gbl08ma

@gbl08ma

I'm a software developer; preferences (descending): Go, Svelte, TypeScript, C#. I built @pds.labeler.tny.im, jungletv.live, underlx.com, and more. Disillusioned with Android; open web FTW. Slightly into game modding. Sometimes I make music. I use Arch BTW.

464 Following305 Followers1,014 Posts

YouTube Bluesky

Is this your profile?Claim it and gbl08ma.dirty.social is yours.Claim profile

gbl08ma@gbl08ma· 20d

I dislike how my list of ideas for blog posts is pretty much all about the impact of LLMs on my profession... ...except for that last one, but it'd need so much visual support it's probably best as a video, and at that point it gets into "requires too much effort" territory.

Blog post ideas

- Excuses to do nothing
- how "AI AI AI" made it worse
- Possibly more generally appealing variation on the above: how "agentic coding" has made it less valuable to share small pieces of software (anything simple enough can be custom built at the distance of a vibecoding session)
- Tangentially, the same "agentic coding" has made Linux more amenable to a wider population, as anything that "can be fixed with a script" is now reachable by more people (with potentially less than desirable side effects, but still)
- It also means that FOSS will suffer (even more than already known) because that "initial simple project" that sparked a bigger one (as it evolved to respond to the requirements of different users) will either not be published at all, or it will get lost in a "sea of vibecoded slop"
- i.e. people won't look for a pre-existing solution to their small problems, they'll just prompt something anew and people will be using a thousand hyper-specialized private solutions rather than contributing to a series of fewer common solutions that solve more problems for more people
- at the same time: this sounds like the Unix philosophy at work... except the "one thing doers" never get published, and they might not "be doing it well"
- Is it still worth it to have a projects portfolio in the age of AI? Won't it just be understood as "here's what I prompted Claude to do?" (And if yes, might there be some value in that?)
- https://www.reddit.com/r/cscareerquestions/comments/1uy5m4u/every_junior_portfolio_i_screen_now_is_incredible/
- Note: this post is basically already written under this page https://gbl08ma.com/works/software/
- CP2077 and the out-of-bounds parkour/exploration meta-game it evokes in me
- Robbie Legume briefly touched on this in one of his videos, how certain areas almost seem designed for this type of activity

gbl08ma@gbl08ma· 24d

Here's mine

4.0K ./did:plc:f32dgb4rl5wgbzx5ia23ercw
364K ./did:plc:y5gazb6lrsk3j4riiro62zjn
1.1M ./did:plc:fkidx24kdwj72crxp7a35pb2
1.9M ./did:plc:cs6o7md2gg3zsde7cjtfn3km
2.4M ./did:plc:uyauirpjzk6le4ygqzatcwnq
7.1M ./did:plc:pco5dul5nqdqdiardqfvyqbv
54M ./did:plc:iq6cty3nsnruczd24vvvnovt
161M ./did:plc:3dr23xsom4xvpxsuwuhj7b7q
171M ./tempt
794M ./did:plc:x5yvecq27wcdvf5qxdi4ru25
1.2G .

gbl08ma@gbl08ma· May 9

BTW you should know that pulling via the web UI is broken after publishing standard.site records

Pull failed: git pull failed: error: cannot pull with rebase: You have unstaged changes. error: Please commit or stash them.

b7a08352aae3:/lichen-root/gbl08ma.com# git status
On branch main
Your branch is up to date with 'origin/main'.

Changes not staged for commit:
(use "git add <file>..." to update what will be committed)
(use "git restore <file>..." to discard changes in working directory)
modified: .standard-site-state.json

no changes added to commit (use "git add" and/or "git commit -a")

gbl08ma@gbl08ma· May 9

Sadly it fails on my particular website when it tries to process a specific post. I don't know enough Rust but to me it sounds you're trying to trim a string in a naive way that Rust is not happy with?

lichen_server | thread 'tokio-rt-worker' (7) panicked at crates/lichen-server/src/publish/standard_site/markdown.rs:152:15:
lichen_server | byte index 10000 is not a char boundary; it is inside '“' (bytes 9998..10001) of `After yesterday’s popular post Windows 10 is unfinished (/posts/windows-10-is-unfinished/), where I bashed said OS, today, I’m going to praise Windows 10 (where possible). This is so we can keep with the opinion diversity people are now accustomed to s`[...]
lichen_server | note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace

gbl08ma@gbl08ma· May 5

You're right, I think it's a logged error but it still pushes. In the meantime it has developed a new error when I try to push via the web UI. When I try it on the terminal the message is more descriptive and actionable... but shouldn't this be something lichen handles itself?

Push failed: git push failed: git-remote-pds: pushing to gbl08ma.com/next.gbl08ma.com git-remote-pds: error: uploadBlob failed: invalid_token: "exp" claim timestamp check failed (401 Unauthorized) error: failed to push some refs to 'pds://gbl08ma.com/next.gbl08ma.com'

329406be938e:/lichen-root/next.gbl08ma.com# git push
2026-05-05T16:14:35.217212Z ERROR atproto_identity::resolve: error=error-atproto-identity-resolve-6 Invalid HTTP resolution response: expected DID format
2026-05-05T16:14:35.641820Z ERROR atproto_identity::resolve: error=error-atproto-identity-resolve-6 Invalid HTTP resolution response: expected DID format
git-remote-pds: error: not logged in for 'gbl08ma.com' — run: git-remote-pds auth login --handle gbl08ma.com
error: failed to push some refs to 'pds://gbl08ma.com/next.gbl08ma.com'

gbl08ma@gbl08ma· May 5

This is the result of resolving the relevant DNS record within the lichen-server container. I'm at a loss...

329406be938e:/lichen-root/next.gbl08ma.com# dig -t TXT _atproto.gbl08ma.com

; <<>> DiG 9.18.47 <<>> -t TXT _atproto.gbl08ma.com
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 47412
;; flags: qr rd ra; QUERY: 1, ANSWER: 1, AUTHORITY: 0, ADDITIONAL: 1

;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags:; udp: 1232
;; QUESTION SECTION:
;_atproto.gbl08ma.com. IN TXT

;; ANSWER SECTION:
_atproto.gbl08ma.com. 6871 IN TXT "did=did:plc:3dr23xsom4xvpxsuwuhj7b7q"

;; Query time: 1 msec
;; SERVER: 127.0.0.11#53(127.0.0.11) (UDP)
;; WHEN: Tue May 05 10:11:25 UTC 2026
;; MSG SIZE rcvd: 98

gbl08ma@gbl08ma· May 5

Hmm, weird... when I run this test on my local machine (not the container where I have lichen-server running), it **passes**. I don't have more time to investigate now but I think I'll have to rule out DNS issues on the machine...

$/// Verifies that handle resolution works for gbl08ma.com. /// /// Uses gbl08ma.com which resolves via DNS TXT. Run with: /// cargo test -p lichen-server --features atproto resolve_handle_gbl -- --ignored #[tokio::test] #[ignore] async fn resolve_handle_gbl() { let did = super::resolve_handle("gbl08ma.com", None).await; assert!(did.is_ok(), "handle resolution failed: {:?}", did.err()); assert!( did.unwrap().starts_with("did:"), "resolved value should be a DID" ); }$

gbl08ma@gbl08ma· May 5

No dice yet, new error message but it still sounds like it is attempting HTTP resolution / not ignoring the error it gets while trying HTTP resolution. I suspect this is a problem somewhere within tangled.org/ngerakines.m... but I don't know enough Rust to be able to quickly put up a repro case.

Screenshot of a terminal showing:

329406be938e:/lichen-root/next.gbl08ma.com# git push
2026-05-05T09:32:00.361791Z ERROR atproto_identity::resolve: error=error-atproto-identity-resolve-6 Invalid HTTP resolution response: expected DID format
Everything up-to-date
329406be938e:/lichen-root/next.gbl08ma.com# git remote -v
origin pds://gbl08ma.com/next.gbl08ma.com (fetch)
origin pds://gbl08ma.com/next.gbl08ma.com (push)

gbl08ma@gbl08ma· May 4

Also, I'm not sure if this is intentional (I suspect not, and I suspect it's kinda broken... it created folders for these on my instance...) the test websites I created on your official instance show up in the listing on my instance

❀ sites
Domain Size Actions
gbltest.lichen.page 369.4 KB
🌐
🖊
⚙
hugogbltest.lichen.page 80.6 KB
🌐
🖊
⚙
next.gbl08ma.com 281.0 MB
🌐
🖊
⚙

gbl08ma@gbl08ma· May 4

Yeah that doesn't look right. The PDS remote thingy is failing to resolve my handle to my PDS and instead is just fetching my old wordpress website at gbl08ma.com. I think it is trying to do HTTP handle resolution, but my handle is set up via DNS records.

error: failed to push some refs to 'pds://gbl08ma.com/next.gbl08ma.com'
6a57c731d671:/lichen-root/next.gbl08ma.com# git push 2>&1 | head -10
git-remote-pds: error: resolveHandle failed (404 Not Found): <!DOCTYPE html>
<html lang="en-US">
<head>
<meta charset="UTF-8" />
<title>Page not found | gbl08ma's personal website</title>

gbl08ma@gbl08ma· May 4

Unfortunately the temp-preview posts are still showing up in post lists

Screenshot showing a Test post with a URL pointing to test-post_temp-preview

gbl08ma@gbl08ma· May 1

My #BandcampFriday haul. All over the place genre-wise, as usual. Feat. works by @deadposssociety.bsky.social, @nthercyte.bsky.social, @arikxu.bsky.social and others + the latest experimental music compilation by @brokecore.bsky.social (@zovi.bsky.social, @masr.bsky.social...) Links in alt text.

'World of Promises' by Michael Oakley - https://newretrowave.bandcamp.com/track/world-of-promises-2

'TRASH CATS: LIVE ON LIMEWIRE' by ASH - https://deadposssociety.bandcamp.com/album/trash-cats-live-on-limewire

'Valuables And Malleables' by Brokecore Recordings - https://brokecore.bandcamp.com/album/valuables-and-malleables

'Reflection EP' by Nthercyte - https://nthercyte.bandcamp.com/album/reflection-ep

'The Night We Changed' by DotMid - https://dotmid.bandcamp.com/album/the-night-we-changed

'REALMTREADER' by DotMid - https://dotmid.bandcamp.com/album/realmtreader

'Vainglory' by Xöchipilli - https://aedis-syopeen.bandcamp.com/album/vainglory

'Foxlore' by ArikXu - https://aedis-syopeen.bandcamp.com/album/foxlore

gbl08ma@gbl08ma· Apr 19

@illegal.trading there's a timezone handling bug on lyda.app. This happens in other places on the app where relative timestamps are shown, too.

screenshot of lyda.app showcasing a notification containing a "56 minutes until" relative timestamp

gbl08ma@gbl08ma· Feb 27

Man, your music and videos are so cool but spamming (un)follows and likes is really uncool.

Screenshot of https://pdsls.dev/at://did:plc:egjdrp7pjac6frfs4exrky7m/app.bsky.graph.follow showing a strategically short amount of inorganic follows happening every few seconds throughout a short period

Screenshot of https://pdsls.dev/at://did:plc:egjdrp7pjac6frfs4exrky7m/app.bsky.feed.like showing inorganic likes happening every few seconds throughout a short period

gbl08ma@gbl08ma· Feb 12

Nothing like a good old unscheduled server downtime to make me waffle about atproto internals to uninterested non-technical users in hopes of making them understand what the "Create a rotation key" in PDS MOOver does 😭

Screenshot of Discord messages from me to someone else:

it is basically a stable identifier that is used across bluesky&co to refer to your account
it's what "holds" your followers
these identifiers are associated with a "DID document" and you can see yours here: https://plc.directory/did:plc:[REDACTED]
this "document" says what your handle is (so every time you change your handle you're actually making a change to this document) and also says what PDS your account is hosted on (so when you migrate your account from one PDS to another, you're also making a change to this document)
to update this document, you need a "private key" for one of the "public keys" configured in the DID document (think of the "private key" as an actual 🔑 and the "public key" as the 🔒 it fits into. everyone is able to see the lock on your door, but you must keep your 🔑 to yourself)
in order to let you change handle, the PDS (in this case, my PDS) has one of the keys to your DID document
by adding a second key that you have, you are later able to change this document yourself without intervention of my PDS (e.g. because it went offline for good and I'm gone)
and this would allow you to keep your did:plc:[REDACTED] and thus all your followers, and all the other "social relationships" that "point to you"
without adding a second key, if my PDS disappears, you won't have the 🔑 for the single 🔒 that is currently in your DID document
by adding a second key, you are adding a second 🚪 with its own 🔒 and its own 🔑 so you can get back home even if my PDS loses the 🔑 to the existing single 🚪
if you go here https://plc.directory/did:plc:[REDACTED]/log/audit - or for a nicer view, click on the Logs tab in pdsls

you'll be able to see all the changes you've made (or which your PDS has made for you) to your identity document
you can see when you added a custom handle

gbl08ma@gbl08ma· Jan 31

In other news,

Renewal confirmation for clouttery.xyz

Total charged: USD $12.98

gbl08ma@gbl08ma· Jan 25

How are you keeping track of this? The impression I got months/weeks ago, from looking at the logs of my labeler, is that many or even most did:webs are broken in some way. Your stats don't seem to deny that 😅

Screenshot of lots of log messages, such as:

ailed to reprocess labels of did:web:asyomei.su: Error: Failed to identify PDS of did:web:asyomei.su for label reprocessing: FailedDocumentResolutionError: failed to resolve did document; did=did:web:asyomei.su
Failed to reprocess labels of did:web:cfb05489.rxi.cx: Error: Failed to identify PDS of did:web:cfb05489.rxi.cx for label reprocessing: FailedDocumentResolutionError: failed to resolve did document; did=did:web:cfb05489.rxi.cx
Failed to reprocess labels of did:web:catstrogen.diy: Error: Failed to identify PDS of did:web:catstrogen.diy for label reprocessing: DocumentNotFoundError: did document not found; did=did:web:catstrogen.diy
Failed to reprocess labels of did:web:atproto.bront.rodeo: Error: Failed to identify PDS of did:web:atproto.bront.rodeo for label reprocessing: DocumentNotFoundError: did document not found; did=did:web:atproto.bront.rodeo
Failed to reprocess labels of did:web:dawn.gaze.systems: Error: Failed to identify PDS of did:web:dawn.gaze.systems for label reprocessing: FailedDocumentResolutionError: failed to resolve did document; did=did:web:dawn.gaze.systems
Failed to reprocess labels of did:web:d.on-t.work: Error: Failed to identify PDS of did:web:d.on-t.work for label reprocessing: DocumentNotFoundError: did document not found; did=did:web:d.on-t.work

gbl08ma@gbl08ma· Jan 25

"You can dismiss it," they say right as I find out there's no way to remove it without using third-party tools. So yeah, today was the day I started using uBO to "fix" the official Bluesky app. It is now joining the ranks of such great apps as Discord, etc.

Screenshot of me using uBlock Origin's picker to create a filter for the "NFL Playoffs" eyesore in the sidebar.

gbl08ma@gbl08ma· Jan 13

I continue to annoy my friend with long-winded explanations of my ambitions for a decentralized approach to plc.directory. Keep in mind this is being sent to someone who barely uses bsky and who, I only realized today, didn't even know what did:plc was. I often wonder how people put up with me.

Screenshot of way too many Discord messages from myself yapping about my didplcbft project to a friend (the friend is not pictured, such is my wall of text)

gbl08ma@gbl08ma· Jan 10

Does anyone know what are the rate limits on GETs from plc.directory? I don't see mentions to them in the spec nor in the API docs.

http://plc.directory/export?after=0 sent back an error.

Error code: 429 Too Many Requests

gbl08ma@gbl08ma· Jan 6

OK I just realized this message is gibberish to 99% of people, but hopefully this makes a bit more sense.

Screenshot of a Discord conversation with myself (I am very normal). Sorry the contents are too big to fit in the alt text. You'll have to use OCR.

gbl08ma@gbl08ma· Jan 2

Screenshot of messages I'm sending to someone else on Discord:

this is a theme you'll see, as you know, throughout the linux userspace ecosystem
did you know arch has its own d-bus implementation
https://wiki.archlinux.org/title/D-Bus#dbus-broker
this will forever be the thing I love about linux and the thing that makes it a nightmare for people who want things to just work
you can't exactly replace, say, the registry in windows with a different implementation
meanwhile on linux there are like 10 different, from-slightly-to-very-incompatible, notions and implementations of what "the registry" should be
as a software developer, you should pick the one you like the most/are most familiar with
and then spend the time you'd take to support the other N-1 implementations arguing with people online about why the solution you chose to support is superior
or at the very least, arguing about why linux is superior to windows
this all used to be a joke
the fact that it is becoming less of a joke by necessity is all due to microsoft deciding to not care about the consumer desktop
or at the very least deciding to care about the wrong metrics (like AI adoption, monthly subscriptions to cloud services, etc.)

gbl08ma@gbl08ma· Dec 24

Cover art for the 2019 indie game Untitled Goose Game, developed by House House and published by Panic Inc.

gbl08ma@gbl08ma· Dec 18

Not only does the official did:plc server implementation not check this, there are operations in the PLC that have duplicated keys, so I'll have to remove this check. The did:plc spec is "great": there are things it doesn't specify, and some of the things it specifies are not actually enforced.

Screenshot of Go code that checks for duplicate rotation keys in PLC operations, with a highlighted comment saying "the typescript implementation doesn't appear to check this at first sight, but the spec says "must include least 1 key and at most 5 keys, with no duplication""

gbl08ma@gbl08ma· Dec 18

On the did:plc front, I've been toying around with some ideas that are likely to be controversial. Some of the reasons will probably be valid, but I bet there is mostly going to be a lot of knee-jerking going around. Maybe it won't matter: I might never get this to a state I'm happy publishing.

$Screenshot of a tree representation of a JSON object where one of the fields contains the string " --- at tangled.org/gbl08ma/didplcbft/plc/impl.go:84 (plcTransaction.ValidateOperation) ---\nCaused by: operation failed validation\n --- at tangled.org/gbl08ma/didplcbft/plc/impl.go:45 (plcImpl.validateOperationWithinTransaction) ---\nCaused by: creation operation not allowed as DID already exists\n --- at tangled.org/gbl08ma/didplcbft/plc/operation_validator.go:81 (V0OperationValidator.Validate) ---\nCaused by: invalid operation sequence"$

gbl08ma@gbl08ma· Dec 11

*11pm takes 4 characters at best duh. I swear I didn't get my spreadsheet wrong

24h time 24h 12h
0 1 4
1 1 3
2 1 3
3 1 3
4 1 3
5 1 3
6 1 3
7 1 3
8 1 3
9 1 3
10 2 4
11 2 4
12 2 4
13 2 3
14 2 3
15 2 3
16 2 3
17 2 3
18 2 3
19 2 3
20 2 3
21 2 3
22 2 4
23 2 4

gbl08ma@gbl08ma· Dec 5

My haul this #BandcampFriday. This time it's too much for me to share individual links, go through my collection instead: bandcamp.com/gbl08ma Feat. work by @debiscomusic.com, @arikxu.bsky.social, @senpimusic.com, @theastonshuffle.com, @elke.cafe, @nthercyte.bsky.social, Kettama, Bad Suns and more

Screenshot of some of the new additions to my Bandcamp collection. Visit my collection for details (everything more recently added than ArikXu's Magic Mints, inclusive, is new)

gbl08ma@gbl08ma· Dec 5

Saying goodbye to my old bio. It all still applies, but I wanted to pack more/different stuff in. One of these days I'll work on a better pinned post, too, and get a banner image like the fancy kids.

Software developer. Most comfortable with Go, can do frontend too. I built jungletv.live, underlx.com, tny.im and more. I know more about the internals of GTA V and Watch Dogs than their average players. Sometimes I make music and high effort shitposts.

gbl08ma@gbl08ma· Dec 5

Within my bubble, I feel this is the state of ATProto decentralization as of late:

gbl08ma@gbl08ma· Nov 30

The other day, I grabbed a Bitwig Studio license for a pretty good price. I figured I could splurge a bit, considering I've spent more hours in DAWs than in most of the few full-price games on my Steam library/backlog. Now, if only turning this into a complete track was as easy as spending money...