Mastodon Feed: Posts

ireneista@irenes.space ("Irenes (many)") wrote:

that trap was annoying to write, due to the constraints we had to write it under, and we were proud that we made it work, but ehhhhhhhh our philosophy on that is we already had the experience, and it lives forever in the version control history, so we should let it go

ireneista@irenes.space ("Irenes (many)") wrote:

and the trap that runs at the beginning of packalign, too. it's always lovely when we can take out the earlier, messier stuff like that.

ireneista@irenes.space ("Irenes (many)") wrote:

yeah this is looking good, we can probably remove the hardcoded metadata entry type for alignment padding

glyph ("Glyph") wrote:

My hopes are not high but I guess it's worth asking at least https://discussions.apple.com/thread/256357711

ireneista@irenes.space ("Irenes (many)") wrote:

hmmm we need to be careful about how it interacts with other words around it, hmm

Boosted by jonny@neuromatch.social ("jonny (nonvenomous)"):
klairelockheart@mastodon.scot ("Klaire Lockheart") wrote:

My dudes, this painting is going to be on display in Denver at the William Havu Gallery! See him for yourself September 18th-November 7th.

#brodalisque #williamhavugallery #minibike #dangermouse #oiloncanvas

ireneista@irenes.space ("Irenes (many)") wrote:

before we had the general-purpose programmability features in the magic-comment syntax, we had it in our head that the hex transform would take responsibility for knowing what all the output-producing words are, semantically

but now that we have that, it makes more sense to have them do it themselves. let's see if we can rework packalign that way...

jscalzi@threads.net ("John Scalzi") wrote:

My fantasy football team won its first game of the season, so I have that going for me, which is nice

ireneista@irenes.space ("Irenes (many)") wrote:

hmm, we'd like to change how alignment padding is displayed, to make it look similar to how suffix comments work... should be easy. doing that next.

ireneista@irenes.space ("Irenes (many)") wrote:

and then at some point after everything else is ready, we also need to add support for showing the high-level flow-control words, which slide code around in memory after it's generated. but there's a clear and obvious way to do that, so we'd like to get the other stuff done first

ireneista@irenes.space ("Irenes (many)") wrote:

that's looking pretty good as far as raw structure goes, it definitely makes us feel like we have enough of a framework to decide how the last few things should look

ireneista@irenes.space ("Irenes (many)") wrote:

hmm, so we used our newfound ability to run magic comments in immediate mode during the label transform to fix up the indentation and section breaks for the overall file

glyph ("Glyph") wrote:

it's part of the general trend of making UIs more "accommodating" and "intuitive" with machine learning. I don't want machine-learning interfaces, I want people-learning interfaces. give me a mostly-intuitive syntax with _rules_ that people can _learn_, that we can get good at and adapt to and build skills with using, not shifting sand that suddenly behaves differently one day when I say "daily at 3PM" instead of "every day at 3PM" because some imperceptible model weight shifted somewhere

glyph ("Glyph") wrote:

it's really grating how "AI" is always making AI less useful. like for example predictive text. predictive text now barfs out entire irrelevant distracting "jab me right in my neurodivergence" full sentences that are extremely annoying. so I have to turn it off. but there's just the one toggle; I either get NO predictive text, or I get LLM barf on my keyboard. so I have to sacrifice whole-word predictive completions entirely now.

Boosted by soatok@furry.engineer ("Soatok Dreamseeker"):
sophieschmieg@infosec.exchange ("Sophie Schmieg") wrote:

@soatok @49016 "yes fine, we get a SIGSEG, is this actually an RCE" was not what I was expecting to read about PGP

Boosted by baldur@toot.cafe ("Baldur Bjarnason"):
uglyreykjavik.bsky.social@bsky.brid.gy ("Ugly Reykjavik") wrote:

The plan to turn Toppstöðin into a recreational centre is still on despite yesterday's fire. The repairs will probably take a while since the building is not in good condition and it contains asbestos.

#Iceland #Reykjavik #photography #streetphotography #abandoned #decay #door #windows #history

Boosted by fribbledom ("muesli"):
joho@mastodon.online ("Joaquim Homrighausen") wrote:

If you ever wondered what happened to Andrew Milner (author of the RemoteAccess BBS software, and many other things), here's his take on it!

https://www.reddit.com/r/bbs/comments/1wgmb8a/the%5Fstory%5Fof%5Fremoteaccess%5Fand%5Fwhat%5Fcame%5Fafter%5Fby/

#andrewmilner #remoteaccess #bbs #entrepreneur #reddit

soatok@furry.engineer ("Soatok Dreamseeker") wrote:

Seeing Werner Koch's lame response on oss-sec is absolutely confirming my suspicions about the GnuPG team

jonny@neuromatch.social ("jonny (nonvenomous)") wrote:

When our instance goes down, just get in the zulip chat and @ me. I think we have had less than like 6 hours of unplanned downtime in almost 4 years

jonny@neuromatch.social ("jonny (nonvenomous)") wrote:

So there are several things that could be wrong: the blacksky PDS could be overloaded/down, but I don't think so since I can't see anything on bluesky at all, even when logged out, and I can't login to another account hosted on their PDS. the feed provider could be down, but that's also unlikely because the same error appears twice, (in two different ways?) for a third party and a first-party feed. The appview could be down, but I also can't access blacksky through the web. So hmm, what went down? And how would I know? And how would I prevent that? It seems like the "feed error" message can't be trusted because it looks like it mixes together errors that should happen in different places, and I have to intuit what an outage means using a variety of indirect observations, since bluesky's status page isn't showing anything. Problem is intermittent, going down and coming back.

These are genuine questions, I want to know what is still a global failure mode for the network, I would have figured an independent PDS and independent AppView would be independent enough that transient failures of the relay might pass by, but I suppose not?

dysfun@treehouse.systems ("gaytabase") wrote:

vimacs: combines all the awkwardness of vim with all the awkwardness of emacs while delivering none of the features of either.

jonny@neuromatch.social ("jonny (nonvenomous)") wrote:

Huh, um, I don't think its supposed to do that.

ireneista@irenes.space ("Irenes (many)") wrote:

https://code.irenes.space/evocation/commit/?id=4137177c016928a6d255bbef762f0d220945083c for the curious

ireneista@irenes.space ("Irenes (many)") wrote:

as we had hoped, this really did make it much clearer to us what it means for one transform to modify another.

ireneista@irenes.space ("Irenes (many)") wrote:

yesssssss okay. we got the hex transform patching the label transform to make the latter handle magic comments. nice.

dysfun@treehouse.systems ("gaytabase") wrote:

we regret that you formalised it.

Boosted by cwebber@social.coop ("Christine Lemmer-Webber"):
vv@solarpunk.moe ("princess bubblegum 💫") wrote:

Patience is such a cool concept for an LFO and it needs some more backers to be funded!

I love the idea of a locking eurorack module

Like, look at this cool backplate

https://www.kickstarter.com/projects/olivia-artz-modular/patience-for-eurorack

#Eurorack #ModularSynthesizer #modular #lfo #kickstarter #synths #module

dysfun@treehouse.systems ("gaytabase") wrote:

on the fruit tier list, we're all clear that apple is at the bottom, right? possibly except for conference pear?

dysfun@treehouse.systems ("gaytabase") wrote:

i am begging corporations to at least grasp the irony of complaining about scraper bots while pursuing AI bullshit.

Boosted by glyph ("Glyph"):
evacide@hachyderm.io wrote:

Just remember that when you opt into allowing Apple to use your customer data to train AI, that means humans seeing and hearing details about your life that maybe you're not so comfortable sharing when you think of it that way.

https://appleinsider.com/articles/26/09/14/apple-has-altered-course-on-using-customer-data-to-train-its-ai