Linus is back

>>106200568 (OP)
hes getting old
also wouldnt an union be better?
this construct implies a runtime cost

I don't even deal with C but he's right, a function that occupies exactly one line is useless.

>>106200568 (OP)
Based chud Linus throwing dildos at WNBA.

>>106200605
there is always a runtime cost to this unless both values are constant in which case it's constant either way

>>106200631
not in c
oftentimes you want a wrapper around your internals to package the spaghetti
in extended C you can force inline so this comes at zero runtime costs. pure syntactical sugar

>>106200662
an union shouldnt incur any runtime cost beyond copying
i expect the other construct to be optimized into a union, but i still prefer to state explicitly what i expect from the computer when i can

>>106200631
>>106200675
cont.
this isnt a function, its a macro
but this is the kind of shit you might want to hide behind an interface for legibility and maintenance sake
also duplicate code is always bad. yes you can use refactoring tools but even if youre nigh-certain you wont make a mistake, the probability is never 0
#define STRINGIFY(target) (t_str *)(&(const struct { size_t size; char text[sizeof(target) + 7]; }){ sizeof(target) - 1, target "\0\0\0\0\0\0\0"})

>>106200728
It's insane that people in 2025 need to think about this shit and write code like this.

>>106200605
>also wouldnt an union be better?
Explain what you mean. In detail.

>>106200728
Nice programming methodology blog post, Rupesh Patel.

>>106200739
believe it or not, people also still have to go manually dig sand from the ground to make your computer that runs your vibe coding stack

>>106200739
why?
the first step of optimization is to tailor your infra to the data youre working with
and thats how you might end up with code like this
even if this is more the result of tailoring a type to the architecture its running on
its a matter of usecases

>>106200739

What’s insane is that “software developers” are still being employed, to the detriment of actual computer programmers who know how everything actually works.

>>106200675
Most obvious nocoder I've seen in a while.

>>106200763
>nothing technical to say whatsoever
>t. nocodeshitter

>>106200748
ez
union bw an array of 2 shorts and an unsigned int

>>106200765
Yes and there are people who movie to the forest and live off the grid. It is doable, but completely unnecessary.

>>106200779
>no technical argument whatsoever
>t. nocodeshitter

btw
why are you seething?
feeling outclassed or something?

>>106200781
>>106200794

>>106200825
>still nothing technical to say
by definition its off topc

brehs I just realized I'm not a programmer should I commit sudoku?

why is he so rude

>>106200858
it makes him feel good about himself

>>106200858
angry old man is yelling at interfaces or something

>>106200858
the only sensible response when you peak at that level

Should've added another line with a pro vax comment, he woulda sucked you off good then

Wonder how many teeths he got left

>>106200568 (OP)
didn't read. not my problem

>>106200568 (OP)
imagine being the linux janitor and after like 25 years you have basically nothing to show for it except that microsoft infiltrated and made a joke out of the idea of what it was supposed to represent. you would probably be grumbly too

>>106200902
>DA VAX
Really, nigger? In the year of our Sneed 2025?

>>106200916
Same, I run bsd

>>106200923
android is linux based. literally the most widespread OS on the planet, so much that nothing comes close to it. you must be lost

>>106200675
>oftentimes you want a wrapper around your internals to package the spaghetti
why would you "package the spaghetti" in the first place? it does not make any sense
>in extended C you can force inline so this comes at zero runtime costs. pure syntactical sugar
wrong, inline does not always work as you would expect as it's just a hint to the compiler.
>>106200728
>but this is the kind of shit you might want to hide behind an interface for legibility and maintenance sake
no you don't, wtf. linus explained why and he is 100% right, read his message again, alos hiding stuff is always a bad reason to extract code into a function or another file, that's yet another reading indirection and add complexity to the mental model of the code.
10 000 lines functions are perfectly fine for this exact reason, code flow is very important and has 0 negative impact (in fact quite the opposite as no inline guessing required) on the generated machine code.
>also duplicate code is always bad.
this is absolutely wrong, why do you have opinions on things when you're clearly a student or an actual nocoder?
duplicating code is very fine and dramatically simplify the code sometimes and more importantly remove ambiguities and allow similar code to diverge with time.
good code is code you can remove easily, your fucntion has a very high chance to see its prototype changes wtih time, taht's a maintainance nightmare introduced for no fucking reason, it's especially true the closer to the machine you are like in these patches.
>yes you can use refactoring tools but even if youre nigh-certain you wont make a mistake, the probability is never 0
don't rely on these things.

why do you reason in absolutes? code duplication is good or bad depending on context, same for obfuscating (yes that's what functions do) code inside functions or macros

>>106201073
>linus explained why and he is 100% right,
kek. Bro doesn't know...

>>106200568 (OP)
Imagine writing all that shit over 1 stupid function.

>>106200694
First of all, using union makes code byte order dependent.
Second, smart compiler will transform this to bit-shift and bit-or anyway. Doing this through memory will be expensive, as CPU cannot forward the data of 2 stores to 1 load internally, it will have to wait for stores to be combined, which means other stores will need to complete first.
Basically, Linus is 100% right.

>>106200568 (OP)
what if you make one argument an u32 so you know it is the high word?

>>106200839
>Low effort, defeated reply
I accept your concession. Thanks for the (you)s.

>>106201093
>Content free post
Fascinating.

>>106201551
It's the call site that matters you dumb nigger.

useless loser, only reason he's not homeless is Bill Gates donating millions to the linux foundation

>>106200923
He made a very successful career out of it so not sure what you're on

>>106201073
>why would you "package the spaghetti" in the first place? it does not make any sense
legibility.
>wrong, inline does not always work as you would expect as it's just a hint to the compiler.
*extended C
as in
__attribute__((always_inline)) inline ...


>linus explained why and he is 100% right,
you dont have the knowledge to discuss this matter
faboyism is tiresome

>>106201245
>First of all, using union makes code byte order dependent.
and a bit shift doesnt, right?

ugh
the absoilute state of fanbois

>>106200568 (OP)
Valid crashout. Retards should not submit shit, especially that late.

>>106200568 (OP)
this is the kid that linus is dabbing on btw

>>106201245
>>106201965
actually
im wrong about the bitshift
didnt know that
'shouldnt have called you a fanboi, have a you

>>106201988
>(((Google))) employed
Deserved

>dabbelt.com
>http://dabbelt.com/~palmer/
I was expecting an Indian. How does a white man stoop so low.

>>106200605
it's heckin' UB which makes trannies dilate, even though every (relevant) compiler on every architecture does with it exactly what any sane person would expect (though then again it is the linux kernel so if not using UB is relevant, ever, it is here)

>>106202087
ub?
the c standard is so convoluted
youre supposed to use a union to avoid type punning
but when it comes to unionizing an array its ub?
completely fucking retarded

thankfully my shit doent need to be portable
i dont need to care about that

>>106201988
I fucking hate young people!

>>106202149
i try to be less grumpy as i get older cause i don't want to turn into my father, but i can see why it happens, once you've dealt with the same shit for the 4000th time it's hard not to sperg out

>>106202295
its interfaces shit
you got two elements
its turbointuitive
theres no excuse to take out your nerves on someone for that
esp the kids position is correct
unless hes breaking a rule in the linux norm

>unless hes breaking a rule in the linux norm
which i suspect is the case because chimping out over an abstraction which most likely is a macro just doesnt make sense to me
what would make sense to me is that theres a mechanism of validation whenever someone wants to add an interface to the linux codebase
and then yeah, i completely understand loonis getting mad over it

>>106200568 (OP)
The old Linus would have insulted the person for writing it
The new Linus says the code is dumb

literally who. why would i possibly care.

>>106200568 (OP)
So, the problem daughter finally got knocked up.

>>106200568 (OP)
>Linus just violated his own CoC with that hostile response
time to remove him from Linux

>>106203338
No, his victim was a white cisgender man so it's fine.

>>106203338
>CoC
lol, LMAO, LELELELELEL

>>106202010
you're wrong about everything you fucking moron, there is no such thing as extended C even the forceinline of MSVC does not force inline it's why C++ guys decide to basically makes it a noop in recent versions of C++

He's Jewish. At least spiritually

>>106200568 (OP)
Isn't Linus jabbed though?

>I am currently employed at Google, where I work on the Android team
Tells me all I need to know.
Good on you Linus.

>>106200568 (OP)
Why not just call it make_u32_from_two_u16(high_word, low_word)

>>106203813
parameters aren't required to be named and named parameters don't really help the person reading it since now they have to look up the function declaration to understand what's happening

>>106200947
I wouldn't be very proud of making India's number 1 OS.

Abstractiontroons absolutely blown the fuck out.

>>106203833
>parameters aren't required to be named
wat?
>named parameters don't really help the person reading it since now they have to look up the function declaration to understand what's happening
What do you mean? You get the parameter name while writing the code from intellisense, and you can hover over a function name to see the declaration.
I understand Linus wouldn't want to add helped functions for this, but I don't understand why he's sperging out about the word order in particular.

>>106203890
>you need intellisense for code to be usable and readable
Kill yourself.

>>106203898
>all troonix devs just know all function declarations by heart

>>106203964
That's exactly the problem you stupid tranny. We don't have time to waste on figuring out how to call a function that's an obscured fucking oneliner that we know by heart.

>>106203989
You still haven't answered the question.

>>106201988
DESERVED
E
S
E
R
V
E
D

>>106200605
He really needs to find a successor, someone with a strong personality because we all know how projects go once a committee takes over.

>>106204052
The answer you need will come once you kill yourself and God will explain.

>>106200568 (OP)
I just imagine him getting progressively madder as it go, all his therapy and relaxation techniques going right out the window and into the shitter

>>106204082
>all his therapy and relaxation techniques going right out the window and into the shitter
usually all it takes is for some utter fucking retard to make your life harder for no apparent reason, crushing what little joy you had left in existing.

>>106204130
Yeah, poor guy.

>>106203715
>jealoius
i know you are
you want to devaluate my skills so that you can cope better with the fact theyre unreachable to you

>no such thing as extended c
ah, so youre the same retard who cant comprehend the concept of abstraction
>MSVC
its garbo. because it doesnt have force inlines, duh
gcc does

>>106204077
whoever its gonna be theyre gonna be 10x more woke

>>106204301
>I have autism

>>106204314
that was very insulting
im a polymath

>>106204271
>you want to devaluate my skills so that you can cope better with the fact theyre unreachable to you
kek
>ah, so youre the same retard who cant comprehend the concept of abstraction
most abstraction in programming are retarded.
>its garbo. because it doesnt have force inlines, duh
no compiler does
>gcc does
absolutely not

>>106201988
I bet he cried and will sue Linus for hurting his feelings

>>106204415
>how to tell someone youre autistic without actually saying it
look it up. always_inline is equivalent to a macro

i wont even address the rest its retarded drivel coming from a handicapped' brain

>Reduced instruction set computer
>~1000 pages

What went wrong?

>>106204665
Humans. Design by committee. All good creations come from 1 to 3 people. What they create is then gradually corrupted and ultimately destroyed by the masses.

>>106204481
>>how to tell someone youre autistic without actually saying it
I'm not a freak, I'd kill myself if I were autistic
>i wont even address the rest its retarded drivel coming from a handicapped' brain
it's true though, gcc and clang won't enforce inlining and MSVC has forceinline but tehre is still no guarantee it works as intended.
facts don't give a fuck about your opinions, not sure what you're mad about

>>106204776
<from the horses' mouth
youre using obsolete programming methods

>>106200568 (OP)
ETL big pharma cocksock NAFO tranny throws yet another embarrassing tantrum. This is why Tranux will never be mainstream.

>>106204776
>>106204842
>no mea culpa
>not even an ACKnowledgment

*puff*
honorless scum

>>106200568 (OP)
I didn't get the part about the cast. What's he supposed to cast? Is a mask a cast?

>>106204842
>>106204936
can'ty you fucking read your own screen? read the last line, it's the same issue with inline..., you CANT rely on inline, forced or otherwise.
inline is worthless if it does not work across compilation units

>>106205053
>indirect calls
>yfw

>>106204679
Obviously I didn't read the whole thing yet, but did they add mmx yet? I'm not even thinking in matrix math and the determinant of deez nuts etc

>>106205053
also i just read the rest of your hysterical kvetching
if you want to use force inlines across compilation units you can use -flto
theres probably a flag that allows for cc/FI specifically but this one does it.

>>106203738
what is this image from

>>106203890
>wat?
jesus christ learn a real programming language, stopped reading your post right there

>>106200568 (OP)
He still has little sperg moments like this, but they're far weaker than they used to be. I don't think he even uses profanity anymore. A far cry from old Linus.

>>106203715
Actually, GCC always_inline attribute (almost) guarantees inlining and generates a warning if it was unable to do so (for technical reasons, like vararg functions are non inlineable). This inclu des -O0.
>why C++ guys decide to basically makes it a noop in recent versions of C++
I am not sure what you're saying, because "C++ guys" (the committee?) do not decide what optimisations to apply. Last time I checked both Clang and GCC did take "online" into account. In Clang's case this is most visible in -Os builds which sets inline threshold to 75, while inline keyword setsmit to 487 for all optimization modes.

>>106203890
>You get the parameter name while writing the code from intellisense
I am sure that the ancient version of uEmacs that Linus maintains personally for himself doesn't do intellisense.

>>106205759
He surely has something open where he looks up how to call functions then. Nobody's going to have thousands of functions memorized, many of which have like 10 fucking parameters.

>>106200782
both digging up sand and writing systems code is completely necessary, retard.

>>106201965
>and a bit shift doesnt, right

>>106205805
yea yea yea i looked it up
i apologized later on
i never had to deal with endianness

>>106200568 (OP)
>That thing makes the world actively a worse place to live
This is why the kernel is such a pleasant piece of software to use compared to nearly anything else. You might not like it but this is what peak open source maintainership looks like.

>>106201988
Of course it's an Android dev.

>>106200568 (OP)
I'm so glad I don't have autism like this.

>>106203338
Correct

>>106200739
>It's insane that people in 2025 need to think about this shit and write code like this.

You are the problem here sir

it's over. The troons took over and only thing Linus can do is seethe against some good willing risc-v developers.

>>106200858
>why is he so rude

Because free code for user freedom should be legible

Why does your spelling teacher get rude when you type in emojis

>>106200568 (OP)

>>106200782
The difference between the off grid guy and you is that I respect him.

>>106201644
>Bill Gates donating millions to the linux foundation

Source? Good for Microsoft man

>>106202017
because incompetence has nothing to do with race

>>106206235
You're severely homosexual and retarded

>>106200568 (OP)
Just read the manual chudnus

>>106206235
It's just simple statistics. Some ethnicities are smarter than others on average.

>>106206235
sure, as long as you ignore statistics

>>106203338
>Code of Conduct
>Which CPU does that execute on?

HAHAHA GOOD ONE FRIENDO

>>106203833
spotted the fentanyl user

>>106203890
Is it really that hard to understand?
On a kernel basic operations like shifts shouldn't be made into static inline function or macros because THEY ALREADY ARE FUNCTIONS + is a function, | is a function << is a function if you use them then you don't have 1000 functions to keep track of. Only once the amount of functions pass what the average programmer can hold in their head should they reach for either a generalization or abstraction

>>106203738
imagine taking some mystery substance pushed questionably hard by your media and government
ngl only alpha males had the balls to take it

>>106206546
That's fine, but the spergout about muh word order in the arguments still seems retarded/irrelevant.

>>106206573
Not really considering that people are more liable to use a random function incorrectly than a primitive.

>>106206588
The argument names can literally just tell you what they are supposed to be.

>>106206602
And they can literally not.

>>106206609
>one of the arguments is called high_word
>the other argument is called low_word
>wonder which one is which

>>106206619
>one of the arguments is called a
>the other argument is called b
>wonder which one is which

>>106206628
a and b ARE retarded, obviously
Cniles love to use naming schemes like that for some reason

>>106203788

>currently employed at Google
>where I work on the art team

>>106206653
high_word and low_word are just as bad they have no semantic meaning and their only purpose is to make some worthless function "usable". Now imagine when they put them in backwards or name the low high and the high low. Seems like a horrible bug.

>>106206677
It's not ideal to do this but like, the high word is literally the high word, and the low word is literally the low word. It's the same pattern as having the AX register divided into AH and AL registers but with words.

>>106206705
It's like talking to a brick wall, no wonder linus just calls you retarded and moves on.

>>106203890

I am not a programmer and can understand that whatever tutorial program you're using isn't what's being discussed (C kernel)

>>106200929
>I'm a cuck
weird thing to announce online, guy.

>>106200923
>nasa uses linux on the mars rover, rockets, satelites, etc
>100% of supercomputers world wide run linux
>80%+ of servers world wide run linux
>the most popular smartphone/tablet operating system runs linux
>microsoft servers like azure all run linux
>linux is the only kernel used by world governments for defensive purposes, e.g. the us army
-
>b-but muh games
you have to be 18 to post here

>>106206663
thought he went to work for facebook

>>106204679
>Humans. Design by committee

Bureaucracy is what we call that in the states. Principled design turns into cash grab by large group.

>>106205789
>Nobody's going to have thousands of functions memorized, many of which have like 10 fucking parameters.

Do faster guys than me just write them as needed?

>>106205619

BLOOD+
Episode 23
00:21:13 - 00:21:17
~96.59% Similarity

>>106207139
imagine posting the one good anime released in 2005-2006

in Windows header files
#define MAKEWORD(lb, hb) ((WORD)(((BYTE)(((DWORD_PTR)(a)) & 0xff)) | ((WORD)((BYTE)(((DWORD_PTR)(b)) & 0xff))) << 8))
#define MAKELONG(lw, hw) ((LONG)(((WORD)(((DWORD_PTR)(a)) & 0xffff)) | ((DWORD)((WORD)(((DWORD_PTR)(b)) & 0xffff))) << 16))
#define LOWORD(l) ((WORD)(((DWORD_PTR)(l)) & 0xffff))
#define HIWORD(l) ((WORD)((((DWORD_PTR)(l)) >> 16) & 0xffff))
#define LOBYTE(w) ((BYTE)(((DWORD_PTR)(w)) & 0xff))
#define HIBYTE(w) ((BYTE)((((DWORD_PTR)(w)) >> 8) & 0xff))


Lintroons are just seething because someone tried to make the code easier to read. Imagine typing out all the crap on the right, imagine trying to read out all that crap on the right instead of just using 1 macro. Memorizing how the macro works is easier than reading that crap every time.
Linus is just an egotistical faggot who thinks that real programmers need to abide by his personal schedule (I'm TRAVELLING ok, that means you need to finish your work FASTER). And if he doesn't like the macro, he should edit it out himself rather than bitching at the people who wrote large amount of valuable code for his project.
This is why only trannies work for open sores. Imagine doing this for free.

>>106203813
>>106207488
KEK exceedingly hilarious that the only two people ITT who thought this was a good idea have swapped parameters.

>>106207517
I didn't swap the parameters. It's from the Windows header files. If you do Windows development you just memorize how the macro works and be done with it. Intellisense will tell you the order of the parameters while you're writing the code and you don't have to worry.
Literally anything is better than ((WORD)(((BYTE)(((DWORD_PTR)(a)) & 0xff)) | ((WORD)((BYTE)(((DWORD_PTR)(b)) & 0xff))) << 8)), and while you might be able to get away from some of the casts in SOME applications, it's still not very good is it.

>>106207517
Well I just did it the way Linus assumed it without thinking about it much, but the Windows way actually makes more sense, so I'd go with that instead.
To me it still doesn't feel as much of a niggerlicious thing to write as Linus makes it out to be.

>>106207582
Those casts are unnecessary

>>106207488
why does it use a and b?

>>106207622
the original macro uses a and b and i changed it to lw and hw so people can see what it does without reading it. of course it doesn't work in that condition because I didn't change the code. they should've used lw and hw to begin with, I think.
>>106207613
they are necessary if someone does something silly like pass a float/double into the macro. you don't need it if you are writing out the code by youself and you already know the data types involved.

>>106207622
Isn't it just a for argument1, b for argument2, etc?

>>106207582
>I didn't swap the parameters. I
>>106207488
Kek you lying fucking nigger, the MAKEWORD and MAKELONG macros use a and b you just added lb, hb and lw, hw.
You literally did swap the parameters from that last guy and lied to me about not adding them. And you clearly added them because you left a and b in the code.

>>106207635
actually they are there for that and 16 bit support, which of course isn't needed on risc-v or anything past win32. Those macros have been unused since 16bit windows.

>>106207660
>Those macros have been unused since 16bit windows
this is not true at all
if you do a search in the leaked XP source code for *.c files that reference MAKELONG, you get 2000+ results, many of which are in NT-specific components (aka not win16 or win16 compatibility).

>>106207488
>This is why only trannies work for open sores. Imagine doing this for free.

>imagine not letting the cop chip tell you what code you can execute on your property

You Apple fags just dont get it

>>106207685
xp supports 16 bit

>>106204665
To funnel govt/VC money into your startup, you need to overcomplicate simple matters enough that your investors get blinded in awe at how smart you are while shutting down any straw of critical thinking they might have.

>>106207720
so? windows "supports 16 bit" up to win10. and I just told you the macros are used in code which only even supports 32-bit and 64-bit. MAKEWORD and MAKELONG are both used extensively in the networking code - anything that calls WSAStartup typically uses MAKEWORD to form the first parameter. Anything that uses any Windows common control you'd use LOWORD and HIWORD frequently. The macros are not just "for 16 bit", they are useful in general.

>>106207635
VS gives you the full definition. It uses a and b in the windows header files, but this same function is used all across github with lowword, lw, l, a, etc.

>>106207766
Why does this show code and not a spec?

>>106202143
>youre supposed to use a union to avoid type punning
What you're suggesting IS type punning. The main use cases of unions in C are:
- Reusing memory without having to reallocate (very important in 1970)
- Data whose type depends on runtime values, e.g:
struct a {
enum {FLOAT, INT, UINT} type;
union {
float f;
int i;
unsigned int u;
} value;
};

>>106207254
I don't have to imagine. I am living the dream.

>>106207759
>>106207488
>MAKEWORD and MAKELONG are both used extensively in the networking code
Yeah you have htons etc for network code so you can write portable code. You're doing the opposite in the kernel, that *is* the abstraction the portable stuff sits on. Linus is right to be upset.

>>106207488
> easier to read
i can read it just fine without coping using defines. and it's not about it looking pretty. it's about introducing needless trash into headers.
> something something trannies
you're entire cringe post reads like it was written by the permanently unemployed that would fail at microsoft BASIC. you're not fooling anybody here, ramashit inachildshand. good lord no.

>>106200858
honestly if you have to deal w/ retards 24/7, you would be rude too.

>>106205053
Stop posting and take the L

>>106205759
tbqh if he uses emacs or vim he should just kill himself honestly

>>106200568 (OP)
kek what a gay cunt

>>106207923
ah
thanks for the clarif. i must have misremembered something
im self taught so i have nigger-tier vocabulary at times

but theres one more "canonical" use for unions when working on gcc at least:
type punning through an union is the preferred way of doing things because its idiomatic to gcc when you wanna tell it to avoid strict aliasing optimization for the resulting pointers
i never thought about that
i always recast like a grug, never had a bug, but it can make things go wrong, apparently

Damn as an embedded systems programmer I'm fucking guilty of that dude's 'helper' functions and I read Linus' reply and it makes sense to me.
So now, when I properly use inline functions? When all these little 'abstractions' are more of a detriment to my code? Or is Linus being a little too spergy right here and better not worry?

Linus needs to go, dude can't stay forever like a dictator.

>>106210953
As long as your argument names are clear, I don't see why it matters honestly.

>wahhh politicsssss wahhh the worldddd heckin is it good is it heckin bad ahhh the worlddddd mattterssss

holy fucking cringe, linux is OVER.

>>106210953
He's a dumbass boomer sucking his own dick. Desperately wants it to still be 1993. If HE doesn't like it then it's LE BAD.

>>106200568 (OP)
this should have a one line review. "Remove this macro/function from this common header", that's all he had to say. Instead, he decided to be an asshole over the most irrelevant shit imaginable. Sometimes I think that Linux succeeded despite Linus, not because of him.

>>106210953
Its fine to have one. Just keep it in your module. Linus argued against it being in a public header where anyone can use it, not against the function itself.

>>106212132
The guy deserved it for trying to push garbage to common headers. Things that end up there become hard to remove or refactor later. Read about the mess min/max macros were in the past...
It is a common pattern among young programmers, so Linus probably sees this a lot. As you get older, seeing those things again and again becomes annoying.

>>106212186
Is that a fact or an opinion? Having common shit that the C language lacks in some common header is perfectly reasonable. The argument that "a << 16 | b" is easier to understand than "make_u32_from_two_u16(a, b)" is frankly retarded, even in the context of the kernel. If he doesn't like that, that's fine, he is the maintainer and he has the right to require contributors to accommodate his preferences, but this is not an absolute and it is not obvious. It just shows a lack of objectivity and empathy.

>>106212268
>Is that a fact or an opinion?
That public things get used and can later be hard to change? It is a fact, yes.

>>106212268
For reference, there is a bit about history of min() and max(): https://lwn.net/Articles/983965

>>106200568 (OP)
> Actively a worse place to live
sounds like he's describing Linux

>>106212340
This is just min/max sucking ass in C. There's nothing inherently wrong with reasonable macros.

>>106212442
>There's nothing inherently wrong with reasonable macros.
Which ones are reasonable then? Some are harder than others, but the function in OP is just not good. Just use shift+bitor.

>>106212468
It's bad because a and b is ambiguous which is easily fixable. It's not going to generate a massive expansion like a typesafe min/max would.

>>106212468
This doesn't even have to be a macro, it could easily be an inline strongly typed function "u32 make_u32(u16, u16)" that avoids all the issues that come from macros and the absolute inability of C to deal with generic functions. The problems of min/max macros are not relevant here.

>>106212499
>The problems of min/max macros are not relevant here.
My point wasn't about macros specifically, but about any "utilities", macros or functions. I just used min()/max() macros as an example of how much of a mess this can be.
>u32 make_u32(u16, u16)
You will need one for making u32 from 4 u8 too, so the name will have to be changed. At this point it will probably be longer than just doing "((u32) a << 16) | b". And you won't have to remember the argument order.

PS: The macro expansions Linux had were even that bad. I have seen expansions so large they crashed multiple compilers due to stack overflow (parsers are usually built using recursion).

If someone replied to me like that whilst I was contributing to their open source project, I would simply stop contributing to that project.
I wouldn't even accept that kind of response at work, where they pay me actual money to do things I wouldn't do by choice. Why do people put up with this for free?

>>106212899
>Why do people put up with this for free?
Because they need this more than Linus needs them.

>>106212915
What are they actually getting out of this? Having their modules in the kernel so they don't have to run out-of-kernel code?
I think if I was this RISC-V person I would start pushing for more out-of-kernel modifications after this. I know it would hurt kernel distribution, but I think it would put pressure on Linus to stop being such a cunt.

>>106200568 (OP)
What a passive-aggressive tranny

>>106200568 (OP)
Linus is black! And he is a woman now too!

>>106212899
>I would simply stop contributing to that project.
you wouldn't get such an email from me if you gave me pajeet level cancer, you'd be fired and i would make sure that nobody in my business acknowledges that you ever worked here. wouldn't even give you a reference for your next employer. you faggots on this board live in absolute fantasy land. it's unreal

>>106213007
Where I work, someone who spoke to me like Linus would be the one in trouble, even if he was 1 or 2 levels above me. My code could me the worst, most dogshit code ever written, and it would still play out this way.
I don't care what you would do, because we don't work together.

>>106200568 (OP)
>Linus is back
No he's not. He's too far gone. Unless he backs up and takes Rust out of the kernel, he's a still retarded senile fool who needs to be gently but quickly moved out to pasture. He's like the original Void "developer": started good, ended.

>>106212268
>Having common shit that the C language lacks in some common header is perfectly reasonable.
maybe for your gay hobby project nobody is ever going to use but i thank god every day that people such as yourself are kept far away from anything important.

>>106213028
> noooo! you can't just remind me of how reality works
you just don't get it, do you? if you submitted such shit code you would NOT get an email, you would be unemployed. what a fucking idiot.

>>106213049
>you just don't get it, do you? if you submitted such shit code you would NOT get an email, you would be unemployed. what a fucking idiot.
But I wouldn't be unemployed. The person who sent me that email would be talking to HR.
Labor laws mean I couldn't be fired for submitting one bad code commit.

>>106200568 (OP) Make it a wallpaper.

Linus is cool and all, but Linux is pure garbage, so even with his fun mails I'll not use his garbage OS.

>>106213028
next time you contribute code to my project, do things the right way

>>106213061
>1920x1080
>2025

>>106213049
Wait do you work at-will? That's so funny. No social contract for you.
>>106213070
That'd be fine. I don't mind criticism.

>>106213061
>but Linux is pure garbage
found another filtered idiot lmfao

>>106213082
>I'm so productive...
you produce nothing with your garbage fanboi OS. deal

>>106213089
what do you know about I produce? Linux is on every single device on planet earth

keep coping and kys while you're at it

>>106213116
You're here wasting your time because you have nothing to do with your life. I came here make a point. Point made. Now I'm leaving. Enjoy your hellish life, loser.

>>106213164
>You're here wasting your time because you have nothing to do with your life
nice projecting, dumbass. are you running away scared now? lmfaoo

i accept your concession.

>>106212899
It's called critisism you fragile little baby. If you can't take the heat, get the fuck out of the kitchen.

Code is WAR!

>>106201988
>My calendar is available online
this is the most cuck shit I've read this year

>>106200568 (OP)
link? the screencap is too low contrast and the line height is too small

> OK, sorry. I've been dropping the ball lately and it kind of piled up as taking a bunch of stuff late, but that just leads to me making mistakes. So I'll stop being late, and hopefully that helps with the quality issues.
NOTHINGBURGER

>>106200568 (OP)
Call him a nigger, Linus. You know you want to.

>>106213627
see >>106201988
dumb retard

>>106200858
Lack of patience, he must have seen some shit.

>>106212132
no, Linus is known for his outrage and that's what makes him great and based

>>106204665
I mean compared to intels 5000+ pages it's basically nothing.

>>106214154
x86 can do 1000x more though

>>106200568 (OP)
he could just have renamed a, b to high, low

>>106210953
When you are designing a headerfile in 2025 you should think very carefully. I treat every headerfile I make like I'm designing an operating system user environment for a savvy but lazy user. You have to think "what would I do in my first month of programming" and then write that solution, then make sure it's obvious how to transform that solution into the correct one.
If that's your main goal then you are rarely going to write extra functions.
https://harmful.cat-v.org/cat-v/unix_prog_design.pdf

antifa lover himself

>>106200568 (OP)
bros what will happen to linux once he is gone

>>106212963
They get what humans crave the most Clout.

>>106214630
Why do you keep posting this richard spencer clip?

>>106200568 (OP)
why did you misrepresnt linus, anon?
the core issue is not the interface
its the fact that helper abstractions get into generic headers
https://lore.kernel.org/lkml/CAHk-=wjLCqUUWd8DzG+xsOn-yVL0Q=O35U9D6j6=2DUWX52ghQ@mail.gmail.com/

you made linus look like a tyrant but his is actually a proportionate response given the context

just submitted a pull request to remove linus from the kernel due to this clear violation of the linux coc

>>106215804
ok, wo what did he do to cross the lgbtranny gang?
is he discussing dropping the support for rust or something?

>>106200568 (OP)
Unprofessional, toxic, unnecessarily aggressive response. Also unnecessarily long. Basically just a rant.
As much as I hate receiving emails written by an LLM, Torvalds would genuinely benefit from passing all his patch rejection mails through chatgpt with a prompt like "rewrite this to be calmer, more professional, and more concise".

>>106215746
>proportionate response given the context
He could have just said "don't include these functions in common headers" like any level headed person would do instead of having a pointless melty.

>>106215921
actually more projects would benefit from the maintainer being a massive fucking asshole

unironically

>>106215956
yeah but imagine having said that
put that into a norm which everyone is supposed to knpw
and then still had to repeat that to 400 people separately
and it isnt some obscure rule
its just- dont put your helpers in generic headers
its literal common sense if you have a modicum of organizational conscience

imagine if among the dozens of contributors nobody respected that rule
and you'd have multiple discrete helpers for the exact same operations

i agree the form isnt optimal but given the circumstances i think its still proportionate

>>106216083
>and it isnt some obscure rule
Says who? There's absolutely nothing unreasonable about thinking some simple byte functions could be in common headers.

>>106216109
>imagine if among the dozens of contributors nobody respected that rule
>and you'd have multiple discrete helpers for the exact same operations
organizationally, its a mess
and then someone would have to audit that...

>>106216109
>>106216118
also its bad fucking manners
implicitly its like saying
>i think my interfaces are better than yours and therefore you should use mine

>>106216118
Having generic functions in generic headers is the opposite of a mess.

>>106216139
what the fuck are you talking about. no such interface existed

>>106216143
>Having generic functions in generic headers is the opposite of a mess.
read my post again
or what you misunderstand is that generic headers are the highers level headers, visible to everyone
you postiviely DONT want to pollute them with helper functions
>>106216143
...
how well do you know c?

>>106216163
Yes the point of having them in headers is precisely to make them visible to everybody so anyone can use them. There is nothing inherently wrong with thinking the functions in questions should be exposed on that level. They are super generic and useful in general. Linus apparently doesn't think so which is fine but he could have just said it instead of dishing out pointless insults to someone making an entirely reasonable technical judgement.

>>106216236
>Yes the point of having them in headers is precisely to make them visible to everybody so anyone can use them.
not only thats extremely rude
but you dont do that
you pollute the namespace
you add lines to audit
you break encapsulation (if the term apoplies)

you dont do that
thats malpractice asf
im way less competent than the linux people and i know that
fuck
everyone knew that when i was still at school

>>106216236
>>106216271
cont
also the guy said sorry, 'tis was an error (bw the lines bc he didnt leave himself enough time to do things tip top)
and maybe thats the story
the guy was at his 14th hr that day, didnt think, had to commit bc he had 1 day left
boom. instant fuckup
evendoe... you dont do that
dont put your helpers in top level headers kids
its rude and is malpractice. not bad practice. MALpractice. procedural error

>>106216271
There is nothing "rude" about proposing entirely reasonable code changes. That's ridiculous. If you don't want the change, you just tell the submitter no. People are too quick to defend Linus whenever he has an entirely useless emotional outburst like this. Being blunt is good but telling people they're making the world a worse place to live when they didn't do anything inherently wrong is stupid.

>>106216373
>There is nothing "rude" about proposing entirely reasonable code changes.
no, ofc
but the guy took the decision unilaterally and thats what makes it rude
loonis found out bc he was reading the thing while traveling, as last proof-read. maybe in a plane, smelly, hungry, bloated from the coffee

like i said
its not optimal
but i still find it proportionate

>>106216373
>but the guy took the decision unilaterally and thats what makes it rude
if youre autistic and dont understand this
heres the rule:
dont override others will, even implicitly so
everybody has their sphere of influence so to speak and its a matter of respect to respect others will withing their sphere of influence

>>106215746
He only looks like a tyrant to retards and trannies. He is just trying to keep Linux safe.

>>106216436
yea, the carrot and the stick
a veeeery large stick so that theres a dissuasive dimension to things
the results speak for themselves
idk how many contributors throughout idk, 30 years now?
and the perhaps the most used piece of software for professionnal and industrial applications

but still
his style creates controversy which tranies then use to create a narrative
its suboptimal. one can destroy the soul of an interlocutor so to speak, while remaining calm and collected
even easier so when in a position of power, and veterancy

>>106216412
The guy that submitted the pull request is not "overriding anything". The whole point is that it's a proposed change for others to review. I don't sperg out if someone proposes changes to code that I maintain that I disagree with. I simply tell them no. It's not some stupid fight over who gets to touch whose turf. It's a collaborative open source project.

>>106216502
thats not how "trying to slip something past linus using the fact that its less than 24h to merge" sounds, anon
that sounds like the guy just put his abstraction in a top level header and didnt tell anybody about it
thats judging form the exchange
if you have more info post sauce

>>106216523
That's just bad faith reading from Linus. The merge window was open and maintainers can still send patches. He doesn't have to accept them like this one for example.

>>106216559
>That's just bad faith reading from Linus.
maybe
but like i said i err on the side guys on vacation and he has extra work bc of the merge windows and boom someone does a nono which csgrads know not to do alr
a guy from google.
for my part i think its a case that the guy was overworked and maybe ctrl c v'd into the wrong window
i'd easily see that coming i oftentimes work on several functions at once and then once im finished i copy all the prototypes into the header at once
and rely on the compiler to check that part for me
if the prototype ends in a top level header the compiler's perefecly happy

also yeah looonis' an asshole but hes right
and his approach is right, just the format sucks
as for 2025, the age of the snowflake

>>106216493
The only ones yelling are trannies so I say Linus is the hero and trannies are eternally seething

>>106216607
kek
youre not wrong

>>106216604
>as for 2025, the age of the snowflake
*esp for 2025, the age of the snowflake

>>106216604
and
>prototype
i think its actually a macro
but point still stands
overworked guy
30.000 contexts open at once
types in the """wrong""" one
compiler says nothing
boom. fuckup.

>>106210093
Ackchully, I read up on it and it turns out that using unions for type punning is perfectly fine in C (but not in C++), though things like endianness will depend on the architecture used etc.

>>106216707
ah
thanks for the clarif

>>106207139
>~96.59% Similarity
is this some program or just made-up bullshit?

Had this been me I would tell the Swede cuck to eat shit and die. Given nobody gave him comeuppance the fag might have finally croaked from that. Fuck troonix.

>>106207488
None of these abstractions dword word etc. bullshit in win32 needed to exist actually. It always was obfuscation...

>>106216604
>also yeah looonis' an asshole but hes right
>and his approach is right, just the format sucks
>as for 2025, the age of the snowflake
Yeah I don't think it's the wrong approach, I find the rant funny. Works as an example for everyone and feels like people will learn something from these exchanges.
Actually I feel like I'd learn a ton in a short span of time from a guy like him.

>>106215956
Hmm no. I've been on both sides of this sort of situation, the guy definitely needed a smack. You don't play games with merges before deadlines like that.

>>106215981
this. people just don't understand the entitlement of retards

>>106200568 (OP)
>not using the big F or R on that retard
>explaining the issue with proper grammar and capitalisation
He's being way too nice imo

>>106200568 (OP)
Linus has a valid contention but I think you would have to be downright retarded to assume anything other than "A" being the high word and "B" being the low word. I would make that assumption regardless of endianess.

Linus chimp outs are fucking stupid and he only does it because dunning kruger redditors egg him on after they just googled what bit shifting is and now they're "in the club"

here you go ceniles, fixed that for you