Ariel Millennium Thornton @arielmt@computerfairi.es

I'm getting desklaunch and deskmenu mixed up. Note to self: See about patching desklaunch to never open ~/.desklaunchrc for writing. Why does it even need to, when its comments & readme explicitly tell users to throw a text editor at the rc?

I remember the age when doing this, adding between a few hundred and a few thousand kilobytes (KB), was considered reckless desktop bloat.

How far we've come, when it's not considered bloat unless it's measured in gigabytes, six orders of magnitude larger.

Well, I suppose I could keep deskmenu for when I'm too lazy to hold down the scroll wheel, and I could also use dmenu for when I'm too lazy even to take a hand off the keyboard.

Wait a minute. If I'm going to use wmx as my window manager, I don't really need a separate desktop menu application, no matter how lightweight.

I can put shell scripts and symlinks in ~/.wmx/ with subdirectories dictating submenu arrangement (and `ln -s .. <submenu-name>` to have a navigation aid), then middle-click-hold to access them in organized menu form! Whatever's there that has +x set gets menu-fied.

All right, practical desktop minimalism is a go!

Oh. It does have a homepage: https://www.oroborus.org/

Interestingly enough, both the source code downloads and the changelog links are from Debian's repo.

Also, because the site hasn't been updated in almost seven years, some of the links have gone stale.

One edit of the Makefile, and it works. And it's as destructively buggy with its rc file as I remember it being.

Has anyone made a Mastodon instance for shouting where, like at dolphin.town you can only post the letter E, you can only post the letter A?

Crap, another package I want turns out to be another piece of abandonware I have to swipe from Debian: desklaunch. The Makefile is a fixer-upper, but there's only a single C file to compile, and the manpage was written by a Debian maintainer, not the author.

Oops. xcal's today pane is the upper-left square. The other two squares are xclock and xload.

I don't expect it matters to too many other people, but I'm excited and happy. Anyway, a screenie (croppped) of xcal running and partly customized on my PC: https://computerfairi.es/media/W40bNZdWr2Bhudlybqw

(My window manager is x11-wm/wmx.)

It built, it installed, and it works exactly on my new FreeBSD 11.1-RELEASE #Unix PC like it did on my old Debian Jessie #Linux PC!

Have the fruits of my labor: http://files.thornton2.com/packages/unix/xcal-4.1-patched-for-freebsd.tar.gz

Inside is the unpatched xcal 4.1 source tarball from Debian Sid (as of last week), my patch file, and my notes for applying the patch.

(Packages screenies & desc: https://screenshots.debian.net/package/xcal )

(Also, no, because I set DESTDIR to my staging directory name, I didn't run `make install` as root.)

So I installed it into a staging directory and ran it, and it ran exactly as expected. But attempting to install it into the staging directory revealed another Makefile showstopper: an extra space between $(DESTDIR) and $(XAPPLOADDIR) in the install:: XCal.help make target. So close yet so far.

Hey! It compiles! Ship it!

Me taking a break from tech support:
*sees call for tech support in local TL*
*sees helpful tech support in local TL*
*sees successful resolution no matter the causes*
*does NOT see any blaming*
*basically, sees computer fairies being computer fairies in local TL*

Have I mentioned yet that I love this place? :3

#Linux ransomware in the wild: https://forums.gentoo.org/viewtopic-t-1060828.html

Linux-based malware not only exists but has existed for longer than most Linux users realize.

Also, not running anything as root that doesn't need root helps with defense in depth, but especially don't run as root anything *designed* to download & run completely unvetted code, like Web browsers do.

Sadly, spoiler: Flash Player was suspected but never confirmed as the ransomware entry vector.

Wow, it's been ages since I used diff and patch. Anyway, I'm making progress.

And finding that GCC is apparently happy with non-void C functions using `return;` without a value, while clang (rightly, IMO) throws an error.