Comments by "George Albany" (@Spartan322) on "Brodie Robertson" channel.

  1. Ah the classic toxic dev problem, probonopd really should be punished by being outed from his own project, either via a fork that everyone rushes over to or by core contributors pushing him to step down, when a dev screws up as much as someone like probonopd, he really needs to be pushed out of the project, specifically when he is intentionally screwing over everyone else simply because he doesn't like them and their tools, even when it cost him no maintenance time and would reduce bug complaints. Literally his own worst enemy.
    142
  2. Once again Linus' rants are entirely and wholly correct, even if he is a bit hostile fact is that folks really should actually listen to Linus even more then they already do, he rarely ever says something, especially of project management or FOSS in general, that is wrong, and when he does its usually only in technicalities.
    87
  3. 75% of the time the problem is also just straight up gnome, perhaps it be better if they simply had less influence over wayland in general given they straight up don't actually care about supporting it so long as they disagree with it.
    85
  4. This is one of the biggest continual issues on Linux, violating backwards compatibility, like most distros are better then Apple, but the fact that every few years someone feels the need to break userspace over and over and then 9 times out of 10 they stubbornly refuse to correct the issue themselves (requiring a patch at least to correct it) is a big contributor to why Linux has such a hard time picking up marketshare, it makes the kernel, and by extension the distros and the things we all rely on said distros, appear (and I would say it would outright just make it) extremely unstable in just the right ways that nobody can correct it.
    79
  5. I used to use Manjaro and its insane how often I suffered from dependency breakages because the package stuff, it was like every week I had to actively manage my packages cause they just kept breaking, once you get to a certain point on Manjaro you can't do anything without the AUR and once you do go into the AUR, your entire system will break and literally nothing will work right, (or even without it, it might just break anyway, I've had many types of non-sense problems no one knew how to fix that only went away once I got off of Manjaro) the Manjaro mentality, alongside utter incompetence, outright violates Arch and as a result routinely will break your system and introduce bugs. (and that's before you consider the non-merged PRs getting merged into the Manjaro packages) As a result of all the suffering Manjaro put me through I actively hate it as a distro.
    63
  6. I had a stroke reading the source and logs being in that abomination style.
    55
  7. I can bet you if I go to a government building, advertise myself as a legit printer, I could absolutely exploit this to get sensitive information, this is a case where there is no possible way to distinguish a legit printer from a malicious fake printer and all it needs is one person to execute the thing on the network and I will have capacity to pwn the network.
    36
  8. More like this aged like debian
    26
  9. That's cope, Rust fanbois are some of the worst types of community for a programming language I've ever seen, they go around trying to spread the gospel of the crab and as a developer it is annoying and I despise them purely for their insistence that everything must be Rust. C elitists don't even compare to the vitriol I've dealt with just telling the Rust guys to leave us alone.
    26
  10. My issue with a lot of the pronoun weirdos is more that they cause fusses like this, they're a lot like vegans, belligerent and obsessed over it, can't shutup about it, and you have to walk on eggshells just to deal with them, everything sets them off and you can't have any fun around them because they'll blow up like they have bipolar. (and I have family with that illness, the comparison is fairly apt) I'm not surprised that a random joke blew up into "hyprland is toxic" type of crap where the pronoun folks try to excise vaxry for something innocuous, I have an even more flippant disregard for it because of religious convictions and I already have a joke on my discord profile where the pronouns thing is that says "are biologically determined" thus when you highlight my "pronouns" it says "pronouns are biologically determined", I've gotten quasi-banned from FOSS project communities for things like that, but I don't care, I still use the software despite that. And no I did not harass anyone, (I never even brought it up there) I was told by leadership in these cases that my profile "may trigger" someone and they told me to change it or leave which I did the latter amicably, hence why I say "quasi-banned" and not actually banned.
    25
  11.  @aricrudd6579  This is what we call a strawman and ad hominem fallacy. Also nothing Christoph said had anything to do with ideology, sounds more like you've never maintained any piece of software in your life.
    20
  12.  @mattelder1971  The OSI is a product of pre-existing open source software and was written as a formal definition of what open source was, contributed to by the exact people that were writing said software at the time, a formal definition coming to exist after its use does not violate the formal definition being well defined. Your claim is fallacious.
    19
  13.  @Nina-cd2eh  But Rust demands semantic knowledge, it gets pretty useless without that, (not really unique to Rust, but it is the most aggressive language I know on that) and the burden of that semantic knowledge is on the C code which the Rust devs either have to refactor or ignore, the latter of which means they have no advantage to that code. (or worse, it makes false promises and then the Rust code is more broken then the C code was) In which case how exactly is Rust advantageous for that case?
    18
  14. Half of the problem with XZ was the distros needlessly externally and silently patching a package just to get a feature they should've been requesting be added as a damn feature, if they had done things through the official upstream XZ wouldn't have been able to activate. The other half of the issue is that XZ was integrating binary blobs into the codebase via the test suite instead of compiling from source and relying purely on source available files. (aside from the lacking maintenance oversight) So long as a package wasn't being compiled with binary blobs that could integrate secret behavior, the XZ problem simply does not exist. Any further security complaints are generally pretty dumb because after those issues you can't distinguish legitimate code from hidden malicious code, they'll look the basically the same, best you get to deal with that is code review. (like how do you distinguish a hidden backdoor from a standard buffer overflow at that point? You can't)
    18
  15. I'm rather happy Hectors gone, he's been the source of so much strife and has been a royal pain in the ass for a while, not to mention he's a toxic developer who wants to ban everyone he disagrees with which includes a lot of people. And I rather despise the R4L project because it really is a fad language, it lacks basic things that all system languages should have and do, (especially for a kernel) the Rust language developers act like children and much of the Rust community worships the language developers like their some kind of "memory safe" gods, any criticism in my experience turns into some personal attack on a religion and no one ever says "fair point" like they'll do when someone criticizes C or C++.
    17
  16. Because Rust fanbois are the worst type of programming language community, when you insist everyone redo everything in your cult language from scratch and people simply ask you to please stop and you turn into an angry nutjob over it, you're gonna piss off a lot of people, I hate Rust purely for its community alone.
    14
  17.  @ps5hasnogames55  "who cares about optics?" Anyone who wants the death of Windows or Apple, which includes Linus Torvalds. "linux doesn't have to serve anyone else" It serves those who use and build it for one. This is a foolish statement. "and especially not proprietary malware like eac" That's your perspective, its not our job to define how others should use the tools, we merely present the tools. Linus agrees, if you break something in userspace, you are the problem, not the userspace. "there's a reason linux kernel has so many bugs and is so bloated - the religion of "don't break userspace!" has created so much technical debt." Then why was Windows 7 considered the epitome of the best Operating Systems that Linux distros are still trying to emulate to this day? Your philosophy does not substantiate anything but a minority opinion that Linus does not agree with. And breaking backwards compatibility has never made Linux less buggy, most of the time it caused bugs. Desktop users don't need to care about bloat for such insignificant things and if you really need it, leave it as an option to remove the bloat, but don't remove it because of your unsubstantiated claims of "less buggy" despite being well more buggy still then Windows and Apple. The fact that both of them were usable for nearly three decades without complaint most of the time and that Linux in only trying to emulate aspects of them that it started to gain traction only demonstrates how worthless your claims are.
    14
  18.  @MiukuMac  Bytecode isn't the problem, its not storing the bytecode into memory that's the problem, this type of crap isn't a problem for things like Javascript and CSS (technically CSS doesn't use a bytecode, but structurally its not too dissimilar) as examples because they don't store it on disk, they keep it in memory no matter how long they wait on execution, the slight memory cost is completely irrelevant compared to the blocking that storing the bytecode would cause.
    13
  19. Still waiting for xfce wayland support, that honestly the only big thing keeping me on X, though I would really love if the streaming issues on wayland would be fixed before then.
    13
  20. When I saw this, I nearly flipped out laughing at the claim.
    13
  21.  @Zhafêu  I still don't get how you can have heap memory safety, which is the majority of the concern, when you can't even panic in the kernel, wouldn't that mean that Rust is no better then at least 20 other languages include C++ for this use case? The whole point of heap memory protections in Rust is that it does panic but if you can't do that, Rust isn't the first language to be able to report errors on memory issues. (although if that's a runtime check that also has a performance cost) And it isn't the only one that protects memory at compile time otherwise.
    13
  22. I don't see how Meta could collect any information from you that you don't already publicly give on a Fediverse account that isn't Threads, like from a technical perspective I don't see how this can even be considered a reasonable concern, I don't see why care about privacy here, if you care about privacy, just connect to the Fediverse without Threads, don't give information you don't want publicly known, sounds pretty simply, there's almost nothing to really get from say Mastodon that isn't already publicly accessible. Also its not like Threads is blocking the Fediverse from having NSFW on it, even then mainstream adoption will result in a lot more bigger Fediverse instances being less NSFW, but that doesn't prevent you from having access to them.
    13
  23. Do kinda wish we had the year as a major version, since I like knowing how out of date I am, but I guess its not that big a deal since generally a kernel shouldn't break backwards compatibility, so semver becomes a little pointless. Though there are cases where things are dropped and it would be at least nice to have that on a clear "major" change so I can be like "oh this is gone now" in case it would brick my system. (unlikely for me since my system is not even remotely niche nor old, but still, I think there are definitely tinkerers and devs that would appreciate it) Perhaps it be better to have dropped things on a yearly basis if you did that yearly number versioning, if I do anything big that I'd probably adopt that if semver doesn't make sense.
    12
  24. I intend to keep costing em as much money as possible for as long as possible and then leave.
    11
  25. IDK why anyone is defending RH in this thread, like the problem has nothing to do with the fact that the fix could cause more problems or reduce stability, its the fact that a contributor was given absolutely no information on fixing a bug "properly" from RH and was instead just told "we won't consider this", if the reason has to do with QA and regression testing, then just say that, foremost really should be asking the community to help with testing because they most certainly would, but even further not giving basic documentation on why the fix can't be accepted and the methodology to get it accepted, that's literally all you need, if someone sees a fix that has issues and says nothing but "I'm not merging this" what the hell do you think is gonna happen? How are contributors supposed to respond? And how does this demonstrate being more secure? Its not like none of the other distros don't also carry QA and regression testing, to say RHEL is unique here is just an outright lie, this response is just simply not okay, it wouldn't be okay to internal teams, its not okay either for external contributors, if you can't say it to a rookie on the production team, you can't just say it to a contributor.
    11
  26.  @robnobert  Fedora having a record of doing things before its stable seems like it make more sense to align itself with KDE, its Wayland support makes a lot more sense.
    11
  27. That doesn't require Threads, they can already do that without it, its social media, if you don't want something publicly accessible, don't give the social media that information, pretty simply.
    10
  28.  @jonnyso1  But most concerns are in dynamic allocations, which Rust doesn't fix without panics. Semantic info is only useful if its kept up to date which fails if any internals of the C functions changes, it will create false promises to Rust which will break it. I don't see where Rust actually helps at the kernel level if it can't do the things that would make it memory safe. (unless it were allowed to crash, then sure, but most exploits are at dynamic memory which requires runtime checks that normally would panic in Rust)
    10
  29. I like the lack of professional behavior, its not like this breaks usability or accessibility of the web page and its not making anything impossible to read, and 1s intervals aren't all that much of a concern, I think its funny because we got people who are enjoying themselves at least over people who demand conformity, which is kinda the whole point of Arch. Not entirely certainly on other things people might be mad at the Arch admins for but I find this hilarious and love watching stupid Arch folks get mad because a lot of them are stupid.
    10
  30. I honestly don't see how you could trigger a DDOS, most especially not a reliable DDOS at that, it would require so much coordination with a multitude of curl users anyway and you could just as well trigger the same exact problem by simply calling curl, and the only case I can see it being a vulnerability is if you openly give access to the retry options, and even then its not going to keep making an immediate attempt. How could that possibly cause a DDOS? I guess if the server is really crap but that would be a problem regardless of an invalid option value.
    10
  31.  @dustycarrier4413  The whole definition and term "open source" came from the FOSS philosophy, its not a monopoly, its an already objectively defined term.
    9
  32. Ew snaps, honestly there is very little reason to use a snap app when you have a package manager.
    9
  33. This is the reason I'm advocating that if you really need to build an app that's easy to cross-platform and develop on with a nice readable and performant language and great UI system that you can actually edit without opening a text file, use Godot, you can even strip it down, there is no need to JIT it, it still works decently as a web app, but most especially it runs as a native app. And if you really need another language or C++, you can just use one of the native bindings that doesn't require recompiling the whole engine. (and even if you're willing to do that, you can shrink it down to a 20-30 meg binary on Linux if all you want is core functionality)
    9
  34. The FSF would've won the war against proprietary code by now if they'd just been reasonable and approach people in an attempt to give them full choice and work as best as is reasonably possible to eliminate proprietary code. They honestly shoot FOSS in the foot the most out of any members of it.
    9
  35. Now that's a strawman if I've ever seen one.
    9
  36. The issue crops up more visibly on lower end systems and with lower refresh rates.
    8
  37.  @anon_y_mousse  Community is also quite toxic and fanatic, we all understand the flaws of C, Rust folks in my experience only ever get angry and throw tantrums if you point any out to them.
    8
  38.  @terrydaktyllus1320  "Just for the record, I can take an entirely "unbiased" attitude to systemd because I have used Gentoo Linux at home since 2003 where I have never had to deploy systemd but, at the same time, I have been in a job working on Red Hat servers for 20 years now and I, of course, have to know how to drive systemd as part of that job." You can't take an unbiased attitude with anything though because that's literally, logically, and physically impossible to do. Facts are inherently biased by nature, to consider yourself unbiased in any manner is by nature a manipulative and deceptive statement no matter how you slice it, anyone who says such things should be immediately distrusted. Also you prove your bias by using experience and anecdotes over specific arguments. Secondly this is complete nonsense, you have a job, in a job you choose to take, you don't get a choice on how to perform your job, you do what you are required of you, you don't get to violate the desires of your boss else you can't have a job, if you don't like that then you shouldn't have chosen the job and can quit said job, you always have a choice, but saying it like that is just try to hide it claiming "my job" as if you didn't chose your job foremost. You were never compelled to do that work, no job is ever forced upon, you have to accept it first. "I started using Linux back in 1996 and it has been my main OS since 2003 but if you asked me to remove systemd from Red Hat, I wouldn't even know where to start. It is core to the functionality of Red Hat and a huge number of other applications depend on it." 95% of those apps are FOSS already, some of them have non-systemd forks anyway, a lot of the time if people actually care the work was already done, in the few other cases its not especially hard for any specific app. Also what do you expect from Red Hat? Its their backend startup system, do you expect them to write everything from scratch when they have an easy to use system right there? Why should they write for anything else when it doesn't gain them anything? "and I would suggest you'll be out of luck there if you asked that of probably all distro developers." Haven't had a problem of this on Arch, I suppose with Debian and Debian-based distros it can be dumb, but that's because they're stupid distros anyway. "but that choice is restricted to what distros are available and how those distros have been built." Not really, choice has nothing to do with availability, this is a conflation of two distinct concepts, its not desirable not interesting and thus has not been done, people don't care enough and by the prospect of market forces nothing has been done. The only distro I can even think of where this is a massive problem is Debian and Fedora, which granted with Debian takes up most of the common users but its not like you don't have a choice of the distro in the first place, and you can distro hop regardless, and with most of the other distros besides them you do still get that choice. "Nothing stops you forking Fedora and rebuilding it to be systemd-less, but that's actually building a new distro, and that's beyond most people." Because its pointless, why would you chose Fedora in the first place if you hate systemd, that's quite stupid, how about instead of doing that just chose one of the distro/distro variants that don't freaking ship with systemd? There's plenty of those, Arch is filled with em, I don't see how this is a problem, it sounds more like making up an issue. And most people don't care anyhow, that's the reason the common use distros are systemd ones anyway. "Please try to be more objective in your thinking in future - so many people only look at the world from their perspective only, in your case "I like systemd, it works for me so everyone else should like it" and then try and "backwards engineer" rationales around such subjective opinion - and then end up being completely wrong." Given you kind of just used a fallacy here, that was a dumb response, you literally did that. Also this sounds kind of like a strawman. I mean you completely violated the fact that you've always had a choice, you're living proof, but yet you then also try to claim that your choice of distro isn't a goddamn choice, how does that make any logical sense? You just made a claim that you fundamentally don't exist.
    8
  39. "No major company with clear staying power is producing and supporting a desktop Linux system" So does this exclude Redhat, Canonical, or Pop_Os? The latter two of which has in fact shipped personal computing desktops in stores with Linux installed. Not to mention Canonical has also tried their hand at the mobile market at least once. And what about Framework? They have a lot of out of the box Linux support.
    8
  40. Legit makes more sense then rust, not to mention its easier for C devs to maintain.
    8
  41.  @MiukuMac  Considerably slower in what way? In a noticeable way? Probably not, if it really mattered, then they'd develop a JIT or AOT compiler, but bytecode can be made fast enough without it if you're not doing anything super complex, (like C++ or Rust level complex) and if you have type information, you can optimize for a lot of performance, also a DSL designed for a specific purpose can make a lot of presumptions that remove performance costs, the small cost of performance otherwise isn't a big deal when the margin isn't noticeable and won't stress out the system, even with a DE you have performance budget to waste a bit, you don't need hyper-optimized performance, you need to prove its a problem with profiling first. A RAM cached bytecode absolutely won't show up in a profiler. (and I'm saying this as someone who has written a bit of language parsers and bytecode stuff before) Often times easy maintenance is superior to having an obsession over performance, performance only matters when you can notice it. The problem of modern systems is that a lot of people ignore it even when it is noticeable, you could easily profile code to see a problem, I agree will addressing those issues, but I'm willing to bet you QML itself would never be the problem here.
    7
  42. 2FA has always felt like a half-assed attempt to reach the "I lost my crypto-wallet" problem for me, least as of late, its kinda frustrating that I need to have a device or I need to know the next device I'm gonna use for the 2FA, IDK I don't really think its the responsibility of the host to provide security, (that violates their convenience without a choice) its up to each individual to decide the risks for themselves. I just don't believe in collective nor centralized solutions, too bad nobody will get over them. And I'm saying this as an active contributor to multiple projects on github and so have to use 2FA, I use a browser addon for my 2FA, honestly I feel there's gotta be a better solution.
    7
  43.  @jonnyso1  That's not actually true though, safety isn't the issue for that point, its semantics, there is nothing in the standard library that would violate those assumptions because the the unsafe behavior below Rust is preserved through the legacy of the C standard library which is always backwards compatible, and if those semantics were to change it will break Rust. I don't expect the Rust standard library to either change semantics unless it wants to break all preexisting code, the C and C++ standard library is full of deprecations they can't get rid of for a reason, unless Rust does that it would fail to be a usable systems language at all. And that requires nearly as many Rust contributors just doing that as C ones, every change needs to be managed by both a Rust and C developer then. (and if that fails, Rust will break) Maintenance to that level is an unmanageable problem without a tool that forces it, which the kernel devs cannot use. (and I speak that from experience, even on a small level maintaining that is an impossible and often stupid task)
    7
  44. She still seems to rave about it like a fanatic which concerns me for if she'll ever actually consider the flaws if you were to argue with her on it, if that's the case, I would then disregard her opinion to the same regard as everyone else. I'm more angry at Rust fanaticism then I ever could be at the language itself, for which I personally despise but have not intended opposition towards. But its for that fanaticism that I don't want Rust anywhere near Linux now.
    7
  45. This is such a confusing exploit, so explicit merge keywords aren't implemented but implicit PR merging is? That's just dumb.
    6
  46. Its so sad that Richard Stallman died of ligma. Who the hell is Richard Stallman? Ligma Balls.
    6
  47. Its a problem with OpenGL because the extensions are basically hardware limited, with Wayland extensions aren't really a problem since any application can have its necessary extensions implemented and installed regardless of hardware and firmware allowing segregated functionality implementations that don't waste anything for those who don't want it, and with a package manager that becomes exceedingly trivial. (and if we rely on a package manager to install things, unless its an optional dependency, you don't even need to perform dependency checks unlike in OpenGL where you more or less had no choice) It also enables competition between the extensions that do similar things meaning a de facto standard will rise to top by being so much better then its alternatives. Another benefit of this is that an agreed upon standard for any specific thing could also devise communication layers so that nothing in the windowing system will step on anything else's toes. If done well, it can be a brilliant cross-section between performance, security, and functionality, however it does start to become harder to manage if you hate package managers, but honestly pretty much all distros are a massive nightmare without them anyway.
    6
  48. I think upstream and downstream package both server necessary purposes, it really depends on the specific application. (course if you're dependency tree is shaky and unstable, this demonstrates the perfect case when not sticking to upstream is a nightmare)
    6
  49. Part of the problem with this is that the specification and concept of Wayland wasn't finished nearly enough before non-demonstrative implementations were implemented. As it stands right now there really should only be one wayland implementation that's literally nothing but the bare basics of wayland's spec for which only really exists to compare functionality, capability, and usability against xorg. But since we have multiple "commercial" implementations its muddying the water and dirtying expectations for wayland. Doesn't help that the wayland designers are so autistic about security that they won't consider someone using the system in a way that they disagree with much like the gnome and KDE devs did/do. I take massive issue with this outlook because that's not up for them to decide, if we want higher security, you can build that, but leave us the option to make it easier or figure out a manner where we can both have our cake and eat it too. If you don't provide either you are ensuring nobody will want to use your mess of a philosophy disguised as a platform improvement. And in the least of things they should give us something that doesn't break everything when it doesn't need to, most functionality that wayland doesn't support isn't because it can't do it in a reasonable way, it really feels more like snobbery for which it does that.
    6
  50.  @tellsackett4572  I would suggest you also learn Zig and C3. (yes they're alpha languages, but honestly I see both as being able to go further since they maintain simplicity and safety)
    5