It’s 2025, so this shouldn’t happen – but it still does.
I’m doing some work on migrating a failed project of sorts- it isn’t a complete failure- but it didn’t accomplish what the goals were, and right now it’s in a holding stasis.. however, needs to be made stable, and trustworthy before other decisions can be made.
I went to audit a few plugins utilized, and I found some code that I found very suspect- so I checked the project’s GPL code. It matched 100%.
This vendor has hard coded user creation and passwords into a project- and published it with those default passwords.
Needless to say, I removed those functions, and emailed them about it. I’m sure they’ll publish an update- and if they don’t, they’re more than just incompetent.
Although this entire service runs on antiquated hardware utilizing QEMUKVM based services, I haven’t really felt the need to migrate it to something new just to possibly slightly lower the footprint on what could likely be self-contained with 2 gigs of RAM including the OS.
Other than software links that have been dormant nearly 20 years, I’ve got everything set to direct to where it had been back when I was actively developing open source, and all of my outdated knowledgebase is still there for the world to have virtually no use for – like how to use Solaris with CUPS using a print server that was already out of date 20 years ago.
Who knows? It might eventually be worth something.
Also, Devuan still isn’t very good about updates for PHP-FPM. It was still the same parent release, but it was silently failing over on the local socket until I did a complete base update, there really was no specific reason given for it, and I decided to just check for updates before pulling out GDB. My guess is some sub-library it was compiled against either had a namespace collision, or changed semantics since that build.
Then again, I don’t know how many people who still willingly run Apache 2+FPM-CGI go out of their way to avoid systemd.
I’ve used TextPattern for nearly the last 20 years. I’ve written a similar RSS export as I did for my own blog software nearly 20 years ago.
It’s as stagnant as this blog. I don’t really feel the need to replace it, but I have burned down and rebuilt before, using only the core for reinforcement. Is it time?
I won’t bother with WordPress; I’d rather burn it down to a holding page than that mess.
Since I haven’t uploaded any images since 2018, I can’t tell you specifically when the images subsystem broke – but I discovered it wasn’t working today.
No clue what happened that it decided that the images were not in an auto_increment field, but that’d explain why it always wanted to index the image at zero and failed the uploading and thumbnailing.
.but now you get to see my interaction with ChatGPT where it decided to start making BOFH quotes in the style of a Stig introduction.
Despite moving most of my place-of-work’s computing into the cloud, I still rent an ancient computer for less than $1/day to host several of my projects, including this blog.
It’s been pretty stable, but since I have been gently pushing the hardware, the network card has been failing and starting to hang until the driver forces a hard reset.
The NOC techician I spoke with offered to move it to the secondary NIC (network port), and I agreed. He was kind enough to probe that I had an open (unbridged) IP available, so I set the physical NIC to that address, while leaving the virtual bridged NICs alone other than adding the new interface as part of the bond. As the bridged interface was the default gateway, I didn’t have to do anything else.
#ifconfig eth1 x.x.x.x netmask y.y.y.y up
#brctl addif br0 eth1
30 seconds later, it cut across seamlessly when he physically moved the cable. Didn’t even get an alert, it was that fast.
Removed the old NIC from the bridge (just in case), took down my spare IP from the physical NIC, and we’re back and rocking.
#brctl delif br0 eth0
#ifconfig eth1 del x.x.x.x
Had I not kept myself used to managing physical hardware, this would have taken longer.
