4 ms·

This will end the moment Linus has to review a giant kernel patch and realizes he has no idea what he is reading. As awful as C is, there is real value in keep

by newaccount2021 5y ago

This will end the moment Linus has to review a giant kernel patch and realizes he has no idea what he is reading. As awful as C is, there is real value in keeping kernel development limited to one language. If people really want a Rust kernel, try Redox

PoignardAzur 5y ago
Rust really isn't that hard to read. The syntax is pretty close to C, and there are very few invisible rules. You need to learn language concepts like '?', derives, traits, dyn traits, etc, but they're all fairly straightforward. Rust's big barrier to entry is learning to write code that passes the borrow checker, but that's not a concern when doing code reviews.
- khuey 5y ago
  > Rust's big barrier to entry is learning to write code that passes the borrow checker, but that's not a concern when doing code reviews. Rust code is actually easier to review in that regard because (unless the code contains unsafe blocks) the reviewer can skip thinking about most aspects of memory management.
panick21_ 5y ago
Time moves on, its of value to be conservative with new languages but it also dogmatic to say it should always only be one. Yes having one has value, but having more then one also can have value. Rust has reached a level of acceptance and adds huge value that it is worth it for the future.
nicoburns 5y ago
What makes you think Linus won't learn Rust? I believe he's said that accepting Rust into the kernel will involve key maintainers learning Rust, and that so far they seem cautiously open to that.
- newaccount2021 5y ago
  he may think he knows Rust until he gets his first whiff of real Rust line noise. to be certain, line noise in C is probably something he sees every day as well, but he also has decades of experience to help him filter it it seems inevitable that the same counterarguments he used to decline C++ will eventually be trotted out again. it is disappointing to see kernel devs be taken in by the Rust Hype
- kzrdude 5y ago
  I fear there will be pushback once some of them learn and don't like it
sidkshatriya 5y ago
Funny! :) However, in reality Linus wont necessarily review the patch line by line. Mostly the reviews will happen on mailing lists. Then if Linus' lieutenants that handle the specific subsystem like what they see, they will probably give a general thumbs up for the change. Finally Linus will take an executive call. He may offer some technical suggestions (as he is technical) and send it back or accept the patch or conditionally accept the patch. The point is that the finer details will have already been dealt with by the time it lands on Linus' desktop. But given that this is going to affect a lot of things for years to come he is probably going to spend a lot more time on it than other features. Also I think Linux may be interested in having Rust in the kernel. See here https://lkml.org/lkml/2021/4/14/1099 https://lkml.org/lkml/2021/4/14/1099 . He says "on the whole I don't hate it". It's clear to see that there is a lot of momentum behind Rust. Allowing Rust in the Linux kernel will help bring a lot more talent and energy into the kernel community (other than the inherent merits of Rust which I think are easy to see).
pavon 5y ago
Linus has already learned Rust and has been actively participating in reviewing previous rounds of patches. For example[1]. His general consensus so far: "I don't hate it"[2]. [1]https://lkml.org/lkml/2021/4/14/1131 https://lkml.org/lkml/2021/4/14/1131 [2]https://lkml.org/lkml/2021/4/14/1099 https://lkml.org/lkml/2021/4/14/1099

PoignardAzur 5y ago

Rust really isn't that hard to read. The syntax is pretty close to C, and there are very few invisible rules. You need to learn language concepts like '?', derives, traits, dyn traits, etc, but they're all fairly straightforward. Rust's big barrier to entry is learning to write code that passes the borrow checker, but that's not a concern when doing code reviews.

khuey 5y ago
> Rust's big barrier to entry is learning to write code that passes the borrow checker, but that's not a concern when doing code reviews. Rust code is actually easier to review in that regard because (unless the code contains unsafe blocks) the reviewer can skip thinking about most aspects of memory management.

panick21_ 5y ago

Time moves on, its of value to be conservative with new languages but it also dogmatic to say it should always only be one. Yes having one has value, but having more then one also can have value. Rust has reached a level of acceptance and adds huge value that it is worth it for the future.

nicoburns 5y ago

What makes you think Linus won't learn Rust? I believe he's said that accepting Rust into the kernel will involve key maintainers learning Rust, and that so far they seem cautiously open to that.

newaccount2021 5y ago
he may think he knows Rust until he gets his first whiff of real Rust line noise. to be certain, line noise in C is probably something he sees every day as well, but he also has decades of experience to help him filter it it seems inevitable that the same counterarguments he used to decline C++ will eventually be trotted out again. it is disappointing to see kernel devs be taken in by the Rust Hype
kzrdude 5y ago
I fear there will be pushback once some of them learn and don't like it

sidkshatriya 5y ago

Funny! :) However, in reality Linus wont necessarily review the patch line by line. Mostly the reviews will happen on mailing lists. Then if Linus' lieutenants that handle the specific subsystem like what they see, they will probably give a general thumbs up for the change. Finally Linus will take an executive call. He may offer some technical suggestions (as he is technical) and send it back or accept the patch or conditionally accept the patch. The point is that the finer details will have already been dealt with by the time it lands on Linus' desktop. But given that this is going to affect a lot of things for years to come he is probably going to spend a lot more time on it than other features. Also I think Linux may be interested in having Rust in the kernel. See here https://lkml.org/lkml/2021/4/14/1099 https://lkml.org/lkml/2021/4/14/1099 . He says "on the whole I don't hate it". It's clear to see that there is a lot of momentum behind Rust. Allowing Rust in the Linux kernel will help bring a lot more talent and energy into the kernel community (other than the inherent merits of Rust which I think are easy to see).

pavon 5y ago

Linus has already learned Rust and has been actively participating in reviewing previous rounds of patches. For example[1]. His general consensus so far: "I don't hate it"[2]. [1]https://lkml.org/lkml/2021/4/14/1131 https://lkml.org/lkml/2021/4/14/1131 [2]https://lkml.org/lkml/2021/4/14/1099 https://lkml.org/lkml/2021/4/14/1099