#Mastodon is decentralised - not just as a protocol, but in DNS terms as well, as this chart shows. Only 2% of Mastodon accounts are on instances that have .com domains (vs ~50% of regular websites), and only 33% on the "legacy" gTLDs (.com, .net & .org). "New" gTLDs are by far the most common TLD for Mastodon server names. Source: https://instances.social/list/old
We might find the tomb of Cleopatra and Mark Anthony!
So the null pointer deref seems to be coming from virtualbox which makes sense, but is no more explicable. What is PHP doing to cause it?!
Looking at a diff of hex dumps of the .phar files created under each scenario, apart from some trailing binary, the only change seems to be some bytes being swapped from 0x33 to 0x12
OK, this is weird, any #php peeps out there who can help me?
I am building a .phar file using "phar pack" to put files into the archive, including dependencies from a separate git repository.
If I use a symbolic link to the separate git repository, the phar works fine. But if I create a git submodule, the phar builds, but causes a kernel NULL pointer dereference when it runs!
What's going on there?
nerdy code stuff
I'm working on some code that will be distributed and installed by a partner, so it needs documentation, build tools, and so on. Example configuration files (YAML and unit files) are in the documentation file, which is written in Markdown.
In my Makefile, which builds a .deb, I do this to extract the sample config files:
pandoc -t markdown -t json < README.md | jq -r '.blocks | select (.t=="CodeBlock") | select(.c=="yaml").c' > /etc/config.yaml.dist
PSA: In yesterday's dpriv WG session Google said that their public DNS will implement draft-ietf-dprive-unilateral-probing real soon now. So if you see weird traffic on port 853 on your nameservers, it's that. #IETF115