C++ Logo

std-discussion

Advanced search

Re: [std-proposals] We should use existing scientific research in building C++ modules

From: Jason McKesson <jmckesson_at_[hidden]>
Date: Thu, 18 Mar 2021 10:52:38 -0400
On Thu, Mar 18, 2021 at 10:25 AM MichaƂ Policht via Std-Discussion
<std-discussion_at_[hidden]> wrote:
>
> Hi,
>
> I am moving this to std-discussion, because I'll be bit off-topic. I am
> also concerned about C++ build systems. I've read the papers referenced
> by Askar and made a remark that all C++ build systems rely on "foreign"
> languages (either a custom DSL is invented or something like Python,
> Haskell, Java....). Wouldn't it make sense to rely heavier on C++ and
> seek for a single C++-based stack for a build toolchain?

Um, why? What does it matter what language your build tools are made
in? You're not using your build tools in the code you're building, so
it's kind of irrelevant.

It's not like the Python compiler/executable is written in Python
(well, it can be, but CPython isn't), nor is the Java compiler written
in Java. Using multiple languages for the build environment of your
language is hardly unusual or confined to C++.

Lastly, it is not the job of the C++ standard to demand that the
language's build system be written in a specific language. So it's
rather out-of-bounds here.

Received on 2021-03-18 09:52:51