Quoting: syxbitIs this also built with Unity?

As a game-dev I would bet it's even built mostly from the same codebase.

View this comment - View article - View full comments

Quoting: dimko

Quoting: soulsourceI was aware that some distributions disable ICMP for regular users, but this is the first time I am wondering why they actually do that.

Does anyone know?

Quoting: soulsourceI was aware that some distributions disable ICMP for regular users, but this is the first time I am wondering why they actually do that.

Does anyone know?

I think it does something else compared to what you think it does.
https://wiki.archlinux.org/title/Sysctl#Allow_unprivileged_users_to_create_IPPROTO_ICMP_sockets [External Link]
I am not a native english speaker, but way I understand it works(suspect) is this:
Their server or whatever sends a ping/packet/whatever from itself to client machine on some port 1-1000(which in linux is a restrictive set of ports, that is ONLY allowed to be opened by default by a root user, by default on most Linuxes anyway).
And by receiving this packet on some port, client machine OS send back a ICMP reply. Which later on that reply is being judged for some glitch/network delay exploit detection.
If set permission is not set an default policy of OS is not to respond - ICMP reply will not be generated/send. Hence it breaks their anticheat.

Its all speculation of course.

The docs you linked sound more like how I understood it though. The setting seems to be about user group numbers, not port numbers:

Quoting: https://wiki.archlinux.org/title/Sysctl#Allow_unprivileged_users_to_create_IPPROTO_ICMP_socketsping_group_range determines the GID range of groups which their users are allowed to create IPPROTO_ICMP sockets.

Also, the docs directly talk about being able to ping other machines and receive their replies

Quoting: https://wiki.archlinux.org/title/Sysctl#Allow_unprivileged_users_to_create_IPPROTO_ICMP_socketsThe IPPROTO_ICMP (icmp(7)) socket type adds the possibility to send ICMP_ECHO messages and receive corresponding ICMP_ECHOREPLY messages without the need to open a raw(7) socket

Furthermore, the server cannot assume that it will be able to ping the user's computer (via ICMP), as many users' computers are behind firewalls and/or NAT, so imho it only makes sense for the user's computer to ping the server, and not the other way.

View this comment - View article - View full comments

I was aware that some distributions disable ICMP for regular users, but this is the first time I am wondering why they actually do that.

Does anyone know?

View this comment - View article - View full comments

Quoting: Villian

Quoting: soulsource

Quoting: poke86the Linux version is an AppImage and

My conclusion from this is that AppImage does not make the dependency situation better. If at all, it makes it more annoying.

the only conclusion you can take from that is tha flatpak is a lot more distro independent than appimage and gentoo is annoying

Yeah, Gentoo is so annoying, it even manages to cause the same issue with libcurl on Steam OS.

Quoting: melkemindSo, I just tried it on my Steam Deck and got a missing libraries error, specifically libcurl-gnutls.so.4. On my desktop, running Manjaro, it started right up with no problem.

View this comment - View article - View full comments

Quoting: poke86

Quoting: soulsourceit does not work out-of-the-box on Gentoo, because the AppImage does not contain all dependencies

Have you reached out to the devs about this? They should fix this.

Not yet. I need to check if they have a proper bug-report form or something... (I only downloaded it yesterday and didn't even play - just tried to launch it.)

Also, I don't think the World of Goo devs can do much about the libfuse dependency. That's a dependency problem introduced directly by AppImage and can only be fixed by the AppImage devs.

View this comment - View article - View full comments

Quoting: lqe5433Is it Vulkan or OpenGL?

I need to check to be sure (am at work...), but on my laptop, which does not have Vulkan set up, it crashes on launch when it tries to render something. That might, however, also be a Box64 issue, as my laptop is ARM-based.

View this comment - View article - View full comments

Quoting: poke86the Linux version is an AppImage and

it does not work out-of-the-box on Gentoo, because the AppImage does not contain all dependencies - the one thing that the AppImage format promises to solve.

So, what does one need to get the game to run on Gentoo?
The AppImage would need the outdated sys-fs/fuse version (slot 2 - luckily it can be installed at the same time as the up-to-date version). However, installing that will just lead to a different error message, this time about libcurl-gnutls, which is not bundled in the AppImage, though it should be.

In order to get the game to run, I had to first unpack the appimage (there is a command line switch for that), and then create a symlink in the game folder from libcurl-gnutls.so.i-forgot-the-version to /usr/lib64/libcurl.so.that-same-version. Then, launching the game with LD_LIBRARY_PATH="./" set (so that it finds the symlinked lib) finally gets it to run.

My conclusion from this is that AppImage does not make the dependency situation better. If at all, it makes it more annoying.

View this comment - View article - View full comments

Quoting: WoodlandorSeems a little strange how they could have programmed it in a way for a simple launcher of sorts to corrupt the user space.

Then again, I remember the original release of Bungie’s Myth 2 Soulblighter had a bug where if you went to uninstall the game it would recursively delete every directory on the C: drive 😂

So all we know is game programmers shouldn’t drink Mountain Dew and program… for everyone’s safety 🤷‍♂️

If I remember correctly, the Steam Linux client also had a bug at some point where it would wipe your whole home folder when uninstalling a game.

View this comment - View article - View full comments

I think I had missed this up to now.

But after seeing the "*cocks gun in Italian*" subtitle, I cannot not wishlist it.

View this comment - View article - View full comments

Quoting: Purple Library GuyThe top ones currently have basically scraped the whole internet, so I'm not sure there's a whole lot more to scrape.

Furthermore, output quality might not scale as well with training dataset size as AI proponents might hope. There's a Computerphile episode on this topic [External Link].

View this comment - View article - View full comments