Debian still having trouble with merged /usr

Posted Apr 6, 2022 2:33 UTC (Wed) by Paf (subscriber, #91811)
In reply to: Debian still having trouble with merged /usr by milesrout
Parent article: Debian still having trouble with merged /usr

I guess I’d like to hear why the merge is incorrect behavior. I see how PATH can be used instead but I don’t see any reason it’s incorrect.

Debian still having trouble with merged /usr

Posted Apr 6, 2022 9:36 UTC (Wed) by anselm (subscriber, #2796) [Link] (5 responses)

It is “incorrect” mainly because / and /usr have been separate for almost 50 years – mostly because the PDP-11 the original Unix people at Bell Labs were using had disks that were too small to hold the content of both / and /usr on one drive, not because it was an inspired design with obvious and far-reaching architectural advantages –, and secondarily because the suggestion of merging them comes from the person who is also responsible for systemd, which implies that anything he says is automatically suspect.

Debian still having trouble with merged /usr

Posted Apr 6, 2022 10:39 UTC (Wed) by pebolle (guest, #35204) [Link]

Thanks for the link to Rob Landley's rant. I knew a rant like that existed but couldn't find it. Great read, that should be read before reading this article, to fully grasp the absurdity of this situation.

Debian still having trouble with merged /usr

Posted Apr 6, 2022 17:54 UTC (Wed) by jccleaver (guest, #127418) [Link] (3 responses)

>the suggestion of merging them comes from the person who is also responsible for systemd, which implies that anything he says is automatically suspect.

Anything LP says *is* automatically suspect, or should be. The systemd cabal has been the source of more needless sysadmin misery than any other aspect of Linux in the last 12 years, and the fact that it was snuck in (on the RH side) as a nearly-invisible transition just like the swap to upstart had been from EL5->EL6 doesn't help with his trustworthiness. (Nor, indeed, does his history before that, but that's a separate matter.)

systemd is the ur-example of the slippery slope, and this extended pain demonstrates how far that slope goes. This was from 2019 in Fedora, but I wouldn't be surprised if there were still more landmines out there thanks to this... https://bugzilla.redhat.com/show_bug.cgi?id=1691825

Debian still having trouble with merged /usr

Posted Apr 14, 2022 19:42 UTC (Thu) by rahulsundaram (subscriber, #21946) [Link]

> Anything LP says *is* automatically suspect, or should be

This is not rational especially given that Lennart wasn't even involved in this change, at all and it had nothing to do with systemd. Evaluate changes on their own merits.

http://0pointer.de/blog/projects/the-usr-merge

"Primarily I put this together to have a nice place to point all those folks who continue to write me annoyed emails, even though I am actually not even working on all of this..."

>This was from 2019 in Fedora, but I wouldn't be surprised if there were still more landmines out there thanks to this...
> https://bugzilla.redhat.com/show_bug.cgi?id=1691825

https://docs.fedoraproject.org/en-US/packaging-guidelines...

This one is just an example of a maintainer fixing up the package they maintain to follow the guidelines as documented after a change. Hardly new or surprising.

Debian still having trouble with merged /usr

Posted Apr 15, 2022 11:30 UTC (Fri) by HelloWorld (guest, #56129) [Link] (1 responses)

Frankly, it's because of comments like yours among other reasons that I've come to the conclusion that when Lennart proposes something, he's probably right.

Debian still having trouble with merged /usr

Posted Apr 19, 2022 4:35 UTC (Tue) by flussence (guest, #85566) [Link]

LP's not always right: ini is an ugly underspecced format (and different things should look different), and for supposedly cleaning up the legacy swamp, some of systemd's defaults are awful - just look at chrony's absurd attempt to enumerate badness in its service file (I use that example because it's one of the few things that even *try*).

But on the whole his involvement in FOSS has made my life as a sysadmin much less miserable, and I don't even use systemd.