READY TO SERVE YOU AGAIN

The Gopher services here are finally back online after they got cut 
off due to Aussies.space not (yet? But it's been six months in 
total now) supporting running cron jobs or CGI on the Gopher server 
anymore. Updates to Firetext (fires burning in the state of 
Victoria, Australia) and Gopher Banker (international currency 
rates and converter) are now handled via automated SFTP uploads 
from elsewhere, which turned out to be more awkward than I thought 
because SFTP commands can be rather limited. It turns out that 
"lftp" implements many of the recursive operations that make the 
UNIX command line bareable, so I probably should have used that 
instead (for reference, it runs on top of OpenSSH SFTP, so eg. if 
you've set up "aussies" as a "Host" in .ssh/config, you can connect 
with "lftp sftp://aussies" - which is not a fact that's immediately 
obvious from the man page).

Of course SFTP doesn't get me past the lack of CGI when it comes to 
converting currencies with the Gopher Banker. Here the Tilde.Club 
account that I set up during my aborted move from Aussies.space 
turned out to be handy after all. The currency rates are still 
listed here, but actual conversion results are done over at 
Tilde.Club, where the Gophers are still free to roam and execute as 
they please (I didn't realise how terrifying that mental image 
could be until I wrote it).

As before, Firetext is here:
gopher://aussies.space/1/%7efreet/firetext

Gopher Banker is here:
gopher://aussies.space/1/%7efreet/currconv

And now Gopher Banker is here as well:
gopher://tilde.club/1/%7efreet/currconv

I've been thinking of setting up a Gopher CGI world clock with 
ASCII clock faces, but I haven't got around to it. It's a bit of a 
pain spreading things over two servers now - in many ways double 
the things to go wrong. I still have to get around to compiling a 
newer version of PuTTY that can connect to the SSH server at 
Tilde.Club. The internet's in a sad state when a maintained SSH 
client won't even keep working for two to three years after it's 
built from the latest release.

Speaking of obsolete things, I think it's funny reading this 
morning about people starting their "Old Computer Challenge", 
limiting themselves to a single CPU core at minimum clock speed 
("cat /sys/bus/cpu/devices/cpu0/cpufreq/cpuinfo_min_freq" shows 
that minimum would be 480MHz on the Atom CPU of my Atomic Pi 
'internet client') and 512MB RAM. I'm posting as usual from a 
single 120MHz CPU with 80MB RAM, and both doubtless still slower 
than a modern computer pretending to have those specs because the 
bus speed is much slower and the CPU doesn't have all that fancy 
modern stuff like speculative execution (and not even because it's 
disabled due to fear of a passing spectre). Actually I recently set 
up a sub-1GHz single-core x86 PC with 512MB RAM with all my regular 
software for general use, which I might get around to talking about 
properly here eventually (I've been neglecting you interweirdos 
lately for some reason, certainly not for a surplus of social 
interaction elsewhere). As my software choices are already biased 
towards low system resource usage, everything was as fast and 
responsive as I could want. The one exception, as usual, being the 
dreaded Firefox, which I wanted to include for websites like Ebay*, 
which aren't very usable in Dillo. Actually I was surprised that 
Firefox 102 ESR runs quite nicely in 512MB of RAM browsing websites 
with all the Javascript turned off via NoScript, but turn on all 
the JS at a heavyweight website like Ebay and it soon slows to a 
crawl while chewing up swap space. Those are the only websites 
where I'm forced to browse in Firefox in the first place, so 
overall it's a bit pointless with 512MB so I maxed out the system 
with 2GB of RAM and now Firefox can do its bloated job.

* And, since a week or two ago, GitHub. Arrgh! Why did so many 
people have to host their projects there, where I now have to start 
Firefox with Javascript enabled just to browse the source code?! 
Horrible Microsoft, idiot users... hatred, fury, KILL KILL KILL!.. 
alright, calm, calm, I guess I'll just have to go away and 
find/write some sort of browser for the Git protocol, assuming it's 
possible to grab individual files/directories with that.

Oh FFS, the answer is no, it's not possible to 'browse' using Git:
https://stackoverflow.com/a/1179728

Except yes on GitHub because they have an SVN-compatible interface:
https://stackoverflow.com/a/38183088

Except no on GitHub because they're disabling it in Jan. 2024:
https://github.blog/2023-01-20-sunsetting-subversion-support/

Proof again that there's no point trying to code your way out of 
these things, all these services are designed against you. The 
internet is for Chrome, Firefox (maybe), and the latest PCs. 
Compatibility is a thing of the past. All I can say is that if 
you must use GitHub to host a project, then mirror everything 
over Gopher as well!

 - The Free Thinker