Skip to main content

I should really change my profile picture. It was meant as a ridiculous form of a bathroom mirror selfie (hence the shower curtain), but it comes across as some pretentious photographer. Pretentious, moi? And also not a photographer, not more than everybody is.

Content warning: Pinned post, about following from multiple addresses

This entry was edited (1 month ago)

Content warning: pinned opinion about nude scenes in films

This entry was edited (1 month ago)

Content warning: crossposting policy

This entry was edited (6 months ago)

Gidi reshared this.

This entry was edited (6 months ago)

Gidi Kroon reshared this.

Meanwhile I dealt with reducing my AWS bill last weekend and also the server is still running.
This entry was edited (1 month ago)
Ok, two weekends ago. This is the elastic file storage part of the bill. That's NFS, still storing the same data. Other services I still need to reduce.
Bar chart showing seven days, with on 15 June a slight increase but on 16 and 17 June a drop to relative zero. Not shown is the next week where on a new scale it clearly drops further.

What is the magic word?


No, try again.


Yes, thank you.

This entry was edited (1 month ago)

It's a good habit to, before you click the shutdown button, scan the menu to see whether it also has a disconnect button.
This entry was edited (1 month ago)

SPOF: very good for performance.

My sites are so much faster now that I introduced a single point of failure. Can highly recommend.

This entry was edited (1 month ago)

It also gets rid of $30 on my Amazon bill. The storage I was using may be only 6 dollarcents (using intelligent tiering) and accessing the data normally only around one dollar (a month), but I had an extra $30 cost because I was using an inefficient setting and was too lazy too figure out how to access the data properly.

(Something was doing more than 100 metadata calls per second to NFS, likely directory listings or date checks, and I still don't know what. Only figured out it's something in php, even after I set opcache to only check every 15 minutes. This is just above the allowed throughput baseline, so I set 'elastic throughput' which scales throughput automatically, but you pay for that too.)

Feeding the bots.
This entry was edited (1 month ago)

"How many visitors does your site draw?"

"Some". "Oh, do you mean with or without the bots?"


"Then none"

There's this one bot that is now at page 8 of a tag on my profile, while I think I only ever made one post with that tag. Another is trying to find whether there is any update to the page that is an overview of posts I made in 2021.

I'm at that point in the rabbit hole where I am disappointed in some widely used software.
This entry was edited (1 month ago)

I mean, how does mod_cache, or varnish for that matter, not take into account the Accept or Content-type headers when caching something?

I mean "I want the html at this address", results in "here have this json because someone before you asked for the json". They are like "the server should have sent an Vary header". Yeah, but what if it doesn't? Shouldn't the cache, as the thing that is taking over serving content, serve the right content according to standard specifications?

And why doesn't the Location directive in Apache's config match with a path like you'd expect? So if you were to add a Vary header to the path /friendica, you can't because it's subject to rewrite rules. You would have to match /index.php?pagename=friendica. But that doesn't work for me either, probably due to the query string. I can only add the header to the whole site, like this:

    <IfModule mod_headers.c>
        <Location "/index.php">
            Header merge Vary "Accept"


        CacheQuickHandler off

and then finally mod_cache serves the correct content types. With a much larger cache.

BTW, the same problem existed on profiles and posts, that are served machine-to-machine as json and to a browser as html, but sometimes from the same address.

The internet has a clear answer to my question how one would horizontally scale Wordpress and Wordpress-like applications: don't.
This entry was edited (1 month ago)

I found that switching between other distributions can better be done in a separate partition. One extra thing to do, though, is that some OS's like Ubuntu only autodetect other OS's like Windows and show these as options in the bootloader screen, but they don't do the same for other Ubuntu-like OS's like Kubuntu. So to get the options to boot into either Ubuntu or Kubuntu (or others) if you have both installed, you need to add the following line to /etc/default/grub and follow the instructions to run update-grup:


EDIT: forgot that I may need to run the following too afterwards, from the OS to be considered the main OS:

sudo dpkg-reconfigure grub-efi-amd64-signed
This entry was edited (1 month ago)

This entry was edited (1 month ago)

So ActiveState has changed the licensing on their products from free to paid, including for old versions, including those you had downloaded and installed before they blogged about that change back in 2021. Other than that blog post they didn't inform anybody of that change. I'm not even sure you can change licensing conditions after someone has installed your product and accepted the license conditions.
This entry was edited (1 month ago)

Messed up media storage on the dev node so much all files have been inaccessible. That's why it's a dev node I guess. Now I get to practise restoring the database, which is also a good thing.
This entry was edited (1 month ago)
That restore went smoothly. It even immediately re-imported the rss feed posts that had happened between the backup and now.

In trying to find out why a new version of a thing uses so much more resources, I've come across an interesting new data point. When I turn it off, the warning that it's off doesn't appear and the things it was supposed to be doing, are still happening.
This entry was edited (1 month ago)

So this site documenting the AT protocol has a frequently asked questions section, but no way to ask questions. I'm suspicious.

That FAQ explains why they couldn't just use ActivityPub, by listing a couple of things missing in it. These things happen to be the exact things the Zot protocol does have, while Zot has existing implementations that do interoperate with ActivityPub. They couldn't have looked long.
#AT #ATProtocol
This entry was edited (1 month ago)

Reminded of the time I wrote a small useful utility that nicely ended with an exit(), which later got incorporated into a web application because it was so useful and which took me an embarrassing amount of time to debug why the server always 'crashed' after clicking my new button.
This entry was edited (1 month ago)

When changing output resolution on a chromecast, selecting ok in the menu does nothing no matter how many times you do it, until you also hit cancel, then it sticks. Just now found out that just doing cancel twice, no ok, also succeeds in changing the resolution.
This entry was edited (1 month ago)
I have to keep doing this kind of thing because my tv and chromecast have a dodgy handshake.

When I switch them on, the tv and chromecast will negotiate resolution, hdr settings and colour encoding and agree upon which to use based on the best available modes. Following this agreement my tv will do one thing and the chromecast another.

Google: please send us all your second factors for our hungry cloud
Me: no
This entry was edited (1 month ago)
You can easily back up and replicate on your other devices.

Thwarted by a pop-under, yet again.
This entry was edited (1 month ago)

I keep finding out that a usb-c plug fits rather well into a usb-a slot. It doesn't connect in any way, but it fits. Enough to fool me into thinking I've plugged it into the correct slot.
This entry was edited (1 month ago)
if it fits it sits 😜

People: these algorithms analyse you and are so smart
Instagram: have a look at popular reels from your country
Me: have I ever, in my decade of life on your app, looked for local content?
This entry was edited (1 month ago)

From a little ago, but several people shared this today, a blog at the FTC (US government thing to regulate businesses and in this case advertising claims), reiterating that just claiming a product has AI is not allowed without making sure of some things. I especially appreciate the risk/responsibility thing: AI tends to be vague and poorly understood, but as a business you are responsible for your product so you better understand how it works, how it gets to its output and how you can test it.
This entry was edited (1 month ago)