Comments by "George Albany" (@Spartan322) on "Brodie Robertson"
channel.
-
3
-
3
-
@arthurmoore9488 You don't need to perform tests for every build and you don't need to include them in your release builds, you shouldn't need to do that in the first place and that was half the problem. I've never seen a case where FOSS projects needed every single release to be compiled at the same time with a binary blob of any kind, if you need tests with binary files, it should be a separated thing from your release CI, literally every big name FOSS project does that, if you can't that's purely incompetence.
Also the makefiles aren't an issue without the binary data, as I told you before, but you seem to have a reading disability, if you can't distinguish between a bug and malice, there is simply nothing you could have done, but the issue here wasn't something that couldn't be distinguished that way, it was the allowance of binary data and the injection of behavior into systemd by distros, the vulnerability was open long ago, it was just a matter of time for someone to use these manners to accomplish this, the makefiles aren't a vector for an attack by themselves, you can't make a makefile malicious by itself in a way that isn't gonna set off alarm bells. (unless you modify make or the other automation tools, which if you're already doing that, messing with makefiles is the least of your damn problem at the point)
3
-
3
-
2
-
2
-
@TechJolt3d
"Rust syntax is verbose as all hell."
Necessary verbosity doesn't bug me, unnecessary verbosity does. Rust has a massive amount of unnecessary verbosity on that front, in C++ we have a solution to most of it coming at least, Rust has yet to even consider it.
"If i had a guess as to why they used let, its probably because they want you to rely on type inference, not on explicitly identifying the type."
C++ has auto and it does literally the same thing with also the capacity to provide type specification modifiers too, like references, mutable, and pointers.
"I think they want you to use the type identifier very, very rarely."
Two reasons this doesn't work and is dumb, I'll use Godot as an example. They banned the use of all type inference in their style guide. Simple reason why, they don't want to force anyone to require reading with a full semantic analyzer just to review code, type inference makes code review a hell because it requires otherwise unnecessary dependencies to perform code review and its easy to hide behaviors in that, especially when you have polymorphic types. This causes problems for reading code especially when explicit declarations could be just as if not shorter then type inference which is necessary for multi-person projects. If a project like Godot which tries to be easy to understand and simple, were to use something like Rust they would still need to ban the use of type inference because of that reason making the whole let thing literally interfering with the design of the project.
That aside it also creates a disconnect between class variables and other variables where none need exist, the distinction makes little sense specifically because they are both variables that share the same everything, the only difference being where they're located which I disagree is a justification for a distinct declaration shape, its more things to remember without qualitative reason to.
2
-
2
-
2
-
To be honest, I don't find "Not attempt to profit from open-source or other software that is otherwise generally available for free" all that reprehensible, like I don't know almost anything about the Windows Store and even I knew that reselling of FOSS projects without the dev consent was a problem, and despite my absolute hatred of Microsoft, how this is read isn't all that bad. Now do I think its a great decision if they enforced it as written for that part? No, it makes no provision for an otherwise free project to officially be distributed for a price on the store like Krita, individuals should be allowed to set their price for that, but I don't think I can hate the idea if it was enforced as written. I don't see how its banning FOSS with this though, seems like massive overreach on what its actually doing, FOSS isn't being banned, even selling FOSS wasn't literally banned, only selling FOSS on the Windows Store when its otherwise free would literally be banned which isn't as bad. As for the second part "nor be priced irrationally high relative to the features and functionality provided by your product." is incredibly vague and describes no objective standard for how that will be enforced hence its stupid, but if that were under an intent, it still doesn't ban FOSS, all it really does imply is stop overcharging people for a below price product, which fair, I don't like having that provision either but outside the vague language I don't have a problem with the intent from a privately managed storefront.
2
-
2
-
2
-
2
-
2
-
2
-
2
-
2
-
@jonnyso1 Maybe it could work, but in my experience managing a FOSS project of nowhere near comparable size to the Linux kernel, (and doesn't need nearly as many guarantees as the kernel does of course either) even having full control over the project managing compatibility with legacy behavior, especially when the language you use doesn't natively provide any semantic awareness, is an absolute nightmare, semantics behavior that is changed by internal behavior changes to a function pretty much throws the entire semantics guarantee out the window, (it could become downright impossible in some cases to work out correct semantics) while I do agree that it should be documented, decentralized management is always better then later centralized management, if you expected someone else to manage accounting for those changes, I am willing to bet a lot of people will have burnout, and if that happens to enough folks, it could kill development to either the Linux kernel or the Rust side, if the Rust side burns out then adding Rust to the kernel will have objectively been a mistake. At that point it basically would become a case of which collective actually has the harder head. Doesn't mean some of those C devs aren't at fault, but we also do need them for Linux to live so there kinda needs to be a limit to pissing them off too.
2
-
2
-
2
-
2
-
2
-
2
-
2
-
2
-
2
-
2
-
2
-
2
-
2
-
2
-
The real problem is that competition is being stunted by the government, and if consumers really cared about "right to repair" (which isn't a right, that's privilege that a government shouldn't be enforcing anyway) then they wouldn't continue to buy from John Deere, Apple, and Tesla as the consumer opinion would've initially started with "this is too locked down to use", but since they don't do this, its clear that these concepts do not actually improve the life of the consumer, only those who have a problem, which is a tiny minority of consumers. In essence its a form of tyranny by a minority.
Also if not for IP law, there would be nothing actually stopping someone from just breaking opening the functionality of something and building a business on circumventing the locked down nature of these companies, for which the company would lose business and would become naturally encouraged to either be more open or keep losing business. The only real reason nobody does this, and thus locked down companies can lock down their "IP" is specifically because selling services to this is extremely damaging as the government and corporate establishment have built the legal infrastructure to punish anyone who can actually pull this off, least to the degree that the company could notice the hole in their pocket.
2
-
2
-
2
-
2
-
2
-
2
-
2
-
2
-
2
-
2
-
Aerospace engineering honestly fails these standards way more then Linux does. These slides kinda deserved to be mocked, no software engineer can follow it to that degree in the first place and its especially stupid because none of the standards apply to any of their tools, none of the compilers follow that, and neither do any aerospace operating systems. There are reasons I could agree that Linux might not be suitable for the case, but what's described here is mostly false in that regard, the problem is complexity hides problems, especially bugs, the larger the source code and the more use cases to deal with the more complexity you have and thus more problems and bugs. As someone with experience in requirements, they don't fix this, requirements aren't designed to address this at all, even in safety standards, they make it easier to diagnosis, trace, and address, but they don't keep your software safe nor secure at all. Only reduced complexity does this and requirements don't enforce such a standard at all. Its conflating a completely different objective in safety-critical software engineering with requirements, which I say is stupid.
1
-
1
-
@Luxalpa
"The `value: type` thing is also a requirement from a language standpoint in order to make it context free"
Which I disagree with, I don't believe in the requirement of parsing necessitate a context free environment either, context is inherently described in all language, it is a requirement to understand language and the same to a degree applies to a parser, even with a parser it works fine for something simple but once you start making anything moderately complex context free becomes more of a hindrance especially for future development, what you gain in simplicity you lose in flexibility.
The same applies to use by people which require context to understand any anyhow, and with a programming language this always applies, in C++ I wouldn't confuse a function vs. a variable syntax just because the context is necessary to know, I have to read the whole declaration to know anything about it anyway. I have not experienced a case of reading nor writing of C++ that served a practical purpose where I would confuse function and variables for example, and its not that difficult to make it easy for human readers to not write it so it could be outright confusing to know such information. I've heard this claim before, most especially with declarations and I have always found it to be an excuse, context free is not nearly as context free with languages and I don't find it all that useful if you are building a new language, you can make the difference easy to recognize without adhering to context free, it just requires deliberate design, which you have do regardless of whether you make it context free anyhow, it saves you no time nor regards.
"syntax highlighting, error checking, type hints and auto-complete are all majorly messed up."
I've yet to find this as a problem on my end, I don't have experience with any language use that already does the things I say having this problem, C, C++, C#, Java, Dart, none of these things have caused me problems like that.
"Another issue is that for example rusts tuple syntax would be conflicting with its function-call syntax here (writing `value(i32)` would be incredibly messed up). A simple statement like `run_code(some_var)` would be unparsable,"
I don't know why you assume my issue suggests that Rust should change one part of itself and nothing else in order to satisfy me, like where you think I wouldn't understand that the language would need to change to accommodate such designs is beyond me, I am entirely aware that a language that does what I like would need to be something designed from the ground up to not be like Rust. Your claim of my character that I don't understand how languages work is quite inane, I don't have a well tested formally designed language under my belt but I have written my own parsers and languages before, I know how language design works in the least.
"Additionally, you have things like the `mut` (or `ref`) keyword making the entire thing even more messy."
Well that was already solved in C++ anyway, modifiers are trivial to resolve this type of stuff, I've written parsers that do just that, context aware parsers aren't really that hard, wiring up semantic understanding might be a bit of a chore but even with a decently produce AST its quite simple.
1
-
1
-
@SeekingTheLoveThatGodMeans7648 That's not even what Ecclesiastes 7:16 refers to, its caution against vain religion one believes to be righteous, the only correct interpretation of that text refers to rituals and ceremony, (say like demanding everyone must dress formally for church) not violations of God's Law, God's Law makes it clear that if one even should dress in a manner that would confuse the sexes he is a disgrace, the same would apply to those who act or speak in such manners. I won't speak in manners to enable confusion hence why I refuse to acknowledge the pronouns.
Also "they" is a terrible substitute because its a plural being used in a potentially plural context obscuring the information of the conversation, and I have direct and repeated experience of this fact causing confusion, including in the cases where people demanded they be called by such. "They" is linguistically a terrible pronoun for use for a singular person, it should be avoid wherever possible. (it also introduces higher capacity for overlapping characters covered thus requiring use of nouns to properly clarify the sentence for every party involved thus further invalidating its use)
As for when to witness, sure there are times of it, its not like I go preaching every time I come across someone doing this or demanding I use their pronouns, it takes them either asking relevant questions or saying something that brings up a moral dilemma, which usually sparks a full conversation in the first place. (unless of course getting into the conversation is banned in which case its to be avoided anyway)
I'm also not sure what you mean by mirroring "the salvational manners of God", only case where I can see that make sense is if you're a Pelagianist or Semi-Pelagianist since salvation isn't voluntarily elective nor a choice.
1
-
1
-
1
-
1
-
Google is not part of the private sector and honestly neither is any of the other browsers, not a single one of them isn't corporately run or funded thus they are not part of the private sector, corporations are literally just syndicates of the state, they're called corporations because they're the body of the state, they enforce government policy (see facebook and twitter colluding with the FBI as mere examples) corpus = body, corporations are a socialist concept, this is socialism. (not to mention that the government itself can decide at any point without any oversight to ban, halt, and nationalize any business it wishes, meaning the government owns the business, not the people running the business, he who can halt the means of production owns the means of production and is thus responsible for what the means of production does)
1