Linus Torvalds, the legend 🔥
-
Are we talking about Linus Torvalds, the person who does not follow GNU coding standards at all?
There are all kinds of hard coded limits in Linux, for instance. I had to completely change the design of something after running into this with overlay fs.
And lately he blew his top at other people who wanted to submit some code for big-endian architectures, because HE, Linus Torvalds, has let people commit code to the kernel that is little endian-only.
Oh, he’s a legend, alright
@chemoelectric @jbc @itsfoss I wouldn't say that. Torvalds has indicated that he does not like to be involved in the naming matter himself, but does consider usage of GNU/Linux or Linux/GNU to be more professional when talking about the combination.
-
@chemoelectric @jbc @itsfoss I wouldn't say that. Torvalds has indicated that he does not like to be involved in the naming matter himself, but does consider usage of GNU/Linux or Linux/GNU to be more professional when talking about the combination.
I was talking about the coding standards, which say that there should be no hard coded limits.
Linux is not a GNU project and so is not subject to the coding standards, but that does not make it hunky-dory that Linux does not follow the standards. It is the de facto kernel for GNU systems and so is used by people like myself who prefer GNU coding standards over hard limits.
And I see no excuse for letting people just plaster in code endian or alignment sensitive.
-
I was talking about the coding standards, which say that there should be no hard coded limits.
Linux is not a GNU project and so is not subject to the coding standards, but that does not make it hunky-dory that Linux does not follow the standards. It is the de facto kernel for GNU systems and so is used by people like myself who prefer GNU coding standards over hard limits.
And I see no excuse for letting people just plaster in code endian or alignment sensitive.
-
@chemoelectric @jbc @itsfoss I see, I would say that the popularity of Linux does speak for it, but indeed I'm not technical enough to judge such aspects. I think from the perspective of Torvalds, at this point hardware design and support should tailor Linux instead of the other way around. This is different from how it used to be and where nowadays other kernels might become more interesting.
Given the recent HURD improvements, possibly it will become an interesting contender down the line.
-
@chemoelectric @jbc @itsfoss I see, I would say that the popularity of Linux does speak for it, but indeed I'm not technical enough to judge such aspects. I think from the perspective of Torvalds, at this point hardware design and support should tailor Linux instead of the other way around. This is different from how it used to be and where nowadays other kernels might become more interesting.
Given the recent HURD improvements, possibly it will become an interesting contender down the line.
@jschwart Torvalds is one of the people who has driven women out of the field.
I do not believe in popularity as a determiner of merit. Linux sucks and should not determine hardware. It is x86-centric precisely because its code quality sucks.
So also Rust sucks and people should not waste time on it. They should instead upgrade their code to C23. If they actually cared about code security they would have switched to Ada long ago. They are switching to Rust only because it is a fad.
-
@jschwart Torvalds is one of the people who has driven women out of the field.
I do not believe in popularity as a determiner of merit. Linux sucks and should not determine hardware. It is x86-centric precisely because its code quality sucks.
So also Rust sucks and people should not waste time on it. They should instead upgrade their code to C23. If they actually cared about code security they would have switched to Ada long ago. They are switching to Rust only because it is a fad.
@jschwart Anyway, these folk want to support PowerPC as a legacy system, not as a competitor with AMD64. PowerPC is a 32-bit CPU, after all. I used to program for it on AIX and Solaris. It is big-endian and requires strict alignment. It is a RISC design and is very fast. To program for it is educational.
People who learn only on Intel/AMD architectures are being cheated. They don’t even learn how to program well on THOSE architectures. It is ADVANTAGEOUS to adhere to alignment rules on them.
-
@jschwart Anyway, these folk want to support PowerPC as a legacy system, not as a competitor with AMD64. PowerPC is a 32-bit CPU, after all. I used to program for it on AIX and Solaris. It is big-endian and requires strict alignment. It is a RISC design and is very fast. To program for it is educational.
People who learn only on Intel/AMD architectures are being cheated. They don’t even learn how to program well on THOSE architectures. It is ADVANTAGEOUS to adhere to alignment rules on them.
@chemoelectric how do you see ARM in that light? Regarding Solaris/SPARC I do have an old Ultra 5 here. Would you suggest I look at learning it's assembly language? (That would be for one future day though, I'm afraid not soon...)
You seem to have a lot of experience and it's interesting to see both our family names like that
-
@chemoelectric how do you see ARM in that light? Regarding Solaris/SPARC I do have an old Ultra 5 here. Would you suggest I look at learning it's assembly language? (That would be for one future day though, I'm afraid not soon...)
You seem to have a lot of experience and it's interesting to see both our family names like that
@jschwart I don’t know ARM nor PowerPC nor SPARC. (Maybe I am wrong about what was on that Sun. But the AIX box was a PowerPC and it was easily our fastest machine.) I don’t know the assembly languages.
The only assembly language I ever knew WELL was Z80/8080. I could do 8086 decently at one time, too.
20-something years ago I worked in porting AS/400 Command Language code (which tied together RPG programs) to POSIX machines. The data had to be binary compatible across architectures...
-
@jschwart I don’t know ARM nor PowerPC nor SPARC. (Maybe I am wrong about what was on that Sun. But the AIX box was a PowerPC and it was easily our fastest machine.) I don’t know the assembly languages.
The only assembly language I ever knew WELL was Z80/8080. I could do 8086 decently at one time, too.
20-something years ago I worked in porting AS/400 Command Language code (which tied together RPG programs) to POSIX machines. The data had to be binary compatible across architectures...
@jschwart I do not remember whether we stored the data big endian or little endian. Probably big endian, because I think HP was the ‘Ur-Host’ and it was big endian. Windows and (after I ported to it) GNU/Linux on x86 were probably our only little endian platforms.
But even then the alignment rules were different. Once I found a bug and had to stick some extra padding in a struct and tell people that on a certain architecture the old files were binary incompatible.
-
@jschwart I do not remember whether we stored the data big endian or little endian. Probably big endian, because I think HP was the ‘Ur-Host’ and it was big endian. Windows and (after I ported to it) GNU/Linux on x86 were probably our only little endian platforms.
But even then the alignment rules were different. Once I found a bug and had to stick some extra padding in a struct and tell people that on a certain architecture the old files were binary incompatible.
@jschwart I have been disabled for over 20 years now but have written things like hash function packages and made sure they followed alignment rules and endianness. Though I can test only on AMD64.
As for my family name, my great grandfather was that rare thing, a Jewish baseball player, LOL: https://www.rollbamaroll.com/2018/8/4/17648800/the-curious-case-of-schwartz-alabama-crimson-tide-mystery-1906-baseball-coach
-
Linus Torvalds, the legend
@itsfoss "to put it in terms you'd understand, vibe coding is a lot like when you give your wife gonorrhea..."
-
@jschwart I have been disabled for over 20 years now but have written things like hash function packages and made sure they followed alignment rules and endianness. Though I can test only on AMD64.
As for my family name, my great grandfather was that rare thing, a Jewish baseball player, LOL: https://www.rollbamaroll.com/2018/8/4/17648800/the-curious-case-of-schwartz-alabama-crimson-tide-mystery-1906-baseball-coach
@jschwart But on my father’s maternal side I go instead to people like minor figures of the Revolutionary War.
My mother’s side is New York City and straight out of Seinfeld, however.
-
@jschwart But on my father’s maternal side I go instead to people like minor figures of the Revolutionary War.
My mother’s side is New York City and straight out of Seinfeld, however.
@jschwart Yeah, I don’t know about the assembly languages.
I run into people who apparently have CS degrees but do not know what in high level languages you have to be aware of endianness and alignment. That these usually are not abstracted out. Certainly they are not in C, which is barely a high level language.
I myself have but 2 credit hours in CS, for the engineering Fortran course that I didn’t really need and where the county college cheated and just stuck us on a Unix PDP-11 with Ratfor
-
@itsfoss Somewhat ironic that the screenshot has telltale signs of being genai...
Mind to provide link the source as should be the norm for good journalism? -
Linus Torvalds, the legend
Linus talks to a pedophile!
Who got STDs from a young girl
then shutdown Epstein's XBox accountI wish this was a joke, but it's real.
-
Linus Torvalds, the legend
@itsfoss Torvalds is very brave for standing next to a pedophile.
-
Linus Torvalds, the legend
