Before getting into the article itself,
I’d like to mention that yesterday I saw a very enterprising YouTube video that described how he set up an XFCE Desktop to run on Ubuntu… Which I thought was extraordinarily ambitious. IMO there would be extreme interest if anyone were to put in an effort to do same on openSUSE. IIRC involved installing an X Server I’d never heard of before, then some further tinkering with X display configuration.
Here is some immediate feedback and observation on this YaST Developmebt Sprint 94 article (Thx!)…
Improved Compatibility with WSL
I don’t remember how long it took to install 15.1 WSL on my old system, it did take awhile but haven’t checked on this recently. I did observe a CentOS installed on someone else’ system that took only about a minute. Noticed with pleasure the Microsoft store cleanup, removing the multiple old apps that pretty much all did the same thing.
I took this opportunity to install WSL on a new system (Intel 8050 1.8mhz, 16GB Optane/1TB 7200 HDD), and it installed in 40 seconds. Inspecting YAST modules, I do suspect implements described improvements removing unnecessary modules but I suspect may have been overdone. I will personally review over next few days. Never know for sure if Optane has an effect, it has a 2x effect installing VBox packages but not always apparent when something is running first time.
Noticed that WSL only has OSS and OSS update repos installed. Notable.
I can’t remember my first experience booting 42.3 the first time, so can’t comment on firstboot performance, but is plenty fast after repos have been refreshed the first time.
Interesting about removing systemd completely (running in Windows init).
I seem to remember limited systemd commands working before, particularly checking status but problematic whether can start/stop. Now, nothing.
In fact,
No services reported running even using the “service –status-all” command
And, Partitioner is reading not just no devices but nothing at all (The module just runs with no content). I see at the bottom this is recognized and reason unknown.
Interesting that “journalctl” is still in the system although it doesn’t work because there is nothing to read.
There is an urgent problem with system messages…
Of course the journal and anything systemd related doesn’t exist.
But, there are no system logs in /var/log either (just some yast and other logs).
You allowed the yast2-module to read system logs to remain, but when it’s run, it’s apparently only a pointer that installs the actual module and that in turn discovers there’s nothing to read again.
Bottom line, although I didn’t check out the full extent of what was possible using systemd in WSL before, I don’t remember running into so many issues, the ramifications of ripping out systemd is likely more extensive than realized.
That’s really cool about YaST identifying sysctl conflicts, AFAIK in general sysctl is the preferred way for low level custom modifications so that if there is a conflict(compare what would happen if the change is made directly to /etc/proc/), the modification will fail without borking the system but doesn’t identify what the conflict might be (might be in the system logs)
Much thx to all the hard work that went into these changes…
]]>Right about time! (no) Last week I happily screwed Windows system partition on my wife’s new laptop during installation of openSUSE 15.1
Everything was just like in this article: YaST allowed me to resize the partition, but the data was destroyed. So an attempt to boot to Windows failed miserably.
And about the label, your are right. “Encryption with Volatile Random Key” would be more correct. We will change it. Thanks for the feedback!
]]>swap-with-random-key already implies a volatile key, at least if I am using the “swap” definition of crypttab(5), and I hope yast does too. So the “protected AES” key has no value over a random volume key, except that it is encrypted again(?) somewhere(?) with AES(?) for what benefit? The explanation is severly lacking, the more so the dropdown box.
At the very least, it should be “Encryption with volatile random key” in the text, because the encryption is not volatile (the data is basically still there after a reboot), but the keys are (gone after a reboot).
]]>For openSUSE everything will keep working as expected, using only one disk by default and with the already known Guided Setup (for the time being, at least).
]]>and that is now mapped to
https://github.com/bmwiedemann/openSUSE/tree/master/packages/a/aubio
‘binaries’ was a bit misleading in what I wrote – the source tarball got replaced by a symlink.
Does not make much sense to track those in git because diff and deltas do not work on them.