sxan

I have a situation where generics would be useful: a server (that I do not control or influence) with many API endpoints that each returns very similar json. There's an envelope with common attributes and then an embedded named substructure (the name differs in the return value of each call) of a different type.

Without generics, you could do something like:

type Base struct {
   // common fields
}

type A {
   Base
   A struct {
      // subtype fields
   }
}

type B {
   Base
   B struct {
      // subtype fields
   }
}

but then you'd need to either duplicate a bunch of API calling and unmarshalling code, or consolidate it and do a bunch of type casting and checking.

Generics to the rescue: subtypes become specific types for a general type:

type Base[T any] {
   // common fields
   Subfield T
}

type A struct {
  // subtype fields
}

type B struct {
  // subtype fields
}

It even looks cleaner! Ah, but the rub is that the marshaled field name Subfield is the same for every type: there's no way to specify a tag for a struct type so that Subfield is un/marshaled with a name specific to the type.

https://go.dev/play/p/3ciyUITYZHk

The only thing I can think of is to create a custom unmarshaller for Base and use introspection to handle the specific type.

Am I missing a less hacky (introspection is always hacky) way to set a default tag for any field of a given struct type? How would you do this?

This pattern - APIs using envelopes for data packets - is exceedingly common. I can't believe the only way to solve it on Go is by either mass code duplication, or introspection.

I do my keyboard configuration with Vial, which may or may not be relevant.

I am unable to momentarily switch layers from a particular layer, and I'm looking for tips.

I have a base Dvorak layer, with all of my layer switches as tap-dance keys under my left hand, with holds triggering a momentary layer switch and all of the other keys under my right hand: a layer for punctuation, a layer for numbers, a layer for function keys, for WM navigation, for tmux navigation... 9 layers in total. It all works well.

Recently, I started playing Factorio again, so I set up a combo switch to the 9th layer, which is bog-standard QWERTY, it being easier to just learn new muscle memory than to reconfigure all 9,000 Factorio key bindings for Dvorak. But now entering numbers was a PITA because my keyboard has no number keys, so I have to switch back to the base layer to use the MO binding to switch to my number layer.

Eventually, I decided this was too much trouble, so I created a tap-dance MO binding for the same physical key in the QWERTY layer... but it doesn't work, in that the layer is not switched to the number layer - except for "0": that combination works. The fact that one key works makes me think it is actually sorta switching layers? But all of the other keys just enter the un-switched QWERTY keys.

I've tried setting the trigger key to a different one, with identical results. All of the keys on the left hand (and under the trigger key) are KC_TRNS on the number layer, so in both cases I've tried the trigger key is KC_TRNS on the number layer. I have not yet tried duplicating the number layer and using that instead.

Does the target layer (the number layer) have to be a layer number greater than the starting layer? Number layer is layer 4, and QWERTY is 9 - do I need to move 4 to 10? Is there some other, common, issue I'm encountering?

Like, not technically how, but emotionally? If I spend too much time messing around on a platform, critters inevitably attack my base. Even if I build a fortress, I worry that something will run out and guns will run out of ammo... or that something will run out and The Factory will grind to a halt. I could just stack up a vast area of capacitors and rely on lasers and a fission reactor, but is this really what you guys are doing?

How do you emotionally detach from Nauvis and commit to not being able to troubleshoot on the home factory? Heck, once I establish factories on other planets, how do I leave them to return to Nauvis and not worry that they'll be overrun??

When Space Age was released I restarted, solo, with a new base, and I'm getting close to building a traveling platform; how do I ensure the security of Nauvis before I depart?

(My first, and as yet only, station attached for giggles)

I'd like to know if I should file a bug, or if this is just something in my config. In particular, I'd be grateful for confirmation about what I'm seeing.

In Raccoon, in this post, I see only the first paragraph of the post. If I "view raw," I can see the entire post text; it also shows up correctly in Thunder, and in the web interface.

I don't see a way to expand the post, and I'm not sure if it's only this post or if it's happening a lot and I just haven't noticed, because the only way to detect missing content is through the "view raw" function.

Is this a setting, or a bug? I

I was thinking about this before the Tholian wave, but it's apropos.

It unscientifically appears to me that TOS had a far higher incidence on non-humanoid aliens than later series. Tholiens, Horta, the flying neural parasites on Deneva; while there were many bipedal aliens sometimes differing only by skins color, many were non-bipeds or were bipedal but radically different from humans, like the Gorn and the salt vampire. In later series, it seems nearly all aliens were reduced to bumpy head species.

TOS ran for only three seasons, and truly different aliens are expensive; I understand the economics of going the prosthetic forehead route. And it's difficult to have recurring truly alien biology in a series.

My question is whether anyone's done a statistical analysis covering the originality of aliens, per series, based on divergence from the humanoid base. Does it only seem like TOS had more different types of aliens (intelligent and non) because it was so short, or was the universe really more diverse in TOS?

Rook provides a secret service a-la secret-tool, keyring, or pass/gopass, except backed by a Keepass kdbx file.

The problem Rook solves is mainly in script automation, where you have aerc, offlineimap, isync, vdirsyncer, msmtp, restic, or any other cron jobs that need passwords and which are often configured to fetch these passwords from a secret service with a CLI tool. Unlike existing solutions, Rook is headless, and does not have a bespoke secrets database full of passwords that must be manually synchronized with Keepass; instead, it uses a Keepass db directly.

Rook is in the AUR and in Alpine testing; binaries are available from the project page.

From the changelog, since the last Lemmy release announcement v0.1.3 on May 20:

[v0.2.0] Fri Oct 11 09:01:03 2024 -0500

Added

• support for password + key file credentials
• show --no-eol option, to strip CRs after, eg, passwords

Changed

• show matches search: it's now case insensitive

Fixed

• successful OPEN with password wasn't clearing the one-time pin, so the DB was staying locked.

stmps is a fork of stmp, under active development and with several additional features. (*) items are PRs which also been accepted by the stmp project.

• mpris support (*)
• improved help text
• improved playlist handling, including concurrent loading in the background
• improved browser behavior, e.g. add all songs by an artist
• global, server-side search
• artist search in the browser (*)
• TUI-less server information query
• queue reordering
• queue shuffling
• randomly add songs to the queue
• randomly add similar songs to the queue, using the Subsonic "get similar songs" feature

It's fast, keyboard driven, and a single executable; it is regularly tested against Navidrome and Gonic.

stmps can be installed by a simple go install command, and it's also in AUR as stmps.

I'm not the author, but am one of the active contributors.

The reactions to most posts are overwhelmingly negative and critical. Ironically, posts to c/unpopularopinion tend to argue that they agree with the post, and are consequently more supportive.

I'm posting here because I have nowhere else to post. If you squint, this meets the community rules because my current keyboard is a Piantor/42, and my issue stems from a combination of 40% and QMK behavior. Although, to be honest, this is mostly about QMK, but using Discord is painful, and I'll go there only as a last resort.

For a long while, I used Kanata on my laptop, and desktop an ErgoDox, having replaced kmonad because of one certain feature: tap-hold key sequence behavior. It's best described here, but the tl;dr is that (press lsft) (press a) (release lsft) (release a) where a is a tap-hold key should output "A" and not "a" -- kmonad outputs "a".

A few months ago, when I got my Piantor, I discovered that this sequence outputs no character, and although there's an option that makes it output "a", I can't find a combination that makes it output "A". I'm asking whether, in the bewildering set of QMK variables, is there a way to configure QMK s.t. the sequence (press lsft) (press a) (release lsft) (release a) outputs "A"?

That's the main thrust of my question. As a sort of addendum, I think this behavior is behind another of my QMK irritations: I'm a reasonably fast typer, and often will be typing the next key before I've completely released the previous key. This means I have to set a large-ish time-out before tap-hold engages, which introduces an annoying delay whenever I want to chord a layer and get at, e.g. numbers. I do understand that this is may be an unsolvable issue, that it's just an unavoidable limitation on small keyboards in having so many common keys (numbers, punctuation, and arrows are the worst -- coding, nearly half the text are characters from layers). Either I have a long timeout and and live with an annoying delay when I want to type (many) punctuation characters or numbers; or I have a short timeout and frequently accidentally shifting layers. However, I feel as if this might be mitigated somewhat with the Kanata-style key sequence handling, because even though my Kanata configuration is nearly an exact mirror of my QMK layer configuration, I never have this problem with Kanata.

I suppose I could give up on using QMK for anything except the most fundamental mapping, and use Kanata instead. However, there's an appeal to the portability of having the programming in the keyboard itself; it makes me a little less dependent on the computer to which the keyboard is attached.

I'm posting here because I have nowhere else to post. If you squint, this meets the community rules because my current keyboard is a Piantor/42, and my issue stems from a combination of 40% and QMK behavior. Although, to be honest, this is mostly about QMK, but using Discord is painful, and I'll go there only as a last resort.

For a long while, I used Kanata on my laptop, and desktop an ErgoDox, having replaced kmonad because of one certain feature: tap-hold key sequence behavior. It's best described here, but the tl;dr is that (press lsft) (press a) (release lsft) (release a) where a is a tap-hold key should output "A" and not "a" -- kmonad outputs "a".

A few months ago, when I got my Piantor, I discovered that this sequence outputs no character, and although there's an option that makes it output "a", I can't find a combination that makes it output "A". I'm asking whether, in the bewildering set of QMK variables, is there a way to configure QMK s.t. the sequence (press lsft) (press a) (release lsft) (release a) outputs "A"?

That's the main thrust of my question. As a sort of addendum, I think this behavior is behind another of my QMK irritations: I'm a reasonably fast typer, and often will be typing the next key before I've completely released the previous key. This means I have to set a large-ish time-out before tap-hold engages, which introduces an annoying delay whenever I want to chord a layer and get at, e.g. numbers. I do understand that this is may be an unsolvable issue, that it's just an unavoidable limitation on small keyboards in having so many common keys (numbers, punctuation, and arrows are the worst -- coding, nearly half the text are characters from layers). Either I have a long timeout and and live with an annoying delay when I want to type (many) punctuation characters or numbers; or I have a short timeout and frequently accidentally shifting layers. However, I feel as if this might be mitigated somewhat with the Kanata-style key sequence handling, because even though my Kanata configuration is nearly an exact mirror of my QMK layer configuration, I never have this problem with Kanata.

I suppose I could give up on using QMK for anything except the most fundamental mapping, and use Kanata instead. However, there's an appeal to the portability of having the programming in the keyboard itself; it makes me a little less dependent on the computer to which the keyboard is attached.

Edit 2024-10-01

Another person posted about a similar need, and I decided to create a matrix document to track it, in the hope that those of us looking for this specific use case could come up with the best solution. The idea here is that, while many OSS social media projects are capable of being used like a Fcbook wall, they don't all necessarily provide an ideal user experience. Feature set is not equivalent to being designed for a specific use case, and the desired workflow should be the primary means of interacting with the service. The (for now) open document tracking this is here.

I'm a little surprised I can't find any posts asking this question, and that there doesn't seem to be a FAQ about it. Maybe "Facebook" covers too many use cases for one clean answer.

Up front, I think the answer for my case is going to be "Friendica," but I'm interested in hearing if there are any other, better options. I'm sure Mastodon and Lemmy aren't it, but there's Pixelfed and a dozen other options with which I'm less familiar with.

This mostly centers around my 3-y/o niece and a geographically distributed family, and the desire for Facebook-like image sharing with a timeline feed, comments, likes (positive feedback), that sort of thing. Critical, in our case, is a good iOS experience for capturing and sharing short videos and pictures; a process where the parents have to take pictures, log into a web site, create a post, attach an image from the gallery is simply too fussy, especially for the non-technical and mostly overwhelmed parents. Less important is the extended family experience, although alerts would be nice. Privacy is critical; the parents are very concerned about limiting access to the media of their daughter that is shared, so the ability to restrict viewing to logged-in members of the family is important.

FUTO Circles was almost perfect. There was some initial confusion about the difference between circles and groups, but in the end the app experience was great and it accomplished all of the goals -- until it didn't. At some point, half of the already shared media disappeared from the feeds of all of the iOS family members (although the Android user could still see all of the posts). It was a thoroughly discouraging experience, and resulted in a complete lack of faith in the ecosystem. While I believe it might be possible to self-host, by the time we decided that everyone liked it and I was about to look into self-hosting our own family server (and remove the storage restrictions, which hadn't yet been reached when it all fell apart), the iOS app bugs had cropped up and we abandoned the platform.

So there's the requirements we're looking for:

• The ability to create private, invite-only groups/communities
• A convenient mobile capture+share experience, which means an app
• Reactions (emojis) & comment threads
• Both iOS and Android support, in addition to whatever web interface is available for desktop use

and, given this community, obviously self-hostable.

I have never personally used Facebook, but my understanding is that it's a little different in that communities are really more like individual blogs with some post-level feedback mechanisms; in this way, it's more like Mastodon, where you follow individuals and can respond to their posts, albeit with a loosely-enforced character limit. And as opposed to Lemmy, which while moderated, doesn't really have a main "owner" model. I can imagine setting up a Lemmy instance and creating a community per person, but I feel as if that'd be trying to wedge a square peg into a round hole.

Pixelfed might be the answer, but from my brief encounter with it, it feels more like a photo-oriented Mastodon, then a Facebook wall-style experience (it's Facebook that has "walls", right?).

So back to where I started: in my personal experience, it seems like Friendica might be the best fit, except that I don't use an iPhone and don't know if there are any decent Friendica apps that would satisfy the user experience we're looking for; honestly, I haven't particularly liked any of the Android apps, so I don't hold out much hope for iOS.

Most of the options speak ActivityPub, so maybe I should just focus on finding the right AP-based mobile client? Although, so far the best experience (until it broke) has been Circles, which is based on Matrix.

It's challenging to install and evaluate all of the options, especially when -- in my case -- to properly evaluate the software requires getting several people on each platform to try and see how they like it. I value the community's experience and opinions.