I am done updating, will stay at current

[Ridrok]

When I see in logs that platformIo is removed, having tested Wayland 10 times without success on various software and reading it will be the only one left with plasma next version I think I am done with updates, I will stick to what I am at after 15 years of Gentoo.

PlatformIO is a thing I can't miss, Damn Wayland is buggy and need to wait another 10 years to be acceptable. So why update when things are working.

I did begin so long ago and lost so many software, but this time it's too much.

Sorry for this rant, but enough is enough in deletion. I prefer to keep an outdated system than loosing yet another vital software or swith to non working windows system that consume also too much cpu for nothing.

[Zucca]

Moved from Installing Gentoo to Gentoo Chat.

[Hu]

dev-embedded/platformio is masked for removal due to:

Code: Select all

# Needs porting to Python 3.13.  Requires old versions of multiple
# dependencies, and reportedly is already broken with its current
# dependencies.  The efforts to revive it do not progress.
# Removal on 2026-02-23.  Bug #944883, #952290, #955765, #968383.

This doesn't seem to be related to Wayland.

You could migrate off Plasma to a system that isn't forcing Wayland so hard.

[wildhorse]

I am not using platformio. But.

https://github.com/platformio/platformio-core/releases

6.1.16
Added support for Python 3.13
...

The current release of platformio is version 6.1.18.
In the gentoo repository, dev-embedded/platformio is stuck at version 6.1.11. And that is why this package gets removed?

I have just built platformio 6.1.18 in less than 10 minutes. I had to rename the ebuild file to 6.1.18 and add python 3.13 to PYTHON_COMPAT in the ebuild file. That was it.

[Ionen]

Note that gentoo has no intention to phase out X support, and there's nothing that will force you to switch to wayland unless your favorite applications happened to drop support upstream. Gnome and KDE are notably planning to drop X sessions support, but other individual applications will likely keep working the same -- if using Gnome or KDE and really do not want to use wayland, may want to switch to something that is more X friendly like xfce4 or not use a DE at all.

Sticking to ancient versions of everything will eventually become a problem when you actually find yourself needing something new, things just won't work together anymore. Like e.g. the web changes a lot and (even if you ignore security) you'll eventually need a newer web browser for websites to work right, and then you'll try to update it and it just won't work right with your ancient system.

wrt platformio, primary issue is that the maintainer for it in Gentoo hasn't been doing anything for it since January 2024 and it kind of fell behind, then this became a problem for the python team and they dropped all these abandoned "stragglers" stuck on python-3.12. It may be possible to get it through some other means (prebuilt, or an overlay -- a local one works too), or maybe someone will be interested in reviving it in ::gentoo later.

[b11n]

It may be possible to get it through some other means (prebuilt, or an overlay -- a local one works too), or maybe someone will be interested in reviving it in ::gentoo later.

I note ::pentoo is one such overlay, if that helps OP any

[Ridrok]

Moving the ebuild to my local repository, renaming to 6.1.18 and editing it to add python 3.13 and deleting entry in /usr/portage/profile/package.mask

Code: Select all

platformio --version
PlatformIO Core, version 6.1.18

For just this missing action, package get ditched from the distribution....

[Zucca]

I contacted the listed maintainer about the update to this platformio ebuild.
Now, at least, the maintainer is aware of the outdated ebuild.
We'll see if it gets updated.

[Ionen]

For just this missing action, package get ditched from the distribution....

It may not seem like much for a single package, but when someone who already takes care of hundreds of packages suddenly has to deal with other packages he's not supposed to be the maintainer of because someone else hasn't been doing their job, it really adds up -- and then, even if fix this, this type of package will need attention again later. To prevent that, it may feel drastic but the simplest is to last-rite the unmaintained package. Best case scenario someone (original maintainer or someone new) will step up and actually maintain that package un-last-riting it, else it'll go away and let us focus on other things.

It's especially a problem with python because the ecosystem and interpreter change things all the time, and the packages constantly need attention and testing plus keeping up given keeping old interpreters support forever would make things unmaintainable. Contrast to e.g. old C applications that rarely break and often do not have volatile dependencies, and then only need trivial fixes maybe once every few years (or decades!), so even if they have no maintainer (may it be gentoo or even upstream) they tend to be left alone.

At least overlays exist, or worst case you always have the option to use a virtualenv and install these packages with pip in your $HOME. So if it's something you really need, shouldn't be entirely out of options even if not ideal.

[asturm]

Moving the ebuild to my local repository, renaming to 6.1.18 and editing it to add python 3.13
[...]
For just this missing action, package get ditched from the distribution....

You would/should hope that the average Gentoo maintainer puts more work into a version bump than "just this missing action". It's not that simple outside of your local overlay.
You were also uniquely motivated as an actual user of this package to make that bump. This is not the case for someone whose own packages are routinely blocked by the inactivity of others, which often comes with potential maintenance backlog, who is not familiar with the package's upstream, or possible build system peculiarities to be on the lookout with version bumps.

Even for KDE Plasma btw, actual phasing out of X11 is one year ahead still.

[moniaqua]

So why update when things are working.

Well, you got a lot of answers I don't really understand because of my lack of knowledge on those topics, but one answer I have: Because things likely won't work any more at one point in future and because there is a security risk with outdated systems.

Gentoo offers a incredibly high number of packages for all flavours; there is no need to stick with desktops that use a as I understand kinda faulty setup of dependencies. I am really happy that there are knowing folks who do the math behind the scenes for me, so that I can use an up-to-date system and say a heartfelt "thank you" to developers and maintainers.

[NeddySeagoon]

Ridrok,

Just for old times sake I have a Gentoo install from April 2003, so its 23 years out of date.
All its security certificates have expired. Most of the crypto algorithms have been retired for being broken or insecure.
In short, it won't work on the internet and its probably full of known security vulnerabilities too.

How long can you live in the past?

[szatox]

How long can you live in the past?

You own example kinda answered it already: as long as you live by yourself.

There are use cases where it is a perfectly reasonable approach, and there are use cases doomed to fail miserably.

[Anon-E-moose]

How long can you live in the past?

Hard to say, I still use sys-fs/eudev-1.10-r2

I have a nas box,

Code: Select all

$ uname -a
Linux nas 5.0.13-zen #4 ZEN SMP PREEMPT Mon Jul 1 12:41:11 CDT 2019 x86_64 Intel(R) Core(TM) i5-3470T CPU @ 2.90GHz GenuineIntel GNU/Linux

that doesn't do much, little web server for local stuff and my local network backups.

Since I move to mini-pcs, my old desktop is used for gaming, early 6 series kernel, vm's for everything from win3.1 to win 11
I don't update it, as there is no real need to.

For my current desktop ...
Mostly it depends on how much work I want to do, balancing keeping old stuff vs running the latest & greatest.

[sam_]

Moving the ebuild to my local repository, renaming to 6.1.18 and editing it to add python 3.13
[...]
For just this missing action, package get ditched from the distribution....

You would/should hope that the average Gentoo maintainer puts more work into a version bump than "just this missing action". It's not that simple outside of your local overlay.
You were also uniquely motivated as an actual user of this package to make that bump. This is not the case for someone whose own packages are routinely blocked by the inactivity of others, which often comes with potential maintenance backlog, who is not familiar with the package's upstream, or possible build system peculiarities to be on the lookout with version bumps.

Even for KDE Plasma btw, actual phasing out of X11 is one year ahead still.

.. which includes e.g. diffing the changes, reviewing the release notes carefully, running the testsuite, adjusting the ebuild for any dependencies, packaging any new dependencies. WORKSFORME is not suitable for committing to the repository. One is free to do that in a local repository, but don't expect us to YOLO things into the main repository.

Anyway, as others have said: put it in a local repo and maintain its dependencies (including possibly old versions it is pinned to, or test if unpinning is safe), or just use a venv w/ python -m venv ~/platformio && . ~/platformio/bin/activate && pip install platformio, or whatever you prefer.

[NeddySeagoon]

szatox,

That 23 year old install I have is a recreation after I found my original minimal CD one day.
It's just a trip down memory lane. See Historical Gentoo on the wiki.
I don't actually use it :)

Finding 23 year old sources to add things is a problem now too.

[wildhorse]

I have a TRS-80 running NewDos/80 and another Z-80 machine running CP/M 2.2.
A lot younger is my VAXstation 3100 running VMS 7.3 and WASD (a wonderful web server).
Sadly, I no longer own the original hardware anymore, but I rebuilt it with FPGAs (this is not a software emulation). It's the spirit that counts, isn't it?