afx Posted June 17, 2021 Report Share Posted June 17, 2021 I quite often get Out Of Sync Errors on a local multiplayer game 2 users against 2 AI opponents, both machines are windows based. Hopefully the logs are useful. ciao afx logs.7z Quote Link to comment Share on other sites More sharing options...
Ceres Posted June 17, 2021 Report Share Posted June 17, 2021 (edited) I played the latest RC with 2 more players and 1 bot 2 times. We 3 (in 1 team ) were beaten up major. It seems to me that a25 has a stronger AI (as before set very easy and balanced). Well, it's me and our kids - all with not much playing experience. But the graphics look great. The only thing I observed were 3 overlapping ships, almost hardly to be recognized as 3 (see screenshot a couple of posts earlier). No crashes so far. A few words about the resolution: The problem about menu entries being overlapped is not restricted to the windowed mode. It also happens in full screen mode. The root cause is setting the magnifying factor too high in Windows 10, e.g. 150x. When set to 125x, the menus are fully visible. Isn't there a setting in some prefs file for the GUI? Oh, another thing: Gfx look "ugly" (no offence intended!) when using a 4K resolution. I found it so bad that I immediately set the resolution back to 1920 x 1080 (and the factor to 125x) that I forgot to take a screenshot. Is 0 A.D. intended to display 4K gfx at all? Please forgive me my ignorance about this, I really don't know. Edited June 17, 2021 by Ceres Quote Link to comment Share on other sites More sharing options...
Stan` Posted June 17, 2021 Report Share Posted June 17, 2021 11 minutes ago, afx said: I quite often get Out Of Sync Errors on a local multiplayer game 2 users against 2 AI opponents, both machines are windows based. Hopefully the logs are useful. ciao afx logs.7z 2 MB · 0 downloads Do you try to rejoin with AIs cause that's not supported. 10 minutes ago, Ceres said: A few words about the resolution: The problem about menu entries being overlapped is not restricted to the windowed mode. It also happens in full screen mode. The root cause is setting the magnifying factor too high in Windows 10, e.g. 150x. When set to 125x, the menus are fully visible. Isn't there a setting in some prefs file for the GUI? Only in the config file. It's gui.scale. 150% meants 540p of height which is inferior to 768 the minimum res. 12 minutes ago, Ceres said: Oh, another thing: Gfx look "ugly" (no offence intended!) when using a 4K resolution. I found it so bad that I immediately set the resolution back to 1920 x 1080 (and the factor to 125x) that I forgot to take a screenshot. Is 0 A.D. intended to display 4K gfx at all? Please forgive me my ignorance about this, I really don't know. We don't really support Highdpi but we're working on it. What do you mean by ugly ? 1 Quote Link to comment Share on other sites More sharing options...
Ceres Posted June 17, 2021 Report Share Posted June 17, 2021 Sorry, "ugly" was not a nice word, nor was it precise. The contours of the moving units show steps and remembered me a bit of the sprites on the Commodore C64. I will post a screenshot tomorrow. Maybe it was simply due to the Windows resolution factor (too high => inferior gfx). Is there a way that 0 A.D. keeps its resolution factor irrespective of the one set for Windows? Can the be different? I know from other games (e.g. SuperTuxKart) that have a similar behaviour, i.e. when the resolution factor in Windows is different than 100%, it shows impact in the game as well (both windowed and full screen). Quote Link to comment Share on other sites More sharing options...
afx Posted June 17, 2021 Report Share Posted June 17, 2021 4 minutes ago, Stan` said: Do you try to rejoin with AIs cause that's not supported. I did not try anything but play with two players against 2 AIs. I just get OoS errors and a popup informing me about them and the inability to rejoin, here is the latest session, logs from host and client. I had that occasionally with a24, but with a25 it is nearly every game. cheers afx logs-host.7z logs-client.7z Quote Link to comment Share on other sites More sharing options...
Ceres Posted June 17, 2021 Report Share Posted June 17, 2021 Do you play via WiFi or LAN? Do have have the OOS errors also with only 1 AI? Does it matter if the AI is in a team with one of the human players or in the opponent's team? Quote Link to comment Share on other sites More sharing options...
afx Posted June 17, 2021 Report Share Posted June 17, 2021 3 minutes ago, Ceres said: Do you play via WiFi or LAN? Do have have the OOS errors also with only 1 AI? Does it matter if the AI is in a team with one of the human players or in the opponent's team? WiFi, 5Ghz, Ubiquiti access point in the same room. Yes, also with one AI only, the last logs I posted have 2 humans against one AI. (the host is an i7 and the other box a Ryzen 7, both 16GB of RAM) We only played AI opponents so far. cheers afx 1 Quote Link to comment Share on other sites More sharing options...
Obelix Posted June 17, 2021 Report Share Posted June 17, 2021 5 hours ago, Ceres said: [...] Yesterday evening I had a quick chat with @Langbart via IRC concerning the fact, that after installing RC 1 (Win), which is 25799, the build info on the startup screen says, it is 25794. The quoted post contains the problematised Info. May someone please confirm, that this is not a bug? Yesterday, one of the theories was, that it might be connected with different OS-builds, but we distanced ourselves from the theorie quickly. 1 Quote Link to comment Share on other sites More sharing options...
Stan` Posted June 17, 2021 Report Share Posted June 17, 2021 8 minutes ago, Ceres said: Is there a way that 0 A.D. keeps its resolution factor irrespective of the one set for Windows? Can the be different? I know from other games (e.g. SuperTuxKart) that have a similar behaviour, i.e. when the resolution factor in Windows is different than 100%, it shows impact in the game as well (both windowed and full screen). Not really because of a dumb bug in the sdl iirc. We don't have a true fullscreen mode so we can't set the resolution. There is a new hidpi option added by vladislav in A24 not sure what it does (Only works on windows) 7 minutes ago, Ceres said: Do you play via WiFi or LAN? Do have have the OOS errors also with only 1 AI? Does it matter if the AI is in a team with one of the human players or in the opponent's team? You cannot rejoin at all. So if you play with AI and get disconnected the game is lost. 1 Quote Link to comment Share on other sites More sharing options...
Stan` Posted June 17, 2021 Report Share Posted June 17, 2021 1 minute ago, Obelix said: May someone please confirm, that this is not a bug? Yesterday, one of the theories was, that it might be connected with different OS-builds, but we distanced ourselves from the theorie quickly. It's not. It's usually off by one between windows and other oses. But since there were no cpp changes the person making the RC did not bother recompiling the game on windows hence the big gap you saw. It's guaranteed however that all the rc files are compatible for a given rc. 1 Quote Link to comment Share on other sites More sharing options...
Gurken Khan Posted June 17, 2021 Report Share Posted June 17, 2021 1 hour ago, Nullus said: I like to spend time making my cities look nice, so I probably play at a slower pace than most players. I try to make my cities look neat. But since I play on turtle speed you're not playing at a slower pace than me. (I look a lot at remaining build times, my focus just isn't on the architecture...) 51 minutes ago, Yekaterina said: In general crashes happen very rarely, and only occurs if my pop cap is set to more than 250. So I have a feeling that it is caused by lack of ram and processing power instead of a genuine bug. I play with 5 AIs at 300 pop cap, also got 8GB RAM* and an i5-core (though a bit faster than yours); didn't have a crash with the RC yet, but maybe I was just lucky and/or you're just getting through a lot more games. * Spoiler I believe there was sth like being a 32 bit app 0AD only uses ~4 GB RAM; so unless you're doing a lot of stuff in the background that probably shouldn't be the problem. BTW I like how ships can navigate around each other now to attack the same target. (Didn't have a fleet on a24, can't tell how new that behavior is.) Quote Link to comment Share on other sites More sharing options...
Yekaterina Posted June 17, 2021 Report Share Posted June 17, 2021 1 minute ago, Gurken Khan said: I play with 5 AIs at 300 pop cap, also got 8GB RAM* and an i5-core (though a bit faster than yours); didn't have a crash with the RC yet, but maybe I was just lucky and/or you're just getting through a lot more games. Are you also on Windows 10? I had a few chrome tabs open, that might be why. Also I use a laptop which has stricty tdp limits. Quote Link to comment Share on other sites More sharing options...
Yekaterina Posted June 17, 2021 Report Share Posted June 17, 2021 Specs (my laptop): i5 8250u (4 cores 8 threads @1.6GHz) 8GB LPDDR4 RAM 250GB disk (sata m.2) Nvidia MX150 Windows 10 Pro My desktop: i7 8700 (6 cores 12 threads turbo up to 4.3GHz) 16GB RAM (DDR4 2666) Nvidia GTX 1660 (MSI armour) WD Nvme m.2 SSD Fedora 34 My desktop has never experienced a single crash on Fedora, not even dev svn A25. There had been a few crashes back in A23 when I was still a Windows user. Ubuntu crashed a few times in A23 as well but less frequent than Windows. 1 Quote Link to comment Share on other sites More sharing options...
Gurken Khan Posted June 17, 2021 Report Share Posted June 17, 2021 24 minutes ago, Yekaterina said: Are you also on Windows 10? I had a few chrome tabs open, that might be why. I'm on Win8.1. Isn't Chrome infamous for its RAM-hunger? 20 minutes ago, Yekaterina said: There had been a few crashes back in A23 when I was still a Windows user. Ubuntu crashed a few times in A23 as well but less frequent than Windows. A23's stability was a catastrophe for me. It crashed so reliably that I tried to save, exit, reload every ~20 mins. Crashes on a24 were rare, and as stated before I'm so far happy with a25, also in that regard. Quote Link to comment Share on other sites More sharing options...
Yekaterina Posted June 17, 2021 Report Share Posted June 17, 2021 1 minute ago, Gurken Khan said: Isn't Chrome infamous for its RAM-hunger? 1 Quote Link to comment Share on other sites More sharing options...
Yekaterina Posted June 17, 2021 Report Share Posted June 17, 2021 3 minutes ago, Gurken Khan said: I'm on Win8.1. I recommend dual-booting your Windows 8.1 with a Linux distro, I recommend Fedora (if you are as dumb as me) or Arch (if you are clever). 1 Quote Link to comment Share on other sites More sharing options...
Gurken Khan Posted June 17, 2021 Report Share Posted June 17, 2021 1 minute ago, Yekaterina said: I recommend Jaja, you're earning yourself a reputation as a linux-fangirl. But I'm dumb and lazy, so as long as the OS (Win) that comes with my machine does what I want it to... To be on-topic: I just successfully pressed the 'disable feedback' button and then the 'enable feedback' button. 1 1 Quote Link to comment Share on other sites More sharing options...
Yekaterina Posted June 17, 2021 Report Share Posted June 17, 2021 I just tested every single button in the game and all of them work. 6 minutes ago, Gurken Khan said: Jaja, you're earning yourself a reputation as a linux-fangirl. I am already known as "Goddess of Linux" in my school 3 1 Quote Link to comment Share on other sites More sharing options...
Ceres Posted June 18, 2021 Report Share Posted June 18, 2021 8 hours ago, Stan` said: It's not. It's usually off by one between windows and other oses. But since there were no cpp changes the person making the RC did not bother recompiling the game on windows hence the big gap you saw. It's guaranteed however that all the rc files are compatible for a given rc. I downloaded the latest installer: https://releases.wildfiregames.com/rc/0ad-0.0.25-rc1-25799-alpha-win32.exe The number shown on my earlier screenshot is 25794, though, as @Yekaterina observed. But now I understand that that's ok and no bug. Correct? Quote Link to comment Share on other sites More sharing options...
zyli Posted June 18, 2021 Report Share Posted June 18, 2021 On 13/06/2021 at 3:48 PM, zyli said: Debian Buster – failed compilation – ERROR: SpiderMonkey build failed. ------------------------------------------------------------------------ File "/media/domowy/exfat_64GB/0ad_tarball/0.0.25_rel_25782/ The cause of a Debian Buster compilation error: compiling on the exfat filesystem. As you can see above, I even gave such a label to the USB stick. Compilation on Linux ext4 was successful. I apologize for this oversight. Quote Link to comment Share on other sites More sharing options...
Yekaterina Posted June 18, 2021 Report Share Posted June 18, 2021 1 hour ago, zyli said: compiling on the exfat filesystem. Yeah, exfat doesn't work. I have tried exfat storage before and I get crashes even when moving files, so I recommend installing Debian properly on a ext4 hard drive. Ext4 probably has better performance as well. Quote Link to comment Share on other sites More sharing options...
Ceres Posted June 18, 2021 Report Share Posted June 18, 2021 (edited) 10 hours ago, Gurken Khan said: BTW I like how ships can navigate around each other now to attack the same target. (Didn't have a fleet on a24, can't tell how new that behavior is.) Yes, I like this, too. So is the overlap of multiple ships a side effects of this? Edited June 18, 2021 by Ceres Quote Link to comment Share on other sites More sharing options...
wraitii Posted June 18, 2021 Report Share Posted June 18, 2021 2 hours ago, Ceres said: But now I understand that that's ok and no bug. Correct? Correct 2 minutes ago, Ceres said: Yes, I like this, too. So is the overlap of multiple ships a side effects of this? Yes, ships have 'pushing' like units do, which makes them slightly more usable, but also overlap quite a lot. 1 Quote Link to comment Share on other sites More sharing options...
Ceres Posted June 18, 2021 Report Share Posted June 18, 2021 Thank you. I think that the new maneuvering (pushing) possibilities offer much advantages. Could you (at least for ships?) think of a way to better indicate that there are more ships (or even show the number) without the need that the player double-clicks on one to get the number of all? In fact, then the number of all ships (of that category only?) on the whole visible screen would be shown, not just of ships close to each other. Not that it matters too much, and maybe such situations are rare. I just saw it be coincidence and only once. The whole "look & feel" of a25 is a great progress - and it seems to run smoother also on machines with integrated GPUs like mine, which I very much appreciate. Quote Link to comment Share on other sites More sharing options...
Ceres Posted June 18, 2021 Report Share Posted June 18, 2021 My Intel NUC i3-8109U PC @3.00 GHz and with 16 GB RAM is connected via HDMI to an LG 4K TV (with speakers). Windows 10 (Pro, 64bit, latest updates and drivers installed) has HDR enabled and a 3840x2160 resolution. When running a25 RC with Windows scaling set to 100%, there is no sound on the TV: After I increased scaling to 125%, sound plays well: I used different (and good) HDMI cables, played with the maximum vertical scan (now set to 60 Hz), but nothing helped except for setting the scaling to 125% or more. Is this problem specific at all for a25 RC? Sorry that I cannot test it on a24 right now. Anyway, I am curious if this is something that could be corrected (in a future version). Thanks for sharing your thoughts. Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.