG team for getting it signed. freyyja, TwainRick and I were able to easily beat 5 "Very Hard, Aggressive" vanilla bots that were using the Gauls civ in alpha 26 rc3, so the bots needed an increase in difficulty. A bot with "Hard" difficulty using kush-extreme is much more difficult. I was able to survive, but it took 24 minutes for me to obtain a battering ram. Maybe I'm not exploiting the bot's low intelligence enough by building diversionary structures, because the bot was wiping out most of my army with each attack. This is good, because it increases the skill ceiling of the bots. It kind of reminds me of Jebel Barkal, where the opposition Kushites civ ("Napata") has quite strong units at mid game, depending on difficulty setting. Is that part of the inspiration for the kushites-extreme mod?
  2. No, it's not what I'm saying. That sounds like a different issue from what I'm observing. What I'm saying is that the game captures the cursor every time I startup the game and/or toggle it to full screen mode. However, it doesn't always keep the cursor captured when the mouse is at the "junction" between two monitors. I'm not using Wayland for X rendering.
  3. Resolved: See the update at the end of this post for a description of how this problem was resolved. There is still a bug where the mouse jumps to a different monitor when scrolling the map. This didn't happen with alpha 0.0.25. Note that I've only tested the release candidates. Steps to reproduce: 1. Ensure that two monitors are connected to your Linux computer, one is 1920x1080, the other is 1280x1024 max resolution. 2. Determine where to move the mouse in order to cause the cursor to appear at the second monitor. Let's call this area the "junction" between the monitors. It's probably the left or right edge of the primary monitor. 3. Move the mouse cursor to the secondary monitor, configure that monitor as the "primary" monitor with an xrandr command that toggles the primary monitor. See the attached shell script for that xrandr command. 3. Start 0ad alpha 26, ensure that it is configured to full screen instead of windowed, begin a game simulation, move the mouse to the edge of the screen, at the "junction" between the monitors, in order to scroll the camera frame, also known as scrolling the map. Stop scrolling, then resume scrolling. Continue alternating this, perhaps also varying the speed that you move the mouse. Expected results: the camera frame scrolls. Actual results: usually the camera frame scrolls. Sometimes the cursor jumps to the other monitor and the camera frame stops scrolling. It's necessary to move the cursor back to the monitor where 0ad is displayed in order to regain the ability to scroll the camera frame. The problem is intermittent. I find that playing the game normally for about 10 minutes, scrolling at the junction occasionally, makes the problem evident. It may be possible to reproduce the symptoms with simpler steps. Update: The problem was caused by the SDL library version that I was using. It was fixed by changing the version of the SDL library that I had installed. Discussion here. set-primary-monitor.sh
  4. Binaries for Linux of these versions of 0ad are available, so compiling from source might not be necessary in this case. 0.0.26 rc3 (rev 27067) as Snap (change version to latest/beta) rev 27077 as AppImage rev 27087 as Snap (change version to latest/edge)
  5. You're right, banning leaked or "pwned" accounts is not the best way to solve the problem. However, it's the only way to maintain security and an effective reputation-based justice system until email addresses are collected, unless WFG wants to start using password reset questions and answers, but that has its own pitfalls such as easily guessed answers. It's impossible to ensure that the original account holder resets the password if the only requirement for resetting the password is the leaked password. Seems like another "legitimate interest" in collecting an email address, for GDPR compliance purposes, specifically for password reset capability. If Linux-loving libertarians think that collection of email addresses is so onerous, then consider making it optional and give incentives such as access to matches that are restricted to "verified" accounts only.
  6. Thanks, Dunedan. @smileyIt seems like you have rushed the process of reading through this thread. I don't know why you're so focused on anti-smurfing, because I'm not. Instead, I'm focused on preventing all player misconduct, including extreme verbal abuse, cheating, griefing, ddosing, etc., by making bans effective. Let's own the problem. Preventing smurfing, defined as unknown players who lie that their skill level is much lower than it actually is, can primarily be done by hosters. But, it is also necessary for lobby moderators to shut down compromised accounts so that a decentralized reputation-based justice system, which is to some extent already established by hosters, can be effective. I don't get this point. What difference does this have from just regular smurfing which is so frowned upon? And if users know the verified account, how does it help with avoiding targeting? Either the actual identity is known, or its not. Okay, let's get this straight. First, I assume that when you use the word "smurfing" you're talking about "duplicate accounts" instead of smurfing. Then, the idea for aliases would involve making the primary identity known only to administrators. This way, it's an effective alias for the average user, but administrators are able to apply effective rule enforcement such as a ban or revocation of "verified" status. If necessary then consider an alternative idea of providing a web-based interface with details on each lobby account that aren't spoofable. Details could include account age, true ID (necessary because lobby pseudonym is currently changeable to anything, which I think should be restricted), rated match history with names, dates and durations of rated matches and their outcomes. And, consider taking inspiration from an idea from the professionals. Valve's Steam service publicly displays past aliases that each account has used and prevents changing alias too frequently. Mutations of this idea could also be useful, such as revealing the primary "verified" username of the player after a match ends. This would allow players to create aliases in order to avoid strategic targeting in-game by other players, but still allow a reputation-based justice system to function. These types of information help to detect artificial rating inflation through thrown matches and show how accurate the rating likely is based on the skill level of opponents. This information would allow hosters, and games limited to "verified" users, to require that new accounts, for example, have played several legitimate rated matches before they can be allowed into "verified" matches. This would increase the cost of creating a new fake account for the problematic users, thereby helping to make bans effective. Other measures would also be useful for this, as I have listed earlier. Please go back and re-read this post. Do you see what my goals are here, now? Which of these goals do you share?
  7. I think this might prevent friends from active players at institutions like schools or universities to join. Where there's a will, there's a way. How about adjusting the policy so that it's flexible? We could require such users to fill out a form requesting an override to the new account creation rate limit. Then an admin reviews the application and allows on a case-by-case basis. Or, simply rely on the "unverified" status for new accounts created, and set the limit for new accounts created per IP address to something like 10 per month.
  8. I'd like to There is just nothing I tried in the past three years to make changes to the lobby that ever worked. That makes it sound like user1 is completely to blame for the lack of software upgrades to the lobby. Let's please avoid placing blame because I'm sure that not only one person is to blame for the impasse. How about let's share responsibility and look for ways to help that are within our own scope of responsibility, such as "find a C++ coder" for the 0 A. D. lobby improvements. Last I checked there's over $30k in the account for WFG in case there is not a C++ coder available willing to donate time. Another idea: offer to help Dunedan with the CI and pipelines on Github for the lobby server software in a way that does not require giving you administrative privileges. Ask what exactly they're stuck on, provide instructions and pointers to articles that answer questions, provide advice on best practices and anecdotes about how similar problems were solved elsewhere.
  9. How do you know why go2die retired from the forum? I think that it is plausible that lobby moderator response time was at least part of the reason, but go2die's last two posts on the forum before retiring were complaining about the difficulty of getting a mod signed for publication on mod.io. (For the record, I think that go2die gave up too early. It looks like user1 is trying to improve right now. But, he's not the only person who has needed to improve, in my opinion.) And, improving the lobby moderation is not something that only involves user1. There are aspects that are outside user1's control, such as improving the legal terms, and whether or not email addresses are collected. Many of the features suggested above, such as showing account age to all users, preventing spoofing of rated games played and rating, and indication of whether an account is "verified" or "unverified", would also involve changes to the 0 A.D. client, which is not (solely?) user1's responsibility. Can we please act more like a team here instead of just pointing fingers at each other?
  10. If there is no one with the official role in WFG of drafting legal terms, then the word responsible is still appropriate. It means who has accepted the responsibility for this task? If collecting email addresses is not the course of action that WFG wants to take then what solutions are being implemented to ensure that bans are effective? Here are some ideas from others, as well as from me. Improve the rate limit for new account creation per IP. Someone said that an IP can create 1 account per hour. Consider adjusting this to, for example, 1 account per month. I think that Gmail allows 10 email addresses to be created from a certain IP before they require a mobile phone number to be attached to a new account. Make account age public for all users to see. Also, make rated game record and rating unspoofable. Regularly scan for weak passwords, and lock accounts with weak passwords. Ban them if there is no password reset capability. Allow free registration without an email address, but establish a policy where all new accounts are "unverified" and have limited privileges. Allow "verified" users to host games that are only open to other "verified" accounts. Changing an account to "verified" requires 3 referrals from existing "verified" accounts, which will lose their own "verified" status if they commit fraud, and also requires an email address and dossier to be filled out. The purpose of the dossier is to have consistency of individual identity across duplicate accounts. Allow use of aliases in order to protect players from bullying or targeting, but require that the aliases are tied to a "verified" account in order to make rule enforcement effective. What's happening with the status quo is that the cost of player misconduct and easy duplicate account creation is being externalized to the player base, especially those players who regularly host games. This still has an impact on WFG, of course, as it causes players to consider leaving or at least withdrawing support for WFG. Case in point: go2die's retiring from the WFG forum yesterday.
  11. How difficult or easy is it to justify the usage of an email address for an online account? Who is responsible in WFG for drafting this language?
  12. Gurken Khan said: Exactly. So, a path forward with making walls more realistic is through modding and/or an alternate game mode. Once it reaches critical mass then players might like to play it sometimes in multiplayer. I would like to see walls actually block projectiles, and offer much more offensive and defensive benefits. I read somewhere that castles could be successfully defended with 1/6th the number of troops as those who were attacking. Of course, siege warfare is the logical counter to this, and that can usually only be broken by a counter-attack by an allied army.
  13. Okay, thanks for the clarification. I think that the player base isn't aware of this policy. Ideally it would be stated in the terms that users read before connecting to the lobby. I'll take this opportunity to announce again that I have rules for the games that I host, and most of the players who have joined my team games seem to like this. What we have is, in my opinion, a bad combination of zero-cost new account creation (no email address, no effort required) and infinite duplicate accounts allowed. Sevda/Yekaterina has multiple times suggested mitigations, such as establishing a primary account for each player, and allowing duplicate accounts that are tied to that account. That way, admins can identify who the primary user is. Others have suggested tasks that are necessary for new accounts to be allowed to join games, such as play several rated games, or complete a training mission. But, as it stands now, bans are completely ineffective on players who can change their IP addresses or use VPNs, and therefore player misconduct (verbal abuse, cheating, smurfing, griefing, ddosing, etc.) is very difficult to control. I see several statements related to GDPR in the Privacy Policy. We can start collecting email addresses now, right?
  14. Click Settings, Video. Ensure that the base (canvas resolution) and output (scaled resolution) are correct. The base canvas resolution needs to be the same as your monitor resolution in order to play and stream with 0ad in full screen mode. The scaled resolution can be lower if you want to downsample the video stream in order to save CPU and/or bandwidth while streaming. Then click Window Capture, then (at the top) Edit, Transform, Fit to Screen. This should cause the red box to expand to the full video frame. You also probably want to add an Audio Input Capture (pulseaudio) for your microphone, and Audio Output Capture (pulseaudio) in order to capture all audio output. Then click Edit, Advanced Audio properties to tune the recording volume and other settings.
  15. I've looked at the code in DllLoader.cpp, but I don't see yet how I can have the complete path output to the console. The base filename is passed to dlopen(); doesn't really make sense to me yet. It looks like it's calling dlopen on merely the filename. That's a relative path that implies "search the current working directory for the file". You could control the current working directory before that point in the code with a bash script or C++ calls. Or, ensure that the installer places the library in the appropriate directory.
  16. Consider setting, and exporting, the LD_LIBRARY_PATH=[absolute path to directory containing libraries] or LD_PRELOAD=[absolute path to specific library] environment variables in a shell script that is called before the pyrogenesis binary. See the manpage for `ld' at the documentation for the switch -rpath-link=dir for a description of the library search path order. But, I haven't worked with AppImages before, so there might be reasons to not do this. Also, you can use strace or ptrace to watch what directories are being searched for libCollada.so. Search paths for shared libraries are configured in /etc/ld.so.conf.d/* If you change the ld.so configuration then don't forget to run ldconfig as root to update the cache. However, changing such configuration might be an even worse idea, as it could have unintended effects on other binaries and libraries.
  17. Good points, smiley. I agree. I think that jurisdiction, governing law of a contract, and which local laws were infringed are actually 3 separate, but potentially mutually affecting issues. Anyway, the pros advise adding a Governing Law clause to legal terms for online services. Who in the Wildfiregames team decides on that?
  18. Could you please include a note about whether these free games are time-limited for claiming the offer but then permanent ownership and use, or else time-limited for playing it even after claiming the offer.
  19. Absence of (obvious) evidence is not evidence of absence. It's possible that the vast majority of users are apathetic about enforcing rules because rule infractions are so rampant. Think fragile state syndrome, despotism, etc. I have noticed many times that users have complained about smurfing in text chat but not in the forum. Here are recent examples, either in lobby or in-game chat About a week ago, in-game chat, Spanish translated to English: (Anonymous): Deanhernandez is the smurf that f____s us In April, 2022... Norse_Harold: we're trying to find out who ProphetMuhammad (741) is a smurf of ProphetMuhammad (741): ProphetMuhammad (741) is NOT a smurf ProphetMuhammad was presumably relatively skilled, but claimed to be accurately rated at 741. The reason this is a problem is because it creates a significant imbalance in a team game. Then players feel like they wasted 20 to 50 minutes playing such a team game. There are more examples, I'm sure. I haven't written them down. I think that creation and enforcement of rules against smurfing can be done primarily by game hosters. Hosters should be asking every unfamiliar player what their skill level or rating is, somehow remembering their stated skill levels, and punishing users that have lied about their skill levels. I think that lobby moderators only need to be involved if there is evidence posted on the forum demonstrating behavior that is defined as a violation of the lobby terms. I define a smurf as "a player who claims to have a significantly lower skill than the player's actual skill". I consider a player who tells the truth about their rating, despite being pseudonymous to NOT be a smurf. I think that some users are not careful about the terminology they use when describing a problem, and some people have used the word "smurf" to complain about duplicate accounts. It's a separate issue and shouldn't be called "smurfing" in my opinion.
  20. But not everyone is in the US and we can just age restrict the lobby meanwhile let USA's children beat up the AI. Although the lobby server is located in Germany, the jurisdiction of the enforcement of the lobby's legal terms is probably the US. This seems like an opportunity to improve the legal terms further, by adding a Governing Law clause. Anyway, I think that one or more admins of the lobby is located in the US, therefore US law is applicable to the conduct of those admins. Therefore, they can't legally collect data on persons under 13 without parental consent. Therefore, the Privacy Policy, Terms of Service and Terms of Use restrict lobby users to 13 or older. Anyway, Germany probably has a similar law to COPPA.
  21. Yes. Another possible cause is an undersized CPU heatsink for the TDP (thermal design power) of the particular model of CPU. When modern CPUs overheat, they temporarily run slower, due to a function called "thermal throttling". It can be detected with software such as Intel Xtreme Tuning Utility (XTU).
  22. In what part of this quote did I say "it's a lot more efficient." ?
  23. Not entirely sure why. I think this game is perfectly appropriate for anyone; young children can play the Pony Ascendant mod and no blood mod. I agree that the single-player game is appropriate for young children. The lobby and lobby-hosted multiplayer games are not, though. Read about the US law COPPA about collection of information on users under 13. Also, "online interactions" can have significant impacts on psychology of young users, due to cyberbullying, frequent use of profanity, impressionable young minds following bad examples. Young users sometimes don't know that they are friends with and/or mimicing psychopathic behavior until a significant amount of time has passed.
  24. Intel i9 gen 11: Passmark score depends on the specific model Intel i9 gen 11: Mersenne score depends on the specific model. i9-9900K @ 3.6 GHz gets 10.43 to 22.44 depending on parameters used NVidia RTX 3060 Ti: 11700 3DMark 10 score And, if you're running a benchmark and the temperatures are within the maximum rated operating temperatures, then the PC will likely "take off due to the fans going crazy", and it's normal. If one doesn't consider it normal then it's a noise optimization problem. What would not be normal is if the temperatures are outside the maximum rated operating temperatures, or else the temperatures and/or fan speeds are too high for the load placed on the system. Some people break their CPU heatsink mounting tabs, which would cause that symptom.
  • Create New...