Microsoft to Explore Using Rust

It&#x27;s interesting that someone in Microsoft is talking about Rust, but the article is very flawed. For example,<p>&gt; C#, a programming language developed by Microsoft, also features some memory access improvements but is not as advanced as Rust. [..] Besides [Rust] being superior to C# in regards to better memory protections<p>That&#x27;s false, isn&#x27;t it? C# is a memory safe language, period. It relies on GC for that.<p>In fact C# has arguably better memory safety than Rust because you can do things safely in C# that you would be forced to use &#x27;unsafe&#x27; in Rust for. (Granted, then you have overhead from GC, but that&#x27;s not what the author is talking about.)<p>&gt; Rust is also more popular with developers these days and might be easier to recruit for.<p>The author has misinterpreted what &quot;most loved&quot; means in the quoted survey: it means that <i>among Rust developers</i>, it gets a very high rating. That says nothing about how big that group is, nor how popular it is in the general population of developers.<p>The Rust community is growing but still very small - it&#x27;s an emerging language. Almost everyone that uses it <i>decided</i> to use it because they like it. (That doesn&#x27;t diminish the accomplishment - there are other emerging languages that are not as loved by their users.)

&quot;70% of all Microsoft patches are for memory-related bugs&quot;.<p>This says a lot. I&#x27;m sure Microsoft has very smart and competent engineers, yet they still get all these bugs. It&#x27;s because manual memory management is really, really hard.<p>I&#x27;m baffled to keep hearing some people saying that manual memory management is okay when the evidence is this huge.<p>Perhaps for very simple programs is okay to do manual memory management, but for complex ones like most of them nowadays, it seems close impossible for humans to write safe programs.

The historic data related to memory access is really damning, 70% of CVE&#x27;s are memory access related, and that hasn&#x27;t changed significantly over the years charted, 2004 to 2018.<p>What&#x27;s annoying in this article is that it doesn&#x27;t come right out and say that MS is considering offering a full Windows SDK with a Rust API. The closest I see the article come to mentioning this is: &quot;Exploring the use of a memory-safe language such as Rust would provide an alternative to creating safer Microsoft apps.<p>But Thomas also argues that third-party developers should also be looking into memory-safe languages as well. He cites reasons such as the time and effort developers put into learning how to debug the memory-related security flaws that crop up in their C++ apps.&quot;<p>It would be interesting to know if a Rust native API is something being actively worked on at MS.<p>BTW, MS does ship Rust in a product today, the VSCode editor&#x2F;ide uses ripgrep for search, which the article fails to mention as well.

I wonder if this applies to Windows at all (and in which parts). Windows Longhorn (using this name for the two years of effectively scrapped work on Vista) was written in c# but abandoned. The best reason I’ve heard is that the GC wasn’t suitable for an OS, especially in edge cases. For example, the C++ code I wrote at Microsoft had to gracefully handle an out of memory situation after any alloc (including setting the STL to not throw exceptions). An out of memory exception in C# can be catastrophic if the handler had to JIT in low level OS code.

&quot;Besides being superior to C# in regards to better memory protections, Rust is also more popular with developers these days and might be easier to recruit for&quot;<p>Citation needed. I&#x27;m quite sure the pool of C#&#x2F;.NET devs right now is quite a bit larger than the pool of Rust devs, to be honest, if anything for the sheer amount of headstart that C# has got over Rust.

If MS is serious about this, I’d like to see them make Rust a first class option for Windows kernel development. There are a lot of drivers that really should be “rewritten in Rust”.

This article is a reblog of Microsoft&#x27;s post: <a href="https:&#x2F;&#x2F;msrc-blog.microsoft.com&#x2F;2019&#x2F;07&#x2F;16&#x2F;a-proactive-approach-to-more-secure-code&#x2F;" rel="nofollow">https:&#x2F;&#x2F;msrc-blog.microsoft.com&#x2F;2019&#x2F;07&#x2F;16&#x2F;a-proactive-appro...</a>

If this goes anything like their &quot;exploring integrating Python into Excel&quot;, in 18 months we&#x27;ll have heard no more about it and it will be presumed abandoned.

This is interesting coming from MS! I&#x27;ve used a lot of VM&#x2F;GC languages and IMO C# is the best of them all. All the things people complain the most about, like type erasure in Java, or lack of generics in Go, are done right in C#.<p>I used it for years and don&#x27;t have any major complaints.<p>Hearing that the company with the &quot;best&quot; language (all subjective) is interested in Rust is exciting.

Get ready for R#.

I thought Microsoft has already been using Rust and contributing to its ecosystem indirectly via its developers such as Actix and Actix-web, although it might just be that the author happens to work for MS while building it in spare time.

Interesting, I would have expected doubling down on c# and .net maybe some compile to native. Obviously not for low level, but the stats are not clear on if it’s general subsystems and applications or low level either

Anyone see an article that says bad things about rust? these articles are shitty.