How's everyone doing?

[Amity88]

I'm kinda revisiting after a big break ( predictably after a fresh Gentoo install ). How's everyone holding up? Neddy Seadragon, pjp, Hu especially. I also wonder about all the guys from OTW.
_________________

[asturm]

Welcome back!

[flysideways]

[Amity88]

[pjp]

Gentoo is still here, as are BSDs.

Most of the OTW regulars left when it was closed in Dec 2020. Every once in a while a couple of them post. OTW2 eventually closed as well. Something may or may not have happened to it's maintainer, or maybe it merely expired.

I have to upgrade a router soon. It's long past due, but has been showing increasing problems, so it can't wait much longer. At least I hope the problems will be resolved by an upgrade. A new computer would be nice, but that's unlikely any time soon. I do have some old, spare laptops I've started messing with, but not much.

More recently I've been working with C again, and the last few days I've been spending time working to better understand make and Makefile. I finally got what I wanted working, but it relies on too much hard coding of file names. Hopefully I can translate that the project I've been working on and have better organization of the project.

In October I'll probably switch to profile 23. Instead of upgrading, I may do a reinstall. It will depend on how big the backup would be. /etc, world, and /home I think should do it. Unfortunately I won't get far enough along with ZFS experiments before I have to upgrade, so I'll probably have to try for the next physical upgrade for that. I have no idea how much a pain it will be to maintain, so that's the uncertainty.

That's about it for the hobby stuff.
_________________
Quis separabit? Quo animo?

[Hu]

I am still here. I still stick mostly to technical threads.

Profile 23 was pretty easy for me.

[Zucca]

[Gatsby]

[pjp]

Hah! I didn't expect a profile 23 discussion. I'm mainly delaying it because there are 3. A binhost / primary workstation / server, a chroot, and my laptop that receives the binaries from the chroot.

Some time before mid-2022 I was seeing some "internal compiler errors" on gcc and clang (and/or something llvm related). After whatever needed to be updated during regular updates finished, the first re-emerge attempt for the failed package would succeed. I'd usually see the failure the next day, and a re-emerge would succeed. On a small number of occasions when the re-emerge would fail, I recall that it happened soon after the first failure. That is, the re-emerge was done within minutes of the failure, not overnight. And it was only those close attempts to re-emerge that failed.

The big packages failing and then succeeding makes me think memory issues. The "close together" failures make me think heat. And no, I haven't yet run memtest on it. I haven't experienced any other odd problems, so if it is memory related, it is rarely encountered (except for that mid-2022 period... I don't recall the last time it happened.

And while I won't stare at the output of compiling, I will want to keep a close eye on whether or not there are any issues, given that the only recommended path is an emptytree rebuild. I typically updated the binhost, then chroot, and finally the laptop.

But since the upgrade procedure also negates any gain that might have come from --buildpkg binaries, it seems marginally different from a reinstall. I'll probably do an upgrade, but I still need to prepare for recovery.

I also thought I might replace the hardware, but that's unlikely before 17.1 is deprecated. And I'd sooner go back to Windows than try and maintain my own profile.

<soapbox-rant>But for the lack of an easier alternative / path, being spammed by the entire profile 23 news update every time portage is engaged, I'd be using something else by now.</soapbox-rant>
_________________
Quis separabit? Quo animo?

[Hu]

You could modify your local Portage tree to un-deprecate the 17.1 profile. The deprecation itself is just so you aren't surprised when it goes away one day, so removing that warning shouldn't hurt Portage's operation. Such a modification will be undone on your next sync, but it's a small enough change that you could redo it.

I chose to take my chances on the 23 migration by not doing an emerge --emptytree. I have not observed any problems from that decision, though I feel compelled to note that this is a deviation from the standard instructions, and should not be attempted without some consideration for the risks versus rewards of such a deviation. Lurking readers should not take my success as an assurance that they too can get away with skipping it.

[pjp]

I get why it's there, but it's exceedingly obnoxious and doesn't really do anything constructive. If I were the type to wager based on past performance, I'd be willing to "guarantee" SOMEONE is going to ask about it because they have a system that hasn't been updated in months or years and they didn't know.

I had hoped reading it would silence it, but no luck. I sometimes leave news items "unread" as a reminder.

I haven't seriously considered not doing the emptytree method, mainly because I don't want the pieces. I have considered a slower approach though where I only update with Gentoo binaries (I suspect most won't be compatible) and then generate an emptytree list of packages that I go through in chunks. But I'm not really sure how "safe" that would work. I presume using nodeps wouldn't be a good solution as it seems that could compound potential problems.

I'll see what I have to do to un-deprecate it
_________________
Quis separabit? Quo animo?

[Gatsby]

[mrbassie]

[GDH-gentoo]

[pjp]

[flysideways]

While on the topic of the profile 23 migration, did you all remain split-usr, or have you also migrated to merged-usr?

[Gatsby]

[Banana]

I've already done the merge a long time ago and I haven't noticed any problems so far.
_________________
Forum Guidelines

PFL - Portage file list - find which package a file or command belongs to.
My delta-labs.org snippets do expire

[GDH-gentoo]

[ian.au]

[Hu]

split-usr versus merged-usr is not about whether /usr is a separate filesystem from /. That distinction is about separate /usr, which the systemd project declared broken years ago. More recently, the systemd project has been pushing merged-usr, where /bin, /sbin, etc. are symlinks into /usr. To further confuse the issue, the systemd wiki page cited in the Gentoo news ( https://freedesktop.org/wiki/Software/systemd/separate-usr-is-broken/ ) now talks about their work pushing merged-usr in Fedora, which I don't think it did when last I read it.

You could argue that merged-usr is more future-proof, since systemd now requires merged-usr, and I have already seen various packages that break on split-usr because of bad assumptions by their authors. It seems plausible that such breaks will continue to appear, and at some point, the maintainers may tire of cleaning up the resulting mess, and declare split-usr unsupported even on non-systemd systems. I have not seen a timeline for such a thing to occur, though.

[pjp]

It almost seems as if Linux (via systemd) is trying to abandon the embedded market. Having /bin and /sbin boot the core system and later mount external /usr for additional capabilities seems extremely useful.
_________________
Quis separabit? Quo animo?

[Zucca]

... and I live on the edge, having separate-etc on few machines. :P
Quite a lot is possible if one crafts own initramfs scripts. ;)

Without initramfs, it's getting harder. Although I'm not sure if one could use EFI partition as pre-real-root environment instead of initramfs... Still it would mostly be the same: pre-init environment does the initial loading and mounting of essential things, then do switch_root.
_________________
..: Zucca :..

[NeddySeagoon]

I'm still here but not as active as I was.
I have a health issue to deal with that's taking my time and energy first.
I expect abnormality to return before spring 2025. Meanwhile I look in on to forums and IRC from time to time.
_________________
Regards,

NeddySeagoon

Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail.

[Gatsby]