Chobbes's recent activity
-
Comment on How bad are Nvidia GPUs for Linux really? in ~comp
-
Comment on The Steam Deck now has over 5,000 Verified games in ~games
Chobbes Link ParentYou know, I hadn't thought about it but... That kind of makes sense! The dock was a huge selling point of the Switch, and it would be incredibly frustrating if it didn't work... And I don't think...You know, I hadn't thought about it but... That kind of makes sense! The dock was a huge selling point of the Switch, and it would be incredibly frustrating if it didn't work... And I don't think I've ever heard anybody complain about it, which is kind of shocking in retrospect.
It's been unclear to me if the problem was that it's just integration hell with different TVs being super picky about something... But surely it has to be possible to handle this otherwise like... Any console with an HDMI port should be in connection hell too. I guess the Steam Deck is a little different in that it's probably less limited in output resolution and stuff than other consoles (which I assume still have like... 4 options for resolutions that will tend to be standard and accepted by TVs), so maybe that's part of the problem. Also part of me can't help but feel like this would be less of an issue if everything just used DisplayPort instead of HDMI (my understanding is that the HDMI licensing can be a bit of a problem for open-source drivers on Linux :|).
-
Comment on The Steam Deck now has over 5,000 Verified games in ~games
Chobbes Link ParentYeah, I had read that too, but I wasn't having any luck with unplugging the power supply either! I really hope they figure out what the issues are, because it's super frustrating when you can't...Yeah, I had read that too, but I wasn't having any luck with unplugging the power supply either! I really hope they figure out what the issues are, because it's super frustrating when you can't get it to work.
-
Comment on The Steam Deck now has over 5,000 Verified games in ~games
Chobbes Link ParentI had all of these problems with the dock as well... Just completely unable to get it to reliably connect to the TV, restarting everything in different orders and just being miserable. I recently...I had all of these problems with the dock as well... Just completely unable to get it to reliably connect to the TV, restarting everything in different orders and just being miserable.
I recently had so much trouble and restarted the Steam Deck like 5 times and then decided "whatever, I'm restarting anyway, might as well try to undervolt the Steam Deck while I'm rebooting"... And ever since I haven't had any issues (knock on wood). It seems like that shouldn't fix it, but maybe going into the bios actually reset some state or maaaaybe the deck draws slightly less power and the dock couldn't handle it before. I really have no idea, but I'm happy it's finally working because it was VERY frustrating.
Honestly, docks are the worst. If they don't work it's so hard to know if it's 1) the dock itself, 2) the dock firmware, 3) the docked device's firmware, 4) the TV itself (TVs suck at being monitors), 5) the HDMI cable, 6) the USB-C / Thunderbolt cable. I have never had an experience with a dock that was just good, and it's such a nightmare to debug.
-
Comment on What email client do you use? in ~tech
Chobbes Link ParentYeah, I use emacs for everything. It's great :).Yeah, I use emacs for everything. It's great :).
-
Comment on What email client do you use? in ~tech
Chobbes LinkI've been using mu4e. It's great having all of your e-mail locally with super fast search, but the main reason I use mu4e is that it's in emacs... Which is really nice because now I can make todos...I've been using mu4e. It's great having all of your e-mail locally with super fast search, but the main reason I use mu4e is that it's in emacs... Which is really nice because now I can make todos and stuff in org-mode that reference e-mails, and I can easily jump to the e-mail referenced in a todo for additional information.
-
Comment on Apple has kept an illegal monopoly over smartphones in US, Justice Department says in antitrust suit in ~tech
Chobbes Link ParentAlso good fucking god, I would love to have the source code for my TV so I could debug it and fix it when it segfaults >_<. Quality.Also good fucking god, I would love to have the source code for my TV so I could debug it and fix it when it segfaults >_<. Quality.
-
Comment on Qualcomm says most Windows games should ‘just work’ on its unannounced Arm laptops in ~games
Chobbes Link ParentOh, I totally agree that the overreach is ridiculous. I don't know if I agree that user-space anti-cheat software could be effective, though. It seems like something you would need some kernel...Oh, I totally agree that the overreach is ridiculous. I don't know if I agree that user-space anti-cheat software could be effective, though. It seems like something you would need some kernel level / hardware support for, to be honest. Not that that means we should accept these things, of course :). I'm honestly a little surprised we haven't seen more usage of TPMs and security extensions for DRM, and frankly I hope we never do because that gets dystopian very fast.
-
Comment on Qualcomm says most Windows games should ‘just work’ on its unannounced Arm laptops in ~games
Chobbes Link ParentI kind of understand the desire for anti-cheat software, because cheating sucks... I've kind of ducked out of online gaming for the past decade, but I feel like part of this problem is that games...I kind of understand the desire for anti-cheat software, because cheating sucks... I've kind of ducked out of online gaming for the past decade, but I feel like part of this problem is that games have moved towards servers which are controlled by the companies instead of community servers so you have to rely on official moderation and stuff. Anti-cheat seems a lot more important when moderation is in the hands of the company and costs them money and there are no community servers which can enforce their own rules (which has its own problems too, of course, but I think it can be a win).
-
Comment on Qualcomm says most Windows games should ‘just work’ on its unannounced Arm laptops in ~games
Chobbes Link ParentThere's a little bit of translation because of differences in calling conventions between Windows and Linux, I think, but yeah with WINE you're essentially just running "native code" (insomuch as...There's a little bit of translation because of differences in calling conventions between Windows and Linux, I think, but yeah with WINE you're essentially just running "native code" (insomuch as native code is a real thing, anyway, haha, the boundaries are a little fuzzy imo).
-
Comment on Qualcomm says most Windows games should ‘just work’ on its unannounced Arm laptops in ~games
Chobbes Link Parentbinfmt_misc is a really cool Linux kernel feature that will let you transparently run executables from different architectures under QEMU and stuff. I imagine it can be put to great use for this...binfmt_misc is a really cool Linux kernel feature that will let you transparently run executables from different architectures under QEMU and stuff. I imagine it can be put to great use for this kind of stuff.
-
Comment on Qualcomm says most Windows games should ‘just work’ on its unannounced Arm laptops in ~games
Chobbes Link ParentI'm not sure how much it will matter to be honest. There will almost certainly be games that don't work or perform poorly, but I'd probably wager a good chunk of games will be just fine with a bit...I'm not sure how much it will matter to be honest. There will almost certainly be games that don't work or perform poorly, but I'd probably wager a good chunk of games will be just fine with a bit of CPU emulation bottleneck, especially if the ARM chips keep getting more performant over time. I imagine it could be a bit like Proton on Linux where most games (particularly if they're not bleeding edge / really CPU bound) will just work fine, with a few notable exceptions.
Still, I do find it disappointing that we have all of these proprietary x86 binaries out in the wild. It'd be really nice if we just had the source code so we could make proper ARM builds out of existing software, haha.
ALSO... I dunno, maybe I'm just paranoid, but I don't really trust Qualcomm? Like I hope these chips are great, but doesn't Qualcomm kind of suck at supporting their old chips? Isn't that a huge part of the reason why Android phones had terrible long term software support? I really don't want to buy a laptop that will only be supported for 9 months T_T. I'm sure it will be better than the Android situation was, but god am I ever bitter about that lol.
-
Comment on Qualcomm says most Windows games should ‘just work’ on its unannounced Arm laptops in ~games
Chobbes Link ParentYeah, this is something that's a bit unclear to me too. There's definitely stuff that's nice about ARM vs x86, but I don't think anybody fully knows how this will shape up, and sometimes I...I haven't seen any compelling arguments that one ISA is inherently more efficient than another, as neither ISA is truly representative of how their respective chips work under the hood anymore.
Yeah, this is something that's a bit unclear to me too. There's definitely stuff that's nice about ARM vs x86, but I don't think anybody fully knows how this will shape up, and sometimes I convince myself that the x86 slander is kind of overblown and the differences don't really matter much in the grand scheme of things. Like... I wouldn't be terribly surprised if ARM ends up coming out on top, but overall I haven't seen great arguments for why it would dominate in all spaces vs x86. Like having a simpler instruction set with fixed-width instructions and more registers available feels like a win... But it's at least not obvious to my amateur eyes that stuff like the number of named registers matters when you start having register renaming and stuff in hardware anyway. All of the computer architecture people I talk to seem to be a little unsure too?
Either way, I'd love for a more power efficient laptop, and the Apple Silicon chips seem really great for that use case... Hoping we can get some of those benefits in Thinkpad land.
-
Comment on Titanfall is still EA's most innovative shooter ten years later in ~games
Chobbes Link ParentMaybe you're just making a joke, but isn't that essentially what Apex Legends is?I'm pretty confident that EA will dig the IP up again sometime, but probably as a life service looter shooter. Come to think of it, I wonder why they haven't done so yet.
Maybe you're just making a joke, but isn't that essentially what Apex Legends is?
-
Comment on Time to delete your Glassdoor account and data in ~tech
Chobbes Link ParentI think I legit just missed that sentence when I was reading it, lol. Apologies :). Yeah, absolutely. In theory the sending mail server could enforce names for senders... but like... no personal...That said, I did address SPF and DKIM when I said the following:
I think I legit just missed that sentence when I was reading it, lol. Apologies :).
...and even then, if your email domain has DKIM and SPF setup correctly, and Glassdoor's email server has done likewise, none of that stops you from saying in the From: header that your name is John Doe when it's actually Jane Smith. So even with proper email standards implemented, just taking the sender name from an email as gospel about who you're talking to is an absurd failure.
Yeah, absolutely. In theory the sending mail server could enforce names for senders... but like... no personal e-mail services will enforce that.
Anyway, this was all just me tangenting to nerding about e-mail. At the end of the day none of this matters, and the point is that this is kind of a crazy decision on many levels. I can understand glassdoor wanting to ensure you actually worked at the company... But people should absolutely be allowed to be anonymous on that platform.
-
Comment on Time to delete your Glassdoor account and data in ~tech
Chobbes Link ParentI don't think there's any protections at all for the name in the From field. But at least for the e-mail addresses themselves, this is getting less true over time. I still wouldn't fully trust it...There's no ensuring that the name in an email To: header is the legit sender of the message.
I don't think there's any protections at all for the name in the
Fromfield. But at least for the e-mail addresses themselves, this is getting less true over time. I still wouldn't fully trust it and it's not something most users are capable of verifying, but if the domain for the sender has set up SPF and DKIM and the receiving mail server actually checks those things, then anybody trying to spoof an e-mail from your address should get sent to oblivion, at least... But it's not quite a guarantee. -
Comment on The more I use Linux, the more I hate every distro in ~tech
Chobbes Link ParentAnd the 5 users that claim they don't want it will inevitably never use --no-update-keyring because there's basically no reason for anybody to ever do that.And the 5 users that claim they don't want it will inevitably never use
--no-update-keyringbecause there's basically no reason for anybody to ever do that. -
Comment on The more I use Linux, the more I hate every distro in ~tech
Chobbes Link ParentI don't do android dev, but it always seems super painful. I hope you find a solution lol.I don't do android dev, but it always seems super painful. I hope you find a solution lol.
-
Comment on The more I use Linux, the more I hate every distro in ~tech
Chobbes Link ParentThe only justification I can think of is that they might want to keep things minimal and not have a special case to update the keyring first or something... But frankly, I just don't respect that...The only justification I can think of is that they might want to keep things minimal and not have a special case to update the keyring first or something... But frankly, I just don't respect that line of reasoning. This is constantly a problem for everybody, and it feels like the package manager is broken if it doesn't take into account the fact that updating any package depends on the keyring. I think they finally addressed this by adding a systemd timer to update the keyring recently, but it just feels like a kludge and the fact that it took YEARS to do that is disappointing. I was a Gentoo user, and I switched to Arch temporarily when I got a new computer because I needed to get set up fast and didn't want to compile everything... But I was pretty disappointed with my Arch experience to be honest. It just feels like there's a bunch of places to stub your toes because Arch users want to have a couple of manual tricks they have to do every so often to feel smug about. It's like the Linux equivalent of having to add eggs to cake mixes so you feel like you're doing something.
I think my other major gripes (which may be wrong, it's been a while) were 1) when a kernel update happens Arch removes all of the old kernel modules immediately, so if you have a USB device that needs a module that you haven't loaded before and you plug it in after running an update it won't work (yes, you should reboot after a kernel update, but no, I'd like it if you don't break the current system while updating), and 2) downgrading packages wasn't a great experience when the packages broke (rollbacks with btrfs or whatever would help, but iirc you just had to hope you still had the old version in your pacman cache, because I think the Arch repos only keep the latest version available? Annoying when the intel wifi drivers break lol). Maybe this is just another "get gud" situation, and you can certainly manually put things in place to ease these problems, but doing that always makes it feel like a hack (and you have to manually do this for every machine, ugh). I also never really understood why 3rd party packages from AUR couldn't be added to pacman, so you have all of these unofficial package managers like
yayandaurato install things from AUR... I get that it's slightly different because you have to build the packages from source, and AUR is an untrusted repository... but it feels like it should be integrated better. I'm probably just bad at Arch, but I honestly don't understand why everybody uses it... I guess it's the main option for a binary minimalistic distribution, but it just feels unpolished to me. It's easily the distro that I've had the worst experiences with personally, but other people love it so fair enough, I guess! -
Comment on The more I use Linux, the more I hate every distro in ~tech
Chobbes Link ParentSounds like you could use some nix in your life! On the plus side you don't have to install NixOS in order to use the nix package manager for development environments on your main machine, it can...This has me thinking it might be worth it. I work from home and I rely on virtual machines to isolate work environments which is a huge pain in the ass because they have to live on an ext4 drive and I have to manually backup important stuff with shared folders and it's pretty gross.
Sounds like you could use some nix in your life! On the plus side you don't have to install NixOS in order to use the nix package manager for development environments on your main machine, it can exist alongside your current distribution, so you can try it out incrementally pretty easily :).
It should be much more lightweight than having a bunch of VMs O_O. I had a huge amount of pain dealing with projects with different dependencies before... I also frequently had to backport fixes to older versions of a project and it was super painful to keep track of all of the valid dependencies from the past and the valid dependencies in the present. Now all of the dependencies are pinned in the git repo using nix so when I switch branches my shell automatically picks up the new dependencies for everything and switches everything over... If I've used that version of things before it's cached locally so I don't have to download anything or build any dependencies and it's instant. It's actually amazing. No VMs, no containers, nothing in the way... I can just use my normal text editor and access everything directly from my normal operating system.
Plus, even if you do want VMs or docker containers you can use nix to build those! One cool thing you can do with nix is build a VM of a nixos system to test it before you even reboot. There's so many nifty things nix lets you do.
Nix looks weird and seems complicated at first, but fundamentally it's not much more than funny looking JSON with recursive functions for specifying dependencies / build steps for a project.
In my experience in the past Nvidia proprietary drivers generally work just fine on Linux. You may occasionally have some random black screen issues when booting up or resuming from suspend or something, but it mostly just works with some small annoyances every so often. If you tend to hold onto your hardware for a long time you may eventually run into an issue where the binary blob driver no longer supports your GPU, which can be annoying, especially since the old drivers that do support your GPU might not support new Linux kernel versions, so eventually you might get to the point where you can't upgrade your OS until you upgrade your hardware... Presumably this can be an issue with amdgpu as well, but since it's open source you can potentially hack it to work if you or somebody else really wants to. Occasionally there is some fight between Linux kernel devs and Nvidia that will break the drivers on bleeding edge kernels, but this isn't super common (and likely won't be an issue for you unless you're on a rolling release distro, which Mint is not).
In general AMD gpus are kind of hassle free on Linux and are more plug-and-play than Nvidia GPUs, but if you want to try Linux out I really wouldn't worry that much about having an Nvidia card. You likely won't have any problems, especially on a stable distro like Mint.