Re: why X

Ivan Shmakov <ivan@siamics.netremove.invalid> wrote:

On 2024-05-14, Computer Nerd Kev wrote:

XFree86 and later X.org used to include the TinyX servers that
stripped out all but the bare essentials for a small, relatively
self-contained, executable which worked (and still works) with
most X software.  The X.org developers lost interest in it many
years ago and removed it, though some forks exist.

 
       Pointers?

The TinyX forks are by the Puppy Linux and Tiny Core Linux distros:
https://github.com/idunham/tinyxserver
https://github.com/tinycorelinux/tinyx

Puppy also has a Tiny Xlib:
https://github.com/idunham/tinyxlib

The Puppy forks might have been abandoned since their latest
commits are from 2014. Tiny Core still uses their TinyX server as
the default X server on their graphical Live CDs. As well as being
smaller, it starts up much faster than current X.org.

       It's not something I think I've ever used myself; and like
       I've mentioned before, I'm not really interested in software
       that is no longer maintained (unless I decide to maintain it
       myself; but there's only so much productive time I can
       dedicate to the effort, you see.)

Publicly maintaining an abandoned software project is likely to be
much more work than tweaking some old source code just enough to
compile on your own modern system. If you're worried about security
issues then the latter prevents them from being caught by others,
but niche software projects don't attract much attention to those
anyway.

       That said, there's tinyx-wscons [1] in pkgsrc, which means
       "maintained" in my book.  Somehow, the only binary listed is
       for the previous NetBSD version (9.3), and only for x86-64 at
       that, which may indicate it doesn't get as much maintenance
       as it really needs.  Mayhaps someone here might look into
       what's wrong with it?
 
[1] http://cdn.netbsd.org/pub/pkgsrc/current/pkgsrc/x11/tinyx-wscons/

I didn't know of that one (I don't use any of the BSDs), though
it's based on the Tiny Core Linux TinyX. Looking at the commits,
it actually seems much more actively maintained than the others,
with a couple of commits this month. One is "Disable the use of
-flto as it causes issues with NetBSD", so perhaps that refers to
an issue which prevented it building for the current NetBSD when
the Pkgsrc binaries were last compiled?

https://github.com/AceHusky12/tinyx-wscons/commits/master/

After fixing various compiler/architecture compatibility issues,
I've built the XFree86 Xfbdev TinyX server for the Raspberry Pi
and it runs fine on a RPi Zero.  Interestingly seemingly every one
of the X libraries built from the XFree86 sources was significantly
smaller than those built from recent X.org on the Pi as well.

 
       Do you have these experiments recorded anywhere?  It might be
       interesting to those dealing with minimalistic or retro systems.

No, I did ask about packaging them for Tiny Core Linux's PiCore
Raspberry Pi distro because their TinyX server doesn't build for
ARM, but although it triggered some discussion nobody said they
were actually interested in using it. That surprised me because
X.org is slow on eg. the RPi Zero, so they're what I would use, but
maybe I'm the only audience looking for it.

Documenting my observations in detail might also be work for the
benefit of an imagined audience. But maybe one day.

       My impression was that there was some work on improving i18n
       support in X.org, and that's both a worthy effort IMO /and/
       tends to lead to bigger binaries.  No idea if that's what's
       happened here, though.

I guess if that's so, it would be something I'd want to disable
since I'm not interested in non-English characters (if I don't
know what it means, then I don't care whether it's rendered). TinyX
will display unicode characters from non-English languages, though
I don't think I've seen an emoji work there except in Firefox
(which I suspect renders such characters itself as graphics). But I
find massive emoji-filled font packages obscene, so that's probably
the first place I've broken them.

--
__          __
#_ < |\| |< _#