Post B8GFuWwTaK8yZKjUGm by wolf480pl@mstdn.io
(DIR) More posts by wolf480pl@mstdn.io
(DIR) Post #B8G4s6tQMiA7hgwsCG by wolf480pl@mstdn.io
0 likes, 0 repeats
Software should be eternal
(DIR) Post #B8G8EJLl1yXAZVTlRI by wolf480pl@mstdn.io
0 likes, 0 repeats
@nullenvk I think in can in principle.IIRC Voyager 2 is still alive.Also, we can read the source code of early Unix and learn how it worked and how its authors solved different problems (not sure if we can compile it though).I think the main reason solved problems don't stay solved, and working code doesn't stay working, is that people working on other code near it accidentally break things.
(DIR) Post #B8GDrtisENKuRczGZk by Suiseiseki@freesoftwareextremist.com
0 likes, 0 repeats
@wolf480pl @nullenvk Main reason "solved" problems don't stay solved is being the so called "solution" is proprietary software, so using it always ends up in disaster.GNU is eternal.
(DIR) Post #B8GDrttrZW7uzi82hk by wolf480pl@mstdn.io
0 likes, 0 repeats
@Suiseiseki @nullenvk so... Python is proprietary software?
(DIR) Post #B8GFfySUzsDDZ9loJc by Suiseiseki@freesoftwareextremist.com
0 likes, 0 repeats
@wolf480pl @nullenvk Python ever since version v3.14.0a4, is proprietary slopware.But, if you use a free fork of python 3.13 without pointless changes, you won't see breakages once software is ported to 3.13.
(DIR) Post #B8GFfyt5P1TmtcD2ky by wolf480pl@mstdn.io
0 likes, 0 repeats
@Suiseiseki @nullenvk yeah, but how many pointless changes that broke software happened between 3.4 and 3.13?
(DIR) Post #B8GFuWwTaK8yZKjUGm by wolf480pl@mstdn.io
0 likes, 0 repeats
@Suiseiseki @nullenvk Unless we look beyond copyright & licenses and treat backwards-incompatible changes themselves as an obstacle to software freedom.I think there's a strong argument for that, since in practice it does limit an individual's ability to keep using a modified version.
(DIR) Post #B8GGJfFTq357ZRbfCC by Suiseiseki@freesoftwareextremist.com
0 likes, 0 repeats
@wolf480pl @nullenvk Literally every 3.X release.At least with free software, you can make breakages cease to happen.
(DIR) Post #B8GGJfV4u3YgLou7VY by wolf480pl@mstdn.io
0 likes, 0 repeats
@Suiseiseki @nullenvk Yeah but that means proprietary software isn't the sole cause of breakage.Even if I choose to keep an old python version, at some point the GNU C Compiler will turn some warning into an error or sth and the old python will no longer compile (so much for GNU being free from the problem).
(DIR) Post #B8GGh4bdj5xwsM8LgW by Suiseiseki@freesoftwareextremist.com
0 likes, 0 repeats
@wolf480pl @nullenvk The GNU Compiler Collection has flags you can set to not treat warnings as errors and you can just use the same version of GCC with your python version.With decently written software, I personally haven't had any issue getting 20 year old software to compile with the latest GCC.
(DIR) Post #B8GGh4t0gVrPkEGDlA by wolf480pl@mstdn.io
0 likes, 0 repeats
@Suiseiseki @nullenvk I've had issues when defining functions and globals in headers broke - I think it started requiring either explicit extern or an explicit static, whereas previously it wasn't necessary.I think that was around gcc 5.x
(DIR) Post #B8GHeNy7Ndku0rviiW by Suiseiseki@freesoftwareextremist.com
0 likes, 0 repeats
@wolf480pl @nullenvk Yes, unfortunately poorly written software can have wrong headers, which later versions of GCC decided maintaining support for wrong headers was too hard - but that issue isn't forbiddingly hard to solve - you fix the headers.
(DIR) Post #B8GHeOBwYEoYhkOlGa by wolf480pl@mstdn.io
0 likes, 0 repeats
@Suiseiseki this can be said about majority of the breaking changes most of the software ever madeeach of them is individually relatively easy to fix once you understand it, but the fact that they always come as a surprise, and there are many of them, makes them a burden