[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Texmacs-dev] Fwd: guile-devel Digest, Vol 212, Issue 4
From: |
Sebastian Miele |
Subject: |
Re: [Texmacs-dev] Fwd: guile-devel Digest, Vol 212, Issue 4 |
Date: |
Wed, 29 Jul 2020 19:18:50 +0200 |
User-agent: |
mu4e 1.4.12; emacs 27.0.91 |
Massimiliano Gubinelli <m.gubinelli@gmail.com> writes:
> I just noted on the guile mailing list that having Guile-2 on Windows
> seems a bit tricky. So we should take this into account when we will
> shift to it (if we do it).
The complete thread starts here:
https://lists.gnu.org/archive/html/guile-user/2020-07/msg00110.html
The gist:
The Guile maintainers generally assume some things that are
provided/fulfilled by POSIX, but not by Windows (per se). Especially
64bit and threads are a (huge) problem.
However, that does not apply when running with a POSIX emulation layer
(Cygwin, MSYS2) or with Windows Subsystem for Linux (WSL2). Cygwin has
the latest Guile (3.0.4) on x86-64, probably with threads / without any
limitations. And seemingly all previous Guiles. I did not find a web
package list for, e.g., WSL2 Ubuntu, but I strongly suspect the
situation is similar. MSYS2 does not have Guile 3, yet. But it is
unlikely that it will not.
As far as I understand www.texmacs.org/tmweb/help/faq.en.html, Cygwin
may already be required to run TeXmacs on Windows. Is that the case?
Possible downsides of depending on Cygwin, MSYS, and/or WSL2:
* Maybe Cygwin introduces a performance penalty? (Probably no big deal.)
Somewhere I read that WSL2 in some cases is even faster than native
Linux, whatever that exactly means. (Even more likely no big deal.)
* There may be slight nuisances for a Windows user: Using emulated POSIX
paths. Having to press the button in Windows 10 that activates WSL2
and installs a WSL2 distribution. Or having to install Cygwin or
MSYS2. (But that should all be a breeze and clearly explainable on one
page.)
* Qt somehow should be used in such a way that it is not required to run
an X Server. I.e. it probably should use the view on the filesystem as
Cygwin/MSYS2/WSL2 does, but nonetheless use the usual Qt-on-Windows
graphics and input methods. It is not unlikely that that already is
solved in the TeXmacs source, at least for Qt4. Is that the case?