Re:Towards Linux Kernel Memory Safety
by vilhelm_s on Thursday November 16, 2017 @12:01AM
(#15710355)
Firefox is 17 million!
[ Parent
| on HN ]
Re:Towards Linux Kernel Memory Safety
by akmittal on Thursday November 16, 2017 @04:49AM
(#15711420)
They rewrote 160,000 lines and it took them more than a year
> It replaces approximately 160,000 lines of C++ with 85,000 lines of Rust.
https://blog.rust-lang.org/2017/11/14/Fearless-Concurrency-I...
Re:Towards Linux Kernel Memory Safety
by mundanevoice on Thursday November 16, 2017 @01:27AM
(#15710618)
The whole Firefox codebase has not been ported to Rust and will not be anytime soon. So your argument doesn't hold true.
Re:Towards Linux Kernel Memory Safety
by daveFNbuck on Thursday November 16, 2017 @01:58AM
(#15710711)
Isn't that a point in the rewrite's favor? You don't have to port everything to gain a lot of the benefits.
Re:Towards Linux Kernel Memory Safety
by mundanevoice on Thursday November 16, 2017 @02:02AM
(#15710727)
> You don't have to port everything to gain a lot of the benefits.
I agree with this. But just saying Firefox has X millions of lines of code doesn't say anything about the possibility for Linux to be ported to Rust.
Re:Towards Linux Kernel Memory Safety
by daveFNbuck on Thursday November 16, 2017 @02:46AM
(#15710852)
It does say that the 20 million lines statistic isn't sufficient to dismiss the idea. That's not saying much, but I think it's a valid response to someone pointing out that the Linux kernel is 20 million lines.
Re:Towards Linux Kernel Memory Safety
by craftyguy on Thursday November 16, 2017 @02:20AM
(#15710770)
you're comparing apples and rutabagas. code line count is a hilariously inaccurate way to measure the complexity required for a portable OS kernel vs a userspace application.
Re:Towards Linux Kernel Memory Safety
by mundanevoice on Thursday November 16, 2017 @03:26AM
(#15711027)
> code line count is a hilariously inaccurate way to measure the complexity required for a portable OS kernel vs a userspace application
I never said that. My whole premise is that you can't say anything like this by just saying code count on two very different projects. Maybe you didn't read my comments properly.
What benefits, precisely, though? New bugs? http://insert-script.blogspot.de/2017/11/firefox-settings-co...
Re:Towards Linux Kernel Memory Safety
by irishsultan on Thursday November 16, 2017 @06:28AM
(#15711852)
What's the relation with Rust?
I don't see it mentioned in the post, it affects Firefox before version 57, but not 57 (not that there wasn't any Rust code in Firefox before 57) and it seems to be a bug in DOMParser, which as far as I know is not written in Rust.