RE: On elisp running native

[arthur miller]

I haven't yet installed WSL, nor do I have any plans. I don't see why should I install entire Linux kernel if I can run applications natively. I don't understand why people are so excited about wsl when they could just run Linux kernel on bare metal? I do it 99% of time myself and I don't miss Windows at slightest.

Mingw and/or Cygwin do just fine for the apps that does not have native win ports already, but most worth-have apps are already written in portable way. Running entire wm and another kernel just to run an app is an overkill on resources in my eyes. 

I would like to see a world where we go towards using less resources instead of just piling platform over platform eating all CPU and ram achievements just because it's easy way out. I think it's madness to run a chat app that requires 1.5 gig or dependency  (slack on blink), or as someone on YT recommends to run a virtual machine on Windows with GNU/Linux distro for the sole purpose of running Emacs. WSL is towards that trend. Of course if Emacs run on WSL, and of course it will,  is not a problem, but if Emacs does not have native port, that will be a problem. 

So I write this just to express that some people value native win32 port and hope native win32 port will not go away or get behind just because it is *easier* to run Emacs on wsl pile of ...

Skickat från min Samsung Galaxy-smartphone.

-------- Originalmeddelande --------

Från: chad <address@hidden>

Datum: 2019-12-29 21:00 (GMT+01:00)

Till: Dmitry Gutov <address@hidden>

Kopia: Eli Zaretskii <address@hidden>, EMACS development team <address@hidden>, Andrea Corallo <address@hidden>

Ämne: Re: On elisp running native

On Sat, Dec 28, 2019 at 5:07 AM Dmitry Gutov <address@hidden> wrote:

Considering WSL is getting more popular and sees some solid development
these days, what are the odds that this is going to be our sole target
on Windows, some years since?

Microsoft is moving the WSL 'platform' away from translating POSIX calls into Windows-native (W32, Win32, I forget the details), and instead implementing a system where the Linux kernel runs in a custom VM (actually, inside a container running inside a VM, if I understand it correctly). The system then provides the necessary software to bridge the two systems from the user POV. The architecture is supposedly very close to what Google (the software part, not the Advertising company part) is doing with ChromeOS and their Crostini project, where the user's choice of distribution runs (kernel and all) alongside the Linux kernel that underlies ChromeOS, and software inside each separate OS is used to bridge (for example) X11 into the Wayland-based display engine, share files between the two, etc. Supposedly, this was the way they found to solve the performance problems with the impedance mismatch between Windows and Linux design issues (especially filesystem operations). 

If this works, then it'll make Windows support easier, since it will be very close to GNU/Linux support.

Hope that helps,

~Chad