# Learn more about twtxt:
#     https://twtxt.readthedocs.io/en/stable/
#     https://www.uninformativ.de/twtxt-mention.html
#
# nick   = movq
# url    = https://www.uninformativ.de/twtxt.txt
# avatar = https://www.uninformativ.de/avatar.png
#
# For out-of-band replies: `echo ed.vitamrofninu@txtwt | rev`
#
# This file is available via Gopher, too:
#     url = gopher://uninformativ.de/0/twtxt.txt
# However, since Gopher has no concept of user agents or any kind of
# headers, I will not be able to see that you're following me and, thus,
# will not be aware of any replies or mentions that you send. Please
# either use the HTTPS feed or drop me an email. :)
#
# Legacy, don't use:
# nick_alias = vain
#
# prev = i2wnvaa twtxt-old_2021-11-08_i2wnvaa.txt
2021-11-08T16:14:26Z	(#xlzkkca) @<stigatle https://twtxt.net/user/stigatle/twtxt.txt> I’m surprised, is that small antenna is enough to do something meaningful?
2021-11-08T16:14:36Z	(#mmkag5a) @<quark https://ferengi.one/twtxt.txt> 🙇
2021-11-08T17:20:34Z	(#xlzkkca) @<fastidious https://arrakis.netbros.com/user/fastidious/twtxt.txt> So bigger isn’t always better? 😁 (Now you know that I have virtually zero knowledge in that area.)
2021-11-08T17:21:33Z	(#ei6zf3a) To this day, I have not understood what Pinterest is. I only know that … it ruins my search results. And that I hate it with a passion. Yup.
2021-11-08T18:23:46Z	(#v4bt6ga) @<lyse https://lyse.isobeef.org/twtxt.txt> Right. `/dev/brain` wasn’t mounted. 🤦
2021-11-09T12:50:46Z	(#v4bt6ga) @<stackeffect https://twtxt.stackeffect.de/stackeffect.txt> You cancel a Git commit by writing an empty commit message. Similarly, you cancel composing a new twt by writing an empty twt. That’s how it works in jenny. (You don’t have to delete *everything*, just everything above the “snip” line. I just find it easier to delete it all.)

Confusion arises from the fact that jenny *pre-fills* parts of a twt when you reply to someone. When you just quit the editor in that situation, then you didn’t write an empty twt (it contains a conversation hash and a mention), so it won’t be discarded.

That’s why I compared it to `git rebase` and not `git commit`. `git rebase` pre-fills your editor with some commands as well, and to abort all this, you have to delete them, save, quit. (A rebase just *happens* to not do anything if you `pick` everything, but it does process your commands – they just happen to be all no-ops. It doesn’t just abort and discard.)

Okay, fine. jenny isn’t Git, so it doesn’t really matter anyway.

Is the current behaviour good UX? I don’t know. “Delete everything, save, quit” is pretty much hard-wired into *my* brain, so it’s what *I* expect. So it’s good UX for *me*.

It would be easy to implement an `mtime` check (so that jenny aborts if you never save the file at all). I already did it in a local branch, it’s a diff that changes three lines. What I don’t like about this approach: It introduces a heuristic that tries to detect if the user changed the file or not. So what if the user does save the file (because they routinely exit their editor with “save-quit” or whatever) but they didn’t change anything? mtime check will fail, so we have to compare the file’s contents now. And now we’re in a situation where some content is *illegal* to have in a twt – you can now no longer write a twt that contains only `(#convhash) @<some person>`. Is that a problem? Maybe not, but it means you can’t just bump a conversation with a mention (we routinely do stuff like that at work). Also, in general, I’d like users to be able to put anything they like in a twt, and that won’t work anymore now.

So … I don’t know. I’m happy with “delete to cancel”, because it is a very explicit command. *And* it’s easy to implement.

If people are truly unhappy with this (I don’t think you are as you said it’s unexpected but okay), then I’d rather implement an interactive prompt after the user has left the editor: “Do you want to publish this twt [Y/n]? ‘$quote_of_twt’” (In fact, my `publish_command` does that. It’s not very UNIX-y, but it’s useful. Sometimes, I change my mind about writing a twt a split second after I left the editor. 🥴 Or stuff like that.)
2021-11-09T16:51:14Z	Can someone explain this?

Open this page in Firefox or Chromium:

https://movq.de/v/f9c0330a69/script.html

Feel free to inspect it using `curl` first. There are 500 `<script>` tags in the header, all of them point to non-existent files, so you’ll just get 404 for everything.

What I’m seeing is this:

https://movq.de/v/a8818b7210/requests.jpg

1.  The browser fetches `script.html` itself.
2.  The browser goes on to retrieve all 500 dead script sources. It does that in parallel using several workers.
3.  Once it’s done with that, it *starts over* and retrieves all 500 scripts *again*, but only using one worker this time.

Why is the browser doing that? 🤔

This only happens in Firefox and Chromium, not in my WebKit browser (WebKit only does 500 requests). It also doesn’t happen with `<link>` tags in the header, nor `<img>` tags in the body:

-   https://movq.de/v/f9c0330a69/link.html
-   https://movq.de/v/f9c0330a69/img.html

I’ve seen this on nginx, OpenBSD httpd, and some node.js “web server” now, so I don’t think it’s related to the server.

🤔 😕
2021-11-10T14:03:52Z	(#hfal6kq) Hmm … I’m really not sure.

Yes, editing twts is problematic because the hash changes. This can lead to broken threading and confusion.

*Deleting* twts is a different story. Why is that a problem? 🤔

Also, are we just talking about the implementation in `yarn` here or twtxt in general?
2021-11-10T17:29:41Z	(#v4bt6ga) Screw it. I added the mtime check. It’s easy to do and hopefully covers most cases.

There are three ways to cancel now:

1.  Save an empty twt.
2.  Quit the editor without ever saving anything to the file.
3.  Quit the editor with an exit code other than 0 (":cq!" in Vim).

I think this is good enough. I won’t add anything that works on the twt’s content (i.e. “same content = cancel”), because I want to be able to write twts with arbitrary content (even if it happens to be identical to that pre-filled content).
2021-11-10T17:32:46Z	(#2kwts7a) 🥳 🥳 🥳

Now we’ll just have to implement it … How hard can it be 🥴
2021-11-10T17:56:40Z	(#2phtura) @<fastidious https://arrakis.netbros.com/user/fastidious/twtxt.txt> Thanks, fixed now.
2021-11-10T18:33:41Z	(#hfal6kq) I have to agree with @<lyse https://lyse.isobeef.org/twtxt.txt>. 🤔

But let’s not forget – and I wasn’t really aware of this, to be honest: It is already impossible to edit or delete older twts on yarn. You can only edit/delete the last one. So this *already* makes me uncomfortable. Dunno, I really really like to have the option to edit or delete old posts. Opinions change over time, maybe I was drunk while posting something, or whatever.

(I don’t really believe in “once it’s public, then it’s too late”. I’ve seen so much stuff vanish from the Internet over the years … It’s only too late if something goes viral and millions of users make copies of your funny face. 🥴 And even *then*, stuff gets lost.)

I’m not a yarn user, so I don’t have a *strong* opinion about this, tbh. I can always just edit my file. 🤔 I just don’t think it’s a good idea to limit the ability to edit/delete.
2021-11-11T13:48:13Z	(#4iszuda) @<fastidious https://arrakis.netbros.com/user/fastidious/twtxt.txt> I accidentally summoned a demon. 😱😱😱
2021-11-11T14:28:36Z	A new Aqual(o)unge release on https://machtdose.de/ 🥳🥳🥳
2021-11-11T14:29:53Z	“no one wants to spend time fixing anything related to X11 now as Wayland is just around the corner” https://gitlab.freedesktop.org/xorg/xserver/-/issues/1241#note_1153655

That guy is scaring me. 😬
2021-11-11T15:00:56Z	This person is following me now: `+{{ .Profile.URL }}; @{{ .Profile.Username }}` 🥴
2021-11-11T15:58:32Z	The Matrix crashed again. I’m just seeing this: https://movq.de/v/31acfb7f36/s.png A plain white wall. Does the world still exist? Am I dead? 🤔
2021-11-11T16:51:59Z	(#te2mlfa) @<lyse https://lyse.isobeef.org/twtxt.txt> Yeah, they keep saying that all the time. I don’t know. 🤔 If Wayland didn’t put so much responsibility on the “compositor”, I’d be much less worried about it. But this makes it super hard for small(er) projects to implement a functional Wayland “server”.
2021-11-11T16:52:18Z	(#rbpcrlq) @<fastidious https://arrakis.netbros.com/user/fastidious/twtxt.txt> It’s an IP of OVH. 🤔 Starts with 198.
2021-11-11T16:53:32Z	(#rbpcrlq) @<lyse https://lyse.isobeef.org/twtxt.txt> Whoops. 😨 Sorry for this Datenreichtum.
2021-11-11T19:47:00Z	(#te2mlfa) @<lyse https://lyse.isobeef.org/twtxt.txt> Here’s my take on it (that’s last time I checked – maybe some things changed):

-   The Wayland wire protocol aside, you have to do everything yourself: Manage display buffers (let’s say “on an OpenGL level”), handle monitors, handle mouse/keyboard/whatever input, … Basically everything that the X server does today. *You* have to write that code now.
-   There is basically only one library that truly tries to help you here: wlroots https://gitlab.freedesktop.org/wlroots/wlroots/ (this is already a “third party” library, it’s from the Sway project; Sway being a port of i3wm to Wayland)
-   On top of that, you implement decorations, window management, …
-   Then there is a myriad of protocol extension and a little battle between GNOME, KDE, and the rest of the world. For example, GNOME wants to enforce client-side decorations, the others don’t. Many of the protocol extensions are still unstable. To give you an idea of what is meant by “protocol extension”: https://wayland.app/protocols/
-   You very much need *an X11 compatibility layer* that runs on top of Wayland for backwards compatibility, and this is usually implemented in the compositor as well.

What X11 gave us was a server that allowed programs to draw windows. And then we could have totally different programs that managed those windows. This was *awesome*. Yes, there are problems. But instead of refining this model and extending it, we are being nudged into having “the Wayland compositor”, which is responsible for everything.

Wayland cares a lot about “each frame is perfect”. Clients draw a full frame into a buffer and then the compositor will use that buffer to assemble the next frame. So there is no more “flickering”, like we have with X11 now. (You’re probably used to it and don’t notice it.) This is one of the main goals and I *suspect* that this is part of the reason why everything has to be done in the compositor? Not sure.

I *think* it would be possible to split all this into different components. Problem is, nobody does it that way. There’s GNOME, which doesn’t care about any of this, they’re a huge project with a ton of manpower, they just do what they want. I don’t know what KDE does. Sway made wlroots, which is probably the best thing we have, but even Sway is just a library that you link against, so you end up with one single binary.

Just look at “TinyWL”, the “minimum viable product Wayland compositor based on wlroots”: https://gitlab.freedesktop.org/wlroots/wlroots/-/blob/master/tinywl/tinywl.c – Look at the size of that thing!

I heard people complain and give up, because Wayland/wlroots is moving so fast that they can’t keep up with their own “toy” compositor. @<z3bra gopher://g.nixers.net/0/~z3bra/twtxt.txt> tried it and eventually gave up.

Note that there were many small Wayland compositors in the beginning, but they all died. I don’t know why, though.

Bottom line is: With X11, it was easy enough to write a window manager, so that one single person could do it. This allows me to run a fully customized desktop environment. With Wayland, there is no “server” anymore and I would have to do *so much more work*, that it’s just not feasible anymore. Result: I’d probably be stuck with *some* desktop environment that I’d hardly be able to customize. And that would suck *a lot*.

But that’s only the status quo. Maybe things will change dramatically, when X11 really dies. I suspect there’s a lot of demand for small niche compositors, and I hope that we will get into a similar situation like today (tons of WMs and DEs to chose from, or you write your own). But we’re a long way from that and things are *not* going in that direction.

</rant>
2021-11-11T19:47:22Z	(#te2mlfa) Oof, what a wall of text. So much for twtxt being a “microblogging” platform. 🤪
2021-11-12T12:37:54Z	(#te2mlfa) @<lyse https://lyse.isobeef.org/twtxt.txt> To be fair, sway https://swaywm.org/ is an i3 clone for Wayland. So I guess you might be fine. 😁 (As for the flickering: If you’re used to it, then you don’t notice it. But it’s there, almost 100% sure of it. Question is, if people don’t notice it, how import is it to fix it …)
2021-11-12T12:39:47Z	(#rbpcrlq) @<prologic https://twtxt.net/user/prologic/twtxt.txt> No idea. I dug up the line in my access.log (I usually run a script on it that *only* shows the twtxt user names and URLs) and the full user agent reads as follows: `jenny/latest (+{{ .Profile.URL }}; @{{ .Profile.Username }})` … so … I don’t know what’s going on here. 🤣
2021-11-12T13:38:41Z	(#cvlfm5a) @<jlj https://twt.nfld.uk/user/jlj/twtxt.txt> If I had kids, I’d probably teach them to floss from day 1. 🤣
2021-11-12T13:59:03Z	(#cvlfm5a) @<fastidious https://arrakis.netbros.com/user/fastidious/twtxt.txt> Good thing I don’t have kids, eh? 🥴
2021-11-12T14:15:31Z	Covid cases in Germany are skyrocketing and I don’t know what to say anymore 😒 https://movq.de/v/54b145f833/s.png
2021-11-12T16:07:34Z	(#hwtxckq) @<fastidious https://arrakis.netbros.com/user/fastidious/twtxt.txt> That’s a pretty good summary of the situation.

We *could* have a much higher vaccination rate (we’re at 67%), but people refuse to do it. And everyone gets tired of dealing with Covid, so they start ignoring the restrictions and/or act like nothing happened. Ya, boom. 😒
2021-11-13T14:24:13Z	Finally, all my OpenBSD boxes have now been upgraded to 7.0. 🥳 🐡
2021-11-13T14:28:26Z	(#wbfguja) Last in line was my beloved Pentium 133 from 1996. 25 years old and still kicking. 👊 Well, kind of. Yes, X11 is hardly usable: https://movq.de/v/6b109d0244 But hey, it’s an OS of this day and age!
2021-11-13T14:30:55Z	(#zlx62ba) @<fastidious https://arrakis.netbros.com/user/fastidious/twtxt.txt> You sure? Moon and Venus are quite a bit apart in Stellarium: https://movq.de/v/d12a6ab3ac/s.png
2021-11-13T15:26:43Z	(#zlx62ba) @<fastidious https://arrakis.netbros.com/user/fastidious/twtxt.txt> Hmm, I thought that doesn’t matter in this case. 🤔 Anyway, switched the view to Sydney and this is how it looked earlier today: https://movq.de/v/89eb36b861/s.png
2021-11-13T15:35:13Z	(#wbfguja) @<fastidious https://arrakis.netbros.com/user/fastidious/twtxt.txt> Yep, it is. But of course, it’s brand new compared to the rest of the machine. (I don’t know the date. Got the display *much* later, it was *probably* built in 2008 as it has a sticker with “84L” as a manufacturing date which, according to The Internet, could mean April 2008.)
2021-11-13T15:36:18Z	(#wbfguja) @<prologic https://twtxt.net/user/prologic/twtxt.txt> Pretty slow, yeah. 😅 But hey, it’s my own window manager that I wrote on Linux in 2016 and it’s really *not* optimized for speed.
2021-11-13T18:12:16Z	(#tatqncq) @<jlj https://twt.nfld.uk/user/jlj/twtxt.txt> Oh my god 🤣

101 years? Oof!
2021-11-13T18:15:04Z	(#wbfguja) @<lyse https://lyse.isobeef.org/twtxt.txt> It basically already is in a museum. 🤣 It’s not like I use it daily.

RAM is the main issue, btw. It’s got 64 MB and it’s swapping quite a bit. Then the hard disk is slow, which makes everything worse.
2021-11-13T19:37:33Z	(#tatqncq) @<fastidious https://arrakis.netbros.com/user/fastidious/twtxt.txt> Well, at least [Kotzen](https://en.wikipedia.org/wiki/Kotzen) still exists. 🤮
2021-11-14T10:47:02Z	I think I’ll give GPG, signed commits, and signed release tarballs another chance. 🤔 Let’s see how big of a headache this is going to be …
2021-11-14T11:10:02Z	(#ptpfohq) @<prologic https://twtxt.net/user/prologic/twtxt.txt> Haha, yeah, I saw that you’re signing commits. Now, first question: How do I get your key? `gpg --recv-keys ...` doesn’t find anything. 😬
2021-11-14T13:22:59Z	It’s pretty easy with jenny to use GPG to *sign* your feed (just do it in your `publish_command`), but of course, there’s no verification. 🤔 Should this be on the TODO list? Is it overkill? Nobody’s using it at the moment …
2021-11-14T13:29:54Z	(#lrmnw3a) Signed this feed now. Let’s see what happens. 🤷
2021-11-14T13:53:36Z	(#lrmnw3a) @<fastidious https://arrakis.netbros.com/user/fastidious/twtxt.txt> The key’s brand new, so maybe it takes a while for key servers to sync? (Dunno if they still do that.) I now pushed the key to keys.openpgp.org, just to be sure. 🤷
2021-11-14T13:55:29Z	(#lrmnw3a) In theory, we could simply run `gpg --verify ...` on all feeds we retrieved and refuse to use the feed if that verification fails. So it shouldn’t be super hard to implement. I’m just not sure if it’s *worth it*. 🤔
2021-11-14T14:32:21Z	(#lrmnw3a) … then again, a better solution might be: Allow the user to specify a `validate_feed_command`. That way, users can use anything they like, be it PGP or OpenBSD signify or whatever. So, feed signing would go into `publish_command`, feed validation into `validate_feed_command`. 🤔 And jenny wouldn’t ever have to deal with any of these crypto tools directly.
2021-11-14T14:32:59Z	(#lrmnw3a) Either way, I’m not going to implement this right away. Feed Archival comes next.
2021-11-14T15:22:46Z	(#ptpfohq) @<prologic https://twtxt.net/user/prologic/twtxt.txt> I don’t. 🤔 🤷  I publish my key(s) on my web site, too, so users have multiple independent sources (key servers + web site + maybe a signed e-mail from me). Not sure if that’s best practice, let alone good UX …
2021-11-14T15:46:21Z	(#ptpfohq) @<prologic https://twtxt.net/user/prologic/twtxt.txt> I just put it on my “contact” page and I’ve seen lots of other people do the same. Of course, that doesn’t allow for auto-discovery, but I get the impression, that GPG frowns upon that anyway and that it wants to you jump through several hoops here (build a web of trust, go to key signing parties, …).
2021-11-14T15:48:43Z	(#lrmnw3a) @<prologic https://twtxt.net/user/prologic/twtxt.txt> So, private messages, essentially? I think that’s totally out of scope for “twtxt as a microblogging platform”, *but* it would be a required feature for “twtxt as a social network”. 🤔 Hmm … I guess I’ll have to play with keys.pub a bit (the weekend’s over, though, and my head is full of GnuPG garbage 🤣).
2021-11-14T18:51:41Z	(#ptpfohq) @<lyse https://lyse.isobeef.org/twtxt.txt> @<prologic https://twtxt.net/user/prologic/twtxt.txt> I’ve never been to a key signing party. 🥴 Got one or two encrypted mails nonetheless! (Nah, it was more than that. But it’s so damn rare. I pretty much gave up on PGP. I’m only getting back at using it because of this blog post https://shibumi.dev/posts/keyless-signatures-with-github-actions/, which triggered my arrogance and made me think: “How hard can it be?!” Haha! 🤣)
2021-11-14T19:09:20Z	(#lrmnw3a) @<lyse https://lyse.isobeef.org/twtxt.txt> I see (and have) the use case, to be honest. Well, kind of. When *I* want to share stuff with my family, I do exactly what you guys said, but the problem is: When *another* family member wants to share data, they have no way of doing it. They usually end up doing horrible things like sharing stuff via Google Drive. Rarely do they post it in our Matrix room (probably because Element is a usability nightmare as well).

Some kind of “private social media”, where people could share their stuff, would be really nice.

Is twtxt the right platform/basis for this? I have no idea. 🤔 Right now, I’m very skeptical, but that might just be my German pessimism. 🤣
2021-11-14T19:13:55Z	(#ptpfohq) @<xuu https://txt.sour.is/user/xuu/twtxt.txt>

> keys.openpgp.org is a descent key server. They only publish a key the at has a valid email.

Mhh, meh. My key currently has no e-mail address attached to it. 😕 I’d like to avoid this. It just makes this address a target for spam, so I’ll have to block it, meaning people can’t reach me anyway …
2021-11-14T20:11:38Z	(#vgyu4qa) … oh my god. I completely forgot what a mess all this was. The key servers as I knew them (SKS) are long dead. I read about this in the news a while back, but since I was hardly using GPG, my brain didn’t decide to store this in long-term memory. 🤦 🤦 🤦 So yeah, I’ll have to read up on how to properly use keys.openpgp.org now. https://keys.openpgp.org/about/news#2019-06-12-launch
2021-11-15T13:10:49Z	(#ptpfohq) @<lyse https://lyse.isobeef.org/twtxt.txt> That sounds weird. 🤔 To be fair, I didn’t dig through all those details. Actually, I stopped reading at “fully automated on Github Actions and fully keyless”, because I figured it’s irrelevant to me anyway.
2021-11-15T13:16:37Z	(#q3z6xkq) @<xuu https://txt.sour.is/user/xuu/twtxt.txt> Yeah, detached signatures are probably the way to go *if* we even implement this.

What I don’t like about detached signatures: They require new metadata fields (URL to sig, method (PGP, signify, …)) and more HTTP requests. Yes, that’s cleaner, but also more involved. (And there’s a slight race condition: Updating the feed and the sig is not an atomic step.)
2021-11-15T13:18:36Z	(#e3i53hq) @<fastidious https://arrakis.netbros.com/user/fastidious/twtxt.txt> $insert_laugh_track_here 🤣
2021-11-15T13:19:08Z	(#5hg654q) @<prologic https://twtxt.net/user/prologic/twtxt.txt> Heh, so I broke everything? 🥴 I can turn this off again. It was just an experiment.
2021-11-15T13:45:15Z	(#e7g2pfa) @<meff https://yarn.meff.me/user/meff/twtxt.txt> Hmm, what do you mean? You fetch my feed and then you want to know *how* to verify the signature (PGP, signify, …)? The most naive approach would probably be this: Run `gpg --verify` on it and, if it fails, try `signify -C`, and so on. Yes, this is not very clean.

Also, no idea about encryption. All I was thinking about was a signed feed. Because … I was playing with GPG that day, I guess. 🤣 I’m not really sure if signed feeds are worth using. I mean, if you fetch via HTTPS, you can already be sure-ish that it comes from my server, can’t you?
2021-11-15T13:48:22Z	(#lrmnw3a) Here’s another thing to consider:

Yarn.social has extended twtxt *a lot* in the past, but all these extensions were backward compatible. Yes, manually constructing threads is awkward, but not impossible. *Encrypted feeds*, on the other hand, are a whole new level. They are so profoundly incompatible with the original twtxt that I’m beginning to wonder: Maybe it’d be worth starting from scratch here? And in the process, solve some of the issues?

Dunno, just a thought.
2021-11-15T13:51:14Z	(#e7g2pfa) … then again, it *could* be implemented in a clean way if we introduced new metadata fields and detached signatures. Then jenny’s hypothetical verification handler could inspect the feed, find metadata fields that point to the signatures (and they indicate if it’s PGP or whatever), fetch that signature, then verify everything. 🤔
2021-11-15T14:10:26Z	Started watching “Another Life” on Netflix. What a *weird* show. It’s like it was written by a 12 year old boy. Utterly stupid mistakes all the time (“hey, let’s visit an unknown planet without protective suits – twice”) and lots of barely clothed women. 🤦
2021-11-15T14:47:48Z	(#4xckrda) @<fastidious https://arrakis.netbros.com/user/fastidious/twtxt.txt> Lol, crap 🤣 I don’t remember any from the top of my head … but I didn’t really pay attention to that 🥴
2021-11-15T14:48:59Z	@<prologic https://twtxt.net/user/prologic/twtxt.txt> I think this is what you/we were looking for the other day: https://wiki.gnupg.org/WKD Appears to allow you to host your GPG key on your own server at a well-known location. I haven’t set it up yet, though.
2021-11-15T16:33:04Z	(#syahp7a) @<xuu https://txt.sour.is/user/xuu/twtxt.txt> Hmm, shouldn’t I be able to do this? `gpg --locate-keys me@sour.is` spits out:

    gpg: error retrieving 'me@sour.is' via WKD: No data
    gpg: error reading key: No data

What am I missing? 🤔
2021-11-15T16:59:28Z	(#syahp7a) @<fastidious https://arrakis.netbros.com/user/fastidious/twtxt.txt> If I understood WKD correctly, then the command I posted should do an HTTPS request directly to sour.is, bypassing any key servers. 🤔 At least it works like that with the little test setup I just made (see https://uninformativ.de/contact.html). I see some "GET /.well-known/openpgpkey/..." pop up in my access log.
2021-11-15T18:21:32Z	(#p2s7oga) @<lyse https://lyse.isobeef.org/twtxt.txt> Huh, whoops. 🥴 I guess it really is necessary to make twt parser very forgiving … (Maybe I should even allow spaces instead of a tab? Some feeds use that. Strictly speaking, they’re broken.)
2021-11-15T18:27:13Z	(#p2s7oga) FWIW, I think she just copies her Mastodon feed over to twtxt: https://occult.institute/@maya This explains *some* empty twts, where the original actually contained an image (see inktober).
2021-11-15T19:42:13Z	Last change before the next jenny release (and then I’ll go for Feed Archival): Finally remove the conversation tag from `Subject:` lines https://www.uninformativ.de/git/jenny/commit/faa3046c007e48f279fc0bb1beb4806a41947da8.html
2021-11-15T19:46:28Z	(#j43fuia) Funny, I’m so used to seeing these tags, it’s almost a bit confusing now that they’re “missing”. 🤣
2021-11-15T20:11:07Z	(#j43fuia) @<fastidious https://arrakis.netbros.com/user/fastidious/twtxt.txt> Yeah, I might re-think that in the future. 🤔 It’s so much nicer to read without all that markup.

I kept the original twt in the mail body, because I wanted to actually see that original twt. (And in early versions, this was very important, so I could manually copy conversation hash tags and/or mentions. 😅) I don’t just want to see “@lyse”, I want to see the URL as well. On the other hand, I don’t need this *all the time*. Hm.

Yeah, dunno. I’m not 100% happy with this either. I’ll put it on the TODO list. 👌
2021-11-16T12:53:52Z	(#syahp7a) @<xuu https://txt.sour.is/user/xuu/twtxt.txt> Yep, it’s working now 👍
2021-11-16T12:56:05Z	(#nzmpuzq) @<prologic https://twtxt.net/user/prologic/twtxt.txt> Lol 😅 Well, I’ll post a pic of my coffee tomorrow morning 😆
2021-11-16T13:18:48Z	(#nzmpuzq) @<prologic https://twtxt.net/user/prologic/twtxt.txt> Does that work? I mean, since it’s an external image, not from your pod … 🤔 Let’s try.

![](https://movq.de/v/a8e7f0ef18/s.png)
2021-11-16T13:21:08Z	(#lalatfq) @<lyse https://lyse.isobeef.org/twtxt.txt> Just go for 12, one per month. 😁

What kind of calendar is it going to be? One of those tear-off calendars?
2021-11-16T13:39:12Z	Spoiler: I put a webcam in one of my windows. It’s taking one picture every day, starting in September. Later this winter, when all the trees have dropped their leaves, I’ll post a GIF. 🍂
2021-11-16T13:39:45Z	(#mjfd2uq) (That is, assuming the Raspberry Pi which takes those pictures doesn’t die first.)
2021-11-16T14:38:39Z	Bwahaha https://drewdevault.com/2021/11/16/Cash-for-leftpad.html
2021-11-16T14:39:36Z	(#mjfd2uq) @<prologic https://twtxt.net/user/prologic/twtxt.txt> Heh, well, here’s a thing: That’s going to be a *big* GIF. I’m thinking about making it a WebM instead for better compression. Can that be inlined as well? 🥴
2021-11-16T15:02:31Z	(#mjfd2uq) @<thecanine https://twtxt.net/user/thecanine/twtxt.txt> Hmm, dunno. Why not, I’ll have a look at it. 🤔
2021-11-16T16:29:58Z	(#d42nbeq) @<lyse https://lyse.isobeef.org/twtxt.txt> Why is that? 😁
2021-11-16T16:31:26Z	(#4tb3nha) @<lyse https://lyse.isobeef.org/twtxt.txt> Maybe npm will self-destruct, who knows.
2021-11-16T16:31:47Z	Sooo, I got a short-scale bass the other day.

![](https://movq.de/v/aafde5f451/s.jpg)
2021-11-16T16:35:16Z	(#5qwmb3q) To be completely honest: Compared to my “regular” bass, it feels a bit like a toy. 🤣 It’s *tiny* (you can’t really tell that from the photo) and *super light*. But since it’s so small, it’s *much* easier to play, especially with my little Trump-like hands. Awesome!
2021-11-16T16:38:50Z	(#5qwmb3q) Despite its size and weight, it’s got a lot of *oomph*:

https://movq.de/v/aafde5f451/badlands.flac

(Lower string tension = more bottom end, apparently?)
2021-11-16T16:40:30Z	(#5qwmb3q) Anyway, this little guy allows me to focus more on music theory now, because I can just play most of the stuff, without having to worry *too much* about not being able reach certain positions.

I love it!
2021-11-16T17:51:17Z	(#5qwmb3q) @<fastidious https://arrakis.netbros.com/user/fastidious/twtxt.txt> Is it too late to start now? I mean, I know the feeling – I could have started like 20-25 years earlier. I probably won’t be able to make up for that time. It’s lost. So, lowering my expectations now. 😁 I’ll never play in a band or anything, but that doesn’t matter, it’s just for me, and a lot of fun. 😊
2021-11-16T17:52:41Z	(#5qwmb3q) @<lyse https://lyse.isobeef.org/twtxt.txt> Yeah, I have to say I like the matte finish much better than the shiny one on the full-scale bass. Shiny stuff just catches fingerprints … 😄
2021-11-16T17:53:17Z	(#ximh6sa) @<fastidious https://arrakis.netbros.com/user/fastidious/twtxt.txt> That must have been pre-Corona, eh? 😁
2021-11-16T18:04:10Z	(#3jlu4sa) @<fastidious https://arrakis.netbros.com/user/fastidious/twtxt.txt> The best sites are those that silently truncate passwords … 🤣 Mine are usually 64 characters long and that’s too much for some pages …
2021-11-16T18:11:39Z	(#j7pnjra) @<lyse https://lyse.isobeef.org/twtxt.txt> 🍿
2021-11-16T19:15:23Z	(#qkq4vva) @<lyse https://lyse.isobeef.org/twtxt.txt> Oh for fork’s sake!
2021-11-16T19:47:06Z	(#xskayra) @<fastidious https://arrakis.netbros.com/user/fastidious/twtxt.txt> Wait. How does this work? You request a plate that says “tux rocks” or something, and in order to get that you have to explain the meaning of it to a person at the DMV? And then what, can they say “no, *that* is not acceptable!”? Like … it all hinges on *your explanation*? 🤨
2021-11-17T15:16:41Z	(#efpt37q) @<prologic https://twtxt.net/user/prologic/twtxt.txt> @<fastidious https://twtxt.net/user/fastidious/twtxt.txt> Yup, exactly. Shorter neck, easier to play, but it can still play the same notes as a full-scale bass.
2021-11-17T15:19:35Z	(#3jlu4sa) @<lyse https://lyse.isobeef.org/twtxt.txt> I can think of one use case for “password truncation”: Avoid DoS. There should be *some* limit to prevent you from using a base64 encoded 5 GB video file as a password. (Of course, that limit could be something like 1024 bytes instead of 20 bytes …)
2021-11-17T15:34:23Z	(#cmwe37q) Damn! I forgot to post my coffee. (I start my day on my work laptop where I don’t have access to twtxt, that’s probably why. 🤣)
2021-11-17T15:37:30Z	(#d2oaqgq) @<lyse https://lyse.isobeef.org/twtxt.txt> “Chlorophilia” reminds me of this old gem: https://www.youtube.com/watch?v=_bvfbK-3PJ8 Although I have a feeling that you’re too young to remember this.
2021-11-17T15:37:55Z	(#o4yc7va) @<jlj https://twt.nfld.uk/user/jlj/twtxt.txt> Lovely! 👌
2021-11-17T15:42:59Z	160 new twts, I can hardly keep up with this. 🤪
2021-11-17T19:58:21Z	(#d2oaqgq) @<lyse https://lyse.isobeef.org/twtxt.txt> Bwaha, yeah, it was super trippy to watch this as a kid. 🤣
2021-11-17T20:01:06Z	(#5qwmb3q) Lol, well it’s better to start at a young age, huh? 🤣 🤯 🥴  https://www.youtube.com/watch?v=PmNyNJFCugo  (That’s also a short-scale bass, so it gives you a rough idea how “big” it is.)
2021-11-18T05:51:57Z	Good morning! ☕ ![](https://movq.de/v/0b85cba6b5/c.jpg) #CoffeeAsPromised