Re:The values of Emacs, the Neovim revolution, and the VSCod...
by monsieurbanana on Sunday January 17, 2021 @06:34PM
(#25816408)
What I often hear is more about Emacs on windows having performance issues in some areas, magit comes to mind. Did you run into any issues?
Re:The values of Emacs, the Neovim revolution, and the VSCod...
by clankyclanker on Sunday January 17, 2021 @07:01PM
(#25816625)
There is one case I know of where Emacs (via Cygwin) has performance issues on Windows: listing directories in remote volumes. Per StackOverflow [0], dired is slow because `df` is slow on NTFS. So, you can eliminate df to speed up dired by a lot.
`(setq directory-free-space-program nil)`
0: https://stackoverflow.com/questions/26128693/emacs-dired-slo...
Re:The values of Emacs, the Neovim revolution, and the VSCod...
by BeetleB on Sunday January 17, 2021 @07:15PM
(#25816724)
Main issue is magit. Currently the repository I work on is small enough that I put up with it. But I can see this being a real deal breaker for many.
There are certain Emacs features that may require some UNIX tools like find or grep. Often I've found ways to configure it to use ripgrep and fd instead (both are trivial installs in Windows, and generally faster than grep and find anyway).
Re:The values of Emacs, the Neovim revolution, and the VSCod...
by omaranto on Sunday January 17, 2021 @07:14PM
(#25816715)
I think git is slow on Windows, and magit doesn't make the problem better or worse.
Re:The values of Emacs, the Neovim revolution, and the VSCod...
by BeetleB on Sunday January 17, 2021 @07:20PM
(#25816767)
Nope - git is mostly fast on Windows. It's magit that's the problem - or rather how magit does things. It launches several git processes to do basic things like status or commit, and that is fundamentally slow in Windows.
See https://magit.vc/manual/magit/Microsoft-Windows-Performance....
Re:The values of Emacs, the Neovim revolution, and the VSCod...
by omaranto on Monday January 18, 2021 @03:47PM
(#25826447)
So you are saying git is slow on Windows, but only because starting any process is slow on Windows. Interesting. I've heard there is work in progress to have magit use libgit2 instead of running the git executable. That should help.
When I was younger, I wanted to work in places that had a passion for these types of things.
Older me realized that was a bad idea. It inevitably seeps into things unrelated to doing the actual work (even performance reviews via 360 feedback). People are paid to do a job, and I've found that whether they love computing or not is mostly orthogonal. Sure, there are some that are wiz's because of it, but they're usually neutralized by those who become ideological and inflexible because of it.
Tangentially related: I'm a Linux guy - it's been my primary OS for almost 20 years now. However, at my current company, I prefer roles that don't involve Linux. The reason is they typically will not give me my own machine - I'll merely get an account on some server, with a poor window manager, and missing a lot of the nicer tools I use on my home machine, and no superuser privileges. In the last job I had, their provided Emacs was a little too old for some of the features I used, and I had to compile my own Emacs from scratch (along with tens of other dependent packages). This was time wasted for me and the company. At least on my Windows machine at work, I have Admin privileges and can make things nicer.
Paradoxically, it is because I'm a "hard core" Linux user that I prefer Windows for work. I've optimized my Linux workflow, and if my work cannot provide me my preferred tools/window managers, I might as well just use Windows.
I don't even use WSL. I use xonsh, which works fairly well on Windows. I have multiple desktops on Windows. And most importantly, I live in Emacs which for the most part works just as it would in Linux :-)
(And no, using/installing Emacs on Windows is not a pain, despite what you often hear - the last time I did it some months ago, it was literally "Download and unzip into directory and you're good to go.")