Comment by LinXitoW
10 months ago
The issue is that Linus put the Rust developers in an impossible position: On the one hand he approved Rust in the kernel, but then never ever ever has the balls to enforce that decision.
Then, the fanatical C developers openly sabotage and work against all the Rust developers efforts. So, the last option for the Rust developers is to take it to social media. Otherwise, the C developers get away with creating a self fulfilling prophecy: Sabotage all Rust efforts, then claim the Rust experiment failed.
Linus didn't seem to ever have the time to actually take a stance, except of course on the social media issue. Fully ignoring all context. It's the equivalent of a school principal suspending the bullied victim for finally snapping and punching their bully.
The thread didn't really have drama before marcan stirred the pot. There was a disagreement, but the individuals pushing for the merge were not attempting to escalate, only try to find a path forward in a way that might make both parties happy with the compromise. The drama and social media outrage arguably did nothing to help, and as far as I can tell, simply makes for good entertainment for onlookers who like to gossip. While it would be nice to have Linus help out here with a clear resolution after escalation, it's clear to me that the behavior marcan displayed is the higher priority problem to address.
I think this isn't the right take. the "disagreement" was a kernel maintainer saying "Rust in the Linux kernel is a mistake and I will do everything in my power to sabotage Rust adoption" (as feedback on version 8 of a patch). The fact that open undermining of Rust for Linux receives no pushback from Linus or anyone else with power in the Kernel is shocking.
100% this. Yes, Hector went nuclear, but he begged Linus to step in and provide leadership (either merge or reject) and instead Linus ignored the whole technical issue with regards to rust being totally blocked.
Even now with Hector out of the picture, there’s still no suitable path forward for rust in Linux. No wonder why people are giving up (exactly what the blockers want).
6 replies →
The rust project is not stalled. There is pushback from maintainers, but as trust is established, things will hopefully change and it'll get easier. Escalating erodes trust and simply will make the process take longer. Anything that fuels us vs them narratives are simply damaging. Everyone needs to focus on the shared objectives of the overall Linux project. Ignoring the maintainers and pretending their opinions are wrong or don't matter won't help. I'm not sure a top down directive is necessarily the right way to establish a healthy space for rust in linux.
I agree. Everyone here seems to be criticising Marcan for not being professional, but it’s very difficult to remain professional when the people you’re working with gloat in public that they intend to completely sabotage your work product despite it being given explicit support by the CEO. Why are you the only one criticising the coworkers? I feel like I’m taking crazy pills reading this thread.
Difficulties notwithstanding, Marcan’s drama is excessive. He has a history of it.
His fellow R4L partners chewed him out for jumping in and spoiling their work. They even quietly but publicly disaffiliated R4L from him.
Because this thread is about Marcan's behavior, not others'. I think it's perfectly fair to claim that everyone behaved badly in this situation, but the person I replied to was insinuating that Marcan didn't do anything wrong. That is not true, and why I highlighted his behavior.
> So, the last option for the Rust developers is to take it to social media.
Social media is an amplifier of interpersonal problems, not a place to seek for a resolution for them - unless your intended "resolution" is to beat down the other side, the people you have to work alongside by necessity, via potshots from random strangers who hardly ever bother to inform themselves fully of the situation. That is never going to be a true resolution, and I think Linus, for all his faults, recognizes that and that's why he draws the line there.
The C maintainer in question had no power to stop the code from being merged, it wasn't in his directory. He was tagged as a courtesy in case he wanted to do a drive-by review since the code was wrapping his subsystem. The Rust code being reviewed wasn't written by marcan, and the other Rust developers called him out for taking the argument to social media when the code was likely going to be merged anyway (see https://lore.kernel.org/rust-for-linux/Z6OzgBYZNJPr_ZD1@phen... and https://lore.kernel.org/rust-for-linux/CAPM=9tzPR9wd=3Wbjnp-... ).