Bluesky Dev
Community discussion of the AT Protocol and Bluesky. (This room is not officially affiliated with the Bluesky team.)
Previous group of messages
  1. moved to @shreyan:beeper.com@shreyanjain:matrix.org
    oh snarfed since you're here, how's lexrpc development going?
  2. snarfed
    moved to @shreyan:beeper.com (@shreyanjain:matrix.org): haven't worked on it in a while. it's close to full lexicon schema validation, but not quite there yet, that'd be nice to finish up eventually
  3. moved to @shreyan:beeper.com@shreyanjain:matrix.org
    cool!
  4. i just got likes and reposts into bskyrb 🙂
  5. snarfed
    nice!
  6. skeetsorg
    When a user is logged into a PDS with its own Bsky app instance, the Bsky instance serves user profiles. If I click on a remote/federated user's profile, do I get to see all their past posts, or just recent ones/posts made after the Bsky instance subscribed to the remote user's PDS? I think Mastodon/ActivityPub only shows the posts made after the instance subscribed, and for older posts you need to visit the remote user's original instance. Would Bsky do the same, or would the BGS graph/indexer solve this problem?
  7. moved to @shreyan:beeper.com@shreyanjain:matrix.org
    i think bgs would solve this for bluesky
  8. with bluesky you ideally will never have to worry about separation of instances
  9. nikolay
    From the context, it was quite obvious that my post was not addressing people outside of the Bluesky team.
  10. skeetsorg
    Does the Bsky app need to be modified to serve directly from BGS/indexer, then? Right now it takes ~1m to import a user's full history into Bsky in my testing; I guess in the future it'd just read directly from BGS instead?
  11. moved to @shreyan:beeper.com@shreyanjain:matrix.org

    In reply to this message

    the reason they are limiting invites right now is because it is a beta. things are not ready for the general public yet. they need a smaller subset of people to test and give feedback before releasing it to the general public
  12. @m_a_s_a_n_a:matrix.org joined the room
  13. @m_a_s_a_n_a:matrix.org removed their display name (OT)
  14. @m_a_s_a_n_a:matrix.org left the room
  15. @caramelpen:matrix.org joined the room
  16. @caramelpen:matrix.org set a profile picture
  17. nedwos joined the room
  18. nikolay

    In reply to this message

    Many projects launch betas on beta servers. Once they are ready for prime time, they launch on official servers. Anyway, I know what a close/private beta is pretty well - I've been closely monitoring the startup scene and participating in it since '99. I'm saying they've repeated a common mistake. It's too late to fix it, I know, but once way to fix this is either stop sending invites, or open up to everybody.
  19. salrides joined the room
  20. @dead10ck:dead10ck.com

    In reply to this message

    I can only guess, as I have very limited knowledge of the architecture so far, but I would imagine it could work like a cache: clients ask the PDS for profile data, PDS says "I don't have it yet", proxies or redirects the request to BGS so the client can get an answer quickly, and then asynchronously syncs the data to its own store.
  21. I'm not sure what was in the profile data you were testing, but 1m sounds not great, esp if it's just mostly short text
  22. TabAtkins joined the room
  23. nzheretic
    Is anyone doing work on proxy servers for aproto communications? E.G. For enterprise/government work use of such communications may require keeping a copy of posts and/or hosting behind firewall/vpn etc?
  24. In reply to this message

    This may require an "Enterprise" level client or even hosting an "internal" server.
  25. Compy (@compy:hazenet.org) joined the room
  26. snek joined the room
  27. snek
    is there a way to see who invited someone, or at least who invited yourself? i guess this isn't part of atproto itself and is just a detail of the first party bluesky instance
  28. moved to @shreyan:beeper.com@shreyanjain:matrix.org
    nah not at the moment
  29. burkard joined the room
  30. Aaron Goldman

    In reply to this message

    Assuming your PDS pulled there repo you would see all content in the repo. So past posts that have not been removed. It would be considerably more expensive to have it pull all past commits for a full history
  31. jsmitty12 joined the room
  32. jsmitty12
    Apologies if this is the wrong venue, does anyone have an invite code they would be willing to share?
  33. moved to @shreyan:beeper.com@shreyanjain:matrix.org
    Aaron Goldman, bnewbold, Daniel Holmgren : do any of you know why .
    bsky.social
    subdomains don't have/need(?) the _atproto txt record?
  34. snek
    anyone know a way to get the "firehose" via the api client?
  35. Takanori Kikuchi joined the room
  36. bnewbold

    In reply to this message

    HTTP XRPC is a fallback for handle lookups, and what we are doing currently for *.
    bsky.social
    . may switch to DNS butt would need to run or own DNS server to handle tons of records
  37. alacer joined the room
  38. In reply to this message

    you can set up powerdns with a custom backend over unix domain socket, basically lets you use language of your choice as a dns server except powerdns does the hard work for you
  39. @10allday:matrix.org
    Is jack dorsey active on Bluesky?
  40. George Antoniadis

    In reply to this message

    Thankfully no
  41. @nazir:matrix.org left the room
  42. jaberwocky joined the room
  43. alrozzi
    Please share an invite code. Please
  44. @planetoryd:matrix.org

    In reply to this message

    no
  45. alrozzi

    In reply to this message

    Why?
  46. alrozzi
    panji.bsky.social: can you please an invite.
  47. @ishan:m.ishanjain.me left the room
  48. @dead10ck:dead10ck.com

    In reply to this message

    Oh interesting. What is it doing now then? Custom DB? And why would you need to run your own nameservers?
  49. Anoop Bhatia left the room
  50. George Antoniadis

    In reply to this message

    My guess would be the limits that some providers impose on number of records per zone/domain, IIRC GCP and Cloudflare for example used to have "relatively" low limits. Not sure about others.
  51. @shellsharks:matrix.org joined the room
  52. @dead10ck:dead10ck.com

    In reply to this message

    Oh wow you're right, Cloudflare is 3500 and GCP is 10k. That seems insanely low for how small a record is
  53. GCP at least says you can contact them to set up higher limits, but even for a default that's very low
  54. George Antoniadis

    In reply to this message

    Cloudflare are a bit more generous with limits if you throw money their way, GCP are really not.
  55. @dead10ck:dead10ck.com
    Yeah I guess it all comes down to 💵
  56. Is there a plan to migrate to DNS for everything in the future though? For a MVP it makes sense to have it as a fallback because of this, but with the protocol's focus on domain names as user handles, it seems appropriate for that to be a longer term goal
  57. vintner joined the room
  58. George Antoniadis

    In reply to this message

    There are a bunch of issues around txt records that have come up in discussions for other protocols, main of which being the size of the each txt record itself. Right now the did plc is a cut down hash of the did document.

    This means there needs to be a registry from which to get the document itself as well as which providers the identity is using etc.

    All that is way too much for a txt record and thus nostr, matrix, and others ended up using well known http endpoints for server or user identity.

  59. @dead10ck:dead10ck.com

    In reply to this message

    Oh right, but the DID document record is separate from the DNS mapping to DID right? If I'm reading

    right, the DNS record just needs to be something like did:plc:12345 right?

    But from there, yeah, a different registry could handle the DID document store

  60. I was thinking DNS could just be used for the domain name → DID mappings. It would serve the purpose of verifying ownership of the domain
  61. Without actually having DNS involved, anyone could make up any domain name as a handle on their PDS right?
    (edited)
  62. Adam Moore joined the room
  63. Jonathan Catchpowle joined the room
  64. @sarangbaloch:matrix.org
    need a bluesky invite code to test the app! kindly someone help….
  65. damon/
    There are no more
  66. @planetoryd:matrix.org

    In reply to this message

    you will never get an invite code
  67. giteauser
    guess we have to run our own 😎
  68. they really ran out of codes?
  69. @sarangbaloch:matrix.org
    how?
  70. we can?
  71. giteauser
    well, i was trying to...
  72. Aaron Goldman
    Scale is hard, to many invites to fast would bring the server to a crawl. Hopefully we get lots of PDS operators
  73. giteauser
    then i was banging my head against a wall until i realised i didn't read this properly
  74. @sarangbaloch:matrix.org
    idts
  75. so now?
  76. how can we get?
  77. any idea
  78. giteauser
    ¯\_(ツ)_/¯ i'm too dumb to know
  79. damon/

    In reply to this message

    I’m hoping so as well
  80. goeo_ joined the room
  81. goeo_
    hey all, i have a question: https://atproto.com/guides/identity says that handles are resolved with the com.atproto.identity.resolveHandle xrpc method, on the handle, and gives a pseudo typescript example. now of course, this seems to work with *.
    bsky.social
    handles. however, i've noticed that for custom domain handles on
    bsky.social
    , users are only asked to set a TXT record, and the described way of handle resolution will not work. thinking the documentation may be outdated, i checked if
    bsky.social
    also sets the TXT record for every user, but this doesn't seem to be the case. are we supposed to just try both?
  82. moved to @shreyan:beeper.com@shreyanjain:matrix.org
    i'm not sure how, but the com.atproto.identity.resolveHandle does support custom domain handles. maybe it does a dns lookup or something?
  83. goeo_
    i think com.atproto.identity.resolveHandle on
    bsky.social
    just works because everyone's on
    bsky.social
    so far - i don't think it could resolve a different pds
  84. or at least i wouldn't expect it to
  85. moved to @shreyan:beeper.com@shreyanjain:matrix.org
    i guess if full repos are synced it might?
  86. xnf0k
    The server resolves with DNS if it's not a domain that's supported by itself (like *.
    bsky.social
    on the official instance)
  87. goeo_
    so i guess we are supposed to try both?
  88. moved to @shreyan:beeper.com@shreyanjain:matrix.org
    resolveHandle by itself works I think
  89. goeo_
    well no, most people with custom domains on
    bsky.social
    don't have xrpc running on their domain
  90. moved to @shreyan:beeper.com@shreyanjain:matrix.org
    try a quick curl https://bsky.social/xrpc/com.atproto.identity.resolveHandle?handle=shreyanjain.net
  91. goeo_
    yeah but the docs say you're supposed to do curl https://shreyanjain.net/xrpc/com.atproto.identity.resolveHandle?handle=shreyanjain.net and that's not gonna work
    (edited)
  92. moved to @shreyan:beeper.com@shreyanjain:matrix.org
    but ya this probably requires that your account is on
    bsky.social
    even if the domain isnt
  93. In reply to this message

    oh true
  94. wait, it only says that for self-hosted, in which case i think it would be accurate?
  95. goeo_
    well is an at protocol implementation supposed to know about bluesky?
  96. i don't think so
  97. it's just another server
  98. moved to @shreyan:beeper.com@shreyanjain:matrix.org
    although it probably doesn't make much practical sense to put your domain username and pds as the same thing
  99. actually sorry
  100. there's probably no problem with that
  101. xnf0k
    I think the example in the docs aren't right
  102. moved to @shreyan:beeper.com@shreyanjain:matrix.org
    yeah they haven't been updated for a long time
  103. goeo_
    well i mean for
    bsky.social
    they're exactly right
  104. for any other domain, they're not
  105. and that's the protocol weirdness
  106. xnf0k
    Also if you see all examples (including
    bsky.social
    ), none of them provide the handle argument anywhere. xrpc.service(pds) is just the server. So all examples need updating I think.
  107. goeo_
    hmm? they do provide the handle argument
  108. xnf0k

    I mean the examples under:

    • Example: Hosting service
    • Example: Hosting service w/separate domain name
    • Example: Self-hosted
  109. lorentzo joined the room
  110. isiah joined the room
  111. Aaron Goldman

    In reply to this message

    Block and Mute are worth being two distinct concepts.

    Block is public, Mute should not let the muted know

  112. Ted Han (than@mozilla.social) joined the room
  113. moved to @shreyan:beeper.com@shreyanjain:matrix.org

    In reply to this message

    yeah this example looks wrong.
  114. Kjartan joined the room
  115. @dead10ck:dead10ck.com

    I did notice in the

    that it does a TXT record lookup on _atproto.{handle}

    I'm a bit confused by this though, since earlier in this thread today, someone said that the DNS lookup is a fallback. But the code doesn't look like it does anything else. Not familiar with the whole code stack or when this code is actually called though

    (edited)
  116. Kjartan
    First: Nope, I don't have any invite codes (but continue reading if you wanted one). But as many want to try bsky, to build clients for it or whatever, and just don't get in, I'm thinking about opening an instance to the public. Because, sad truth is: if you didn't get in by now, there is little hope you won't make it in before it goes public in general. On the instance I'm thinking about opening, zou would be able to communicate with the others on that instance, but not with those who are on the official server (at least not yet - might change once federation is there - but likely you want to move then anyway to a different server (the way the protocol is built, you should be able to take everything with you)). The question is: is there even demand for that (or do all those pretend to need a code for development, actually just want to get into the main instance?) Just react with some emoji to this text, so I can check later ot tomorrow, how many would be interested. I have some capacity on my server available, but likely I would have to spin up a new machine, and I have no idea whether I could handle the demand or if there is little to no demand at all. That's it.
  117. George Antoniadis

    In reply to this message

    there is little hope you won't make it in before it goes public in general

    Not sure what this is based on. Everyone invited gets invites every week, at some point a friend of a friend will get an invite, and soon after that you'll get one too.

    (edited)
  118. That been said, the more servers the better :D There are already couple of servers online with thousands of users each.
  119. Kjartan
    Absolutely correct. Although it's 1 invite every 2 weeks. So "eventually" might still mean "never before it gets opened to the public". for many if not most. But what I meant that moment was: if you didn't make it in yet VIA THE WAITLIST don't hope you will ever get into it via the waitlist
  120. And I fully understand it, if one wants to get their hands on some client or whatever - access to a server is basically a must. The documentation of the protocol is very thin, and even if it were better, it's still a different expereince to actually be able to try things. If I can help those, I'm happy to help
  121. George Antoniadis
    retr0id do I even wanna know what skeetstudio is? :P
  122. retr0id
    it lets you use markdown link syntax
  123. and tag people similarly
  124. [example.com](https://www.youtube.com/watch?v=dQw4w9WgXcQ) and all that good stuff
  125. martinhn joined the room
  126. George Antoniadis
    actually sounds cool
  127. retr0id
    you can also have zero-length labels to pull off @everyone-like tricks
    (edited)
  128. George Antoniadis
    still not logging in, but sounds cool :D
  129. retr0id
    [](@retr0.id)
  130. I guess I'll put it on github at some point, it's just index.html and main.js, will work offline if you download those two files
  131. moved to @shreyan:beeper.com@shreyanjain:matrix.org
    wait whats the link to skeetstudio
  132. George Antoniadis
    that would be appreciated :D
  133. moved to @shreyan:beeper.com@shreyanjain:matrix.org
    awesome
  134. @10allday:matrix.org
    Is it a gimmick to be able to transfer your data from one server to another?
  135. retr0id
    huh?
  136. George Antoniadis

    In reply to this message

    Depends how often you get banned or how often you wanna move to a place with nicer people.
    If I had the option to move my twitter/mastodon acc to a different without losing the ability to post/read my friend's stuff I'd do that on a hearbeat.
  137. Two of the mastodon servers I was in folded due to server costs and mods not being able to handle the pressure of their servers getting popular, and had to figure out where all the pees I was following went to, make sure it's really them, that it's their primary server etc.
  138. snarfed
    George Antoniadis: just curious, did you not migrate your account and export/import your posts before they closed? https://docs.joinmastodon.org/user/moving/#migration
  139. Mastodon account migration is nonstandard and awkward, but it's at least something
  140. George Antoniadis
    snarfed I did, but the main issue was not my posts, I'm mostly lurking, it's losing my connections.
  141. moved to @shreyan:beeper.com@shreyanjain:matrix.org
    yeah i agree
  142. the connections are an issue
  143. snarfed

    In reply to this message

    this is just for non-default (ie non- .
    bsky.social
    ) handles
  144. George Antoniadis
    Same thing with. I'm using my own domain cause so that I can jump from provider to provider without being tied to google/microsoft/fastmail/whatever.
  145. snarfed
    still confused though, Mastodon's account migration migrates your followers
  146. George Antoniadis

    In reply to this message

    Not really, cause that's your handle, your identity comes with you (did plc)
  147. When you move to a new PDS, it just creates a new handle.
  148. But anyone who is following you will just see the new handle, cause the underlying DID is the same.
  149. moved to @shreyan:beeper.com@shreyanjain:matrix.org
    does it? handles are disconnected from pds afaik
  150. snarfed
    right. maybe we're talking about different things. @dead10ck:dead10ck.com was asking about a DNS lookup in the code, not how ATP account migration works
  151. @10allday:matrix.org
    Can the account transfer functionality be used in conjunction with game distribution platforms?
    (edited)
  152. snarfed
    (and again, the DNS-based TXT lookup in the current PDS isn't part of the currently documented protocol)
  153. George Antoniadis
    correct, but the DNS part is just validation for the domain. let's say that you are following me on
    geoah.me
    when you do that, the server does a DNS query to the domain, and gets my underlying DID it then uses my DID of everything on the protocol. and only uses the handle for displaying stuff on the ui.
  154. @10allday:matrix.org

    In reply to this message

    I don't know why we are just building a protocol just to clone twitter, and have someone create a separate protocol just as a steam clone.
  155. moved to @shreyan:beeper.com@shreyanjain:matrix.org

    In reply to this message

    alr, let me try to understand this:

    are all .

    subdomains pointing to a wildcard to the
    bsky.social
    pds, which using http xrpc actually does use this lookup method?

  156. George Antoniadis
  157. moved to @shreyan:beeper.com@shreyanjain:matrix.org
    i'm probably not putting this into words well but that's the idea I'm getting
  158. George Antoniadis
    but even that, just resolves to a DID
  159. moved to @shreyan:beeper.com@shreyanjain:matrix.org
    yeah, ultimately the did is the actual account/repo
  160. snarfed

    In reply to this message

    right. and afaik that DNS lookup doesn't even happen on follow. it happens once (and/or periodically), in the background, by the PDS, to confirm the domain/handle binding
  161. regardless yeah I think we're all agreeing
  162. George Antoniadis
    there is a missing part in all of this that we don't really understand, and that is the plc directory.
  163. moved to @shreyan:beeper.com@shreyanjain:matrix.org
    plc directory i think is where all the DID documents are stored for now
  164. so whenever you see a did:plc:xxxxxxx you can find its did document at https://plc.directory/did:plc:xxxxxxx
  165. snarfed
    the interesting part will be if/when did:web's start showing up, which are an alternate way of using URLs/domains. the main PDS doesn't use them yet afaik, but they say they'll be supported. https://atproto.com/specs/atp#dids , https://atproto.com/guides/identity#example-self-hosted
  166. moved to @shreyan:beeper.com@shreyanjain:matrix.org
    then that has an alsoKnownAs which should contain your PDS host and domain username
  167. In reply to this message

    definitely! i don't understand did:web as well as i understand did:plc though. is there any simple explanation out there?
  168. @10allday:matrix.org
    How are we going to communicate with other protocols, if they don't interact using lexicons?
  169. @danilo:chat.ex06.nohost.me joined the room
  170. George Antoniadis

    In reply to this message

    I think my confusion is that last time I checked plc directory didn't tell you which pds to talk to about this -- docs seem updated though, might need to check again.
  171. snarfed

    In reply to this message

    yup, they're just URLs. after the leading did:web:, replace all :s with /s. then resolution is looking up a JSON doc at a well-known sub-path under that URL. https://w3c-ccg.github.io/did-method-web/#read-resolve
  172. George Antoniadis

    In reply to this message

    The hosting service: https://alice.com

    mmm that's nice

  173. In reply to this message

    what do you mean by "don't interact using lexicons"?
  174. moved to @shreyan:beeper.com@shreyanjain:matrix.org

    In reply to this message

    oh that’s pretty neat 🙂
  175. @10allday:matrix.org

    In reply to this message

    If there's a decentralized game marketplace protocol, they'll need to speak lexicons???
  176. George Antoniadis
    I'm really not sure about the need for the plc directory and dns. since everything is http they might as well have gone with well known endpoints and get it over with. this way we could have self hosted our own did documents and where they point and gotten it over with.
  177. moved to @shreyan:beeper.com@shreyanjain:matrix.org
    you mean like webfinger?
  178. George Antoniadis

    In reply to this message

    if they want to leverage other atp features then... yes.
  179. In reply to this message

    webfinger, nostr, matrix, etc
  180. this would remove the need to have a centralized did registry and just use it as more of a witness / backup / convinience thingie
  181. @10allday:matrix.org

    In reply to this message

    It's strange that atp only allows you to interact with users using atp, but there are no standardized ways to talk to other protocols.
  182. moved to @shreyan:beeper.com@shreyanjain:matrix.org

    In reply to this message

    yeah true, but based on what snarfed sent did:web would solve that so
  183. George Antoniadis

    In reply to this message

    correct, I'm still not sure when they started accepting did:web -- I'll shut up and go read the docs.
  184. @10allday:matrix.org
    Why aren't we using oauth2 authentication? Why do we need DID?
  185. Kjartan

    In reply to this message

    Can't say it for sure, but I would say, at least(!) 10 days ago
  186. George Antoniadis

    In reply to this message

    which of the hundreds of thousands of protocols should they support?
    maybe the top 10k ones?
    and why?
  187. retr0id

    In reply to this message

    wait is did:web supported by bsky right now?
  188. George Antoniadis

    In reply to this message

    I feel these two are a bit different concepts. to the best I can quickly explain this:
    did is just a way to identify a user/service/whatever
    oauth2 is a way to request/provide authorization

    you can use dids for identity and still use oauth2 for authz

  189. moved to @shreyan:beeper.com@shreyanjain:matrix.org

    In reply to this message

    i don't believe it is right now unfortunately
  190. @dead10ck:dead10ck.com
    One might think we need a protocol
    ...
    to talk to all protocols
  191. George Antoniadis

    In reply to this message

    That's the part I'm confused about -- did:web is in the docs FOR REALZZ
  192. moved to @shreyan:beeper.com@shreyanjain:matrix.org

    In reply to this message

    lmao
  193. George Antoniadis
  194. ^ I'm sorry but I had to
  195. @dead10ck:dead10ck.com
    Lol I have this link in the ellipses 🙂
  196. retr0id

    In reply to this message

    taking a brief skim of the source, it does seem to be supported in the current atproto source, but I don't think there's a way to actually like, import a did:web account right now (is import the right word?)
  197. flancian 🍄
    hi all! so excited about the direction of [[bluesky]]! would people be interested in using threading to organize discussion?
  198. George Antoniadis

    In reply to this message

    bahahaha, sorry element's dark mode isn't doing a great job of highlighting it
  199. flancian 🍄
    (if you don't see this response as a thread, only a quote, you might need to upgrade your matrix client.)
  200. George Antoniadis

    In reply to this message

    I have threads enabled but still don't see this as a thread.
  201. In reply to this message

    owh
  202. In reply to this message

    it ALSO shows it in a thread as well as in the main channel
  203. In reply to this message

    the fuck element
  204. In reply to this message

    people keep trying threads but not element clients support it and it gets messy
  205. In reply to this message

    honestly I'm more inclined to get out of channel and stick to the discord
  206. flancian 🍄

    In reply to this message

    https://app.element.io supports them on my device, have you updated recently? if so I could report this bug to element
  207. we cross-post discord and matrix for the Agora project rooms
  208. I would beseech the blueky project to stick to matrix or some other open source network if at all possible :)
  209. I think there are sometimes happy little glitches we can optionally fix ;)
  210. George Antoniadis
    issue is that this channel is just invite-request-spam at this point and most of us are in the touchers discord anyway
  211. moved to @shreyan:beeper.com@shreyanjain:matrix.org
    i think the touchers discord should remain hidden from "outsiders" to prevent invite spam
  212. George Antoniadis
    yap and also they actually have mods :P
  213. and rules
  214. paul/daniel/why are no longer active here so we might as well bail at this point
  215. retr0id do you know what we're talking about?
  216. retr0id

    In reply to this message

    the what discord
  217. moved to @shreyan:beeper.com@shreyanjain:matrix.org
    i do prefer matrix to discord as a protocol but yeah
  218. retr0id
  219. (found it)
  220. George Antoniadis
    flancian 🍄 bopped you on bsky.
  221. @dead10ck:dead10ck.com

    I mean this seems like a moderation problem, rather than a platform problem. We need mods to ban people like @10allday:matrix.org and those asking for invites who just turn the room into garbage.

    I prefer matrix as a protocol too. If all the people actually involved in bsky leave, this room will truly have little reason to exist :/

  222. George Antoniadis
    I don't agree with the banning of @10allday:matrix.org but I do agree with the rest :D
  223. Michelle Tomasko joined the room
  224. @10allday:matrix.org

    In reply to this message

    I will just leave instead. ✌️
  225. @10allday:matrix.org left the room
  226. George Antoniadis

    In reply to this message

    that was a bit on the the harsh side maybe :/
  227. @dead10ck:dead10ck.com

    In reply to this message

    Yeah, so is behavior like this. Which I've seen them do pretty much every day since I joined this room. So 🤷
  228. @dead10ck:dead10ck.com
    Another thing that might help is making a Matrix Space with a few different topical rooms to keep discussions more focused
  229. Kjartan
    Sounds like a good idea imo. It will be also easier to search for/find already answered questions
  230. Actually, I think a forum would be just so much more suitable in general.
  231. Maybe someone can quickly build one - based on the atproto :D
  232. or just some <whatever>BB
  233. flancian 🍄
    so, to be clear, the development team has moved to discord?
  234. oh, I just read another thread -- has the dev team moved to discord altogether?
  235. @dead10ck:dead10ck.com

    I'm an OSS contributor myself to the Helix text editor, and the other day we got a complaint about a user with "Groyper" in their name. And the maintainers just collectively sighed. We just want to work on our text editor, we did not sign up to be chat room moderators.

    But I think one of the consequences of success is that it comes with a whole bunch of stuff you never wanted to do. So I get it

    (edited)
  236. Kjartan
    I kinda get it, too. But I have asked several questions here. Which were either never read, answered or I maybe I have never read the response because I missed it
  237. George Antoniadis

    In reply to this message

    the dev team are way to busy right now, the discord I mentioned is a community thing as well
  238. flancian 🍄

    In reply to this message

    does this mean development is not currently public? where can we follow along/engage with devs? on bluesky proper?
  239. George Antoniadis

    In reply to this message

    mostly bsky, doc and server commits
  240. flancian 🍄

    In reply to this message

    found https://github.com/bluesky-social, thank you; do you know which repository has most of the 'community level' / federation related activity? I'm guessing 'atproto' proper?
  241. added link to https://anagora.org/bluesky, a public knowledge base
  242. mettty joined the room
  243. snek
    is there a way to subscribe to the subset of events relevant to you on the websocket?
  244. George Antoniadis

    In reply to this message

    atproto is the main thing, but federation is not there.
  245. indigo is also nice as it's cleaner to read and understand, but not a full blown pds/client
  246. Kjartan

    In reply to this message

    Yeah. Totally agree. Sadly I have no Go writing skills (/and/or motivation)
  247. @alexl:kde.org joined the room
  248. squishthoughts
    I've been interested in exploring the platform per-se, but that doesn't require an account on the main instance - I'm exactly the use case I think you are describing. And I have enough manners to know not to beg folks here or elsewhere for invites. Let me know if you decide to do this, as I'd be interested in access to a playpen at least. I'm a go dev, and curious about the proto, what's built so far, etc.
  249. Anonymous joined the room
  250. Anonymous
    Can I get an invite code please?
  251. Wahid changed their display name to Linad
  252. OT changed their display name to Anonymous
  253. Anonymous changed their display name to OT
  254. @danilo:chat.ex06.nohost.me
    Message deleted
  255. Anonymous
    They seem to be hard to come by, we still didn't get one.
  256. Was also looking for self-hosting docs if anyone has a guide.
  257. The official docs have zero guides on how to do it.
  258. mikuhl
    Should the cdn server have cors mode?
  259. I just fired up my client in web app to just see if its working, it is except for images don't load.
  260. ryangallagher joined the room
  261. moonwalker358 joined the room
  262. @spencerthayer:matrix.org joined the room
  263. epuerta9 joined the room
  264. @spencerthayer:matrix.org

    In reply to this message

    Is the code for the development sever available for people to run on their own servers?
  265. @spencerthayer:matrix.org
    Sick
  266. snarfed
  267. federation isn't turned on yet though, so for now your server would be its own island
  268. 河越靖代 joined the room
  269. ・ / changed their display name to 河越靖代
  270. retr0id

    In reply to this message

    would an instance like this be able to federate with other instances that don't understand handshake domains?
  271. snarfed
    retr0id: unclear. I vaguely remember someone on the team saying no, maybe, but I don't remember details
  272. snarfed
    finally pulled this out into its own project: https://github.com/snarfed/arroba . Python library that implements PDS and ATP, including repo, MST, and sync methods. still in progress!
    (edited)
  273. xnf0k

    In reply to this message

    Likely no since there would be no way for them to read the TXT and get the did.
    But one future we see is where stems federates with other Handshake resolving servers.
  274. retr0id
    what if they nominally used a "gateway" like
    hns.to
    , at the atprotocol level, so other instances see domain.name.here.hns.to etc. validated thru traditional DNS, but the handshake instances would strip the hns.to internally and do their stuff via handshake
    (edited)
  275. although I suppose that introduces reliance on traditional dns and defeats the whole purpose of handshake
  276. xnf0k
    Not just reliance but a huge security problem. That single server can read&modify every single request and response. That's why there's a big red banner with a warning that
    hns.to
    (and other gateways) are just to demo Handshake to new folks.
  277. @gordonforas:matrix.org joined the room
  278. joshlacal
    Has anyone on here had any luck setting up and configuring a PDS for testing purposes? I’m able to get it running on
    localhost
    and I’ve tried creating a user using the shell referenced in the makefile but i’m unsure how to set a password so I can try to login using the bsky app. i’m able to connect to the PDS, but no idea how to login, or configure the server properly
  279. @spencerthayer:matrix.org

    In reply to this message

    I’ve been trying to find that in the documentation. No luck.
  280. @reiver (CHARLES ILIYA KREMPEAUX) joined the room
  281. @dead10ck:dead10ck.com

    In reply to this message

    So this came up a couple of times over the last few days, and I did a bit more research and learned that

    , so if I'm understanding it right, this does not necessarily get handled by any application layer; it just happens in the DNS layer (naturally), so ATproto would play no part. This means it will be up to the instance operator to configure their DNS to protect their users from identity spoofing.

    I just

    domain through Route 53 and it was actually quite easy. A few clicks through the UI and then copy / pasting some fields in the registrar's web UI and that was it. So luckily this isn't a big hurdle for operators to overcome to protect their users.

  282. @reiver (CHARLES ILIYA KREMPEAUX)
    I'm trying to understand the at-protocol. How do the plc DIDs get resolved? ... For example — My domain is "
    example.com
    ". So I do a DNS TXT lookup on "_
    atproto.example.com
    ". I get back something like: "did=did:plc:abcde12345" How do I then resolve this DID? Do I use WebFinger? And if "yes" (I use WebFinger), which host do I call WebFinger on?
  283. xnf0k

    In reply to this message

    Currently, it is fetched from a central directory: https://plc.directory

    There's regular discussions here to replace this "placeholder"

  284. @gordonforas:matrix.org
    Message deleted by Administrator
  285. @reiver (CHARLES ILIYA KREMPEAUX)

    In reply to this message

    Is the resolution done using WebFinger?

    I tried, and it doesn't seem to work:

    https://plc.directory/.well-known/webfinger?resource=did:plc:m2jwplpernhxkzbo4ev5ljwj

    I get this error:

    "Cannot GET /.well-known/webfinger"

    Did I do something wrong?

  286. moved to @shreyan:beeper.com@shreyanjain:matrix.org
    nope
  287. it's not webfinger. instead it used DID documents
  288. so when you get back did=did:plc:abcde12345 you can go to https://plc.directory/did:plc:abcde12345
  289. @reiver (CHARLES ILIYA KREMPEAUX)

    In reply to this message

    Awesome! Thank you.
  290. ichigoro joined the room
  291. Frankely Diaz joined the room
  292. grabb0id joined the room
  293. laford_public joined the room
  294. laford_public
    Hello! Just popping in to check out what's happening.
    (edited)
  295. himazindes joined the room
  296. yazankittaneh joined the room
  297. @turing_k:matrix.org joined the room
  298. 八雲 joined the room
  299. thepudds joined the room
  300. Kjartan

    In reply to this message

    I didn't try the shell. But if you start the pds (and set temporarily that no invites are required) you can simply create an account via client (like the web client or the mobile apps).
  301. @dsenjoyer:matrix.org

    In reply to this message

    I see. Many thanks for all your answers!
  302. arnobchak
    Notice: Indian Government banned Element message app Because it promotes terrorism and national security concerns. Any alternative way
  303. John Ngugi joined the room
  304. lindort joined the room
  305. 2j86 joined the room
  306. Max (@mf:dendrite.matrix.org) joined the room
  307. @planetoryd:matrix.org

    In reply to this message

    typical fascism
  308. for the children and national security
  309. In reply to this message

    the spam is getting specific
  310. jimberkman joined the room
  311. jimberkman
    Just joined
  312. djamesp joined the room
  313. @dsenjoyer:matrix.org
    Message deleted
  314. Mateus Nunes joined the room
  315. alacer
    Dear(s), is there any guide for SelfHosting a network Node
  316. Melchior joined the room
  317. @planetoryd:matrix.org

    In reply to this message

    get an invite code
  318. @interpol:cutefunny.art joined the room
  319. Mark Foster SSI: @mfoster.io
    01001001 01110011 00100000 01100101 01110110 01100101 01110010 01111001 01101111 01101110 01100101 00100000 01110011 01101011 01100101 01100101 01110100 01101001 01101110 01100111 00100000 01101001 01101110 00100000 01101000 01100101 01110010 01100101 00111111
  320. sylphrenetic

    In reply to this message

    search for another Matrix client, maybe a small one that they don't know about
  321. Mark Foster SSI: @mfoster.io

    In reply to this message

    Wow
  322. There are bridges also
  323. @danilo:chat.ex06.nohost.me left the room
  324. Mark Foster SSI: @mfoster.io

    In reply to this message

    I have been thinking a lot lately about a Human Readable UI name space utilizing DIDs that broad casts through a multicast session in a scatter attempt to connect beyond traditional DNS

    https://docs.google.com/presentation/d/1-N-qZs4eEup70ZcGsxPf-SY6BbL0qn-UebWNMuCi5MU/

  325. klinark joined the room
  326. Mark Foster SSI: @mfoster.io

    In reply to this message

    There are also some interesting concepts around low latency networks utilizing Erasure Coding one being Liquid Data Networking an interesting concept for rural networks that could bind to a mesh network of iot devices

    https://youtu.be/DSlL2MuF97E

  327. klinark

    In reply to this message

    Any luck with finding those self-hosting docs/guide? Also looking for an invite code but do not wish to flood this place with invite spam so no worries if this needs to be removed.
  328. @planetoryd:matrix.org

    In reply to this message

    they dont remove. just spam
  329. if you spam enough, you'd get an invite code
  330. klinark
    I don't want to be a bother so no worries
  331. @planetoryd:matrix.org

    In reply to this message

    that's how you will never get an invite code
  332. klinark
    I might check back tomorrow then. Thanks for the advice.
  333. klinark set a profile picture
  334. Mark Foster SSI: @mfoster.io
    Have any of you respawned your Gitter.I’m profile since the move to Matrix?
  335. @dead10ck:dead10ck.com

    In reply to this message

    Hey, how about not encouraging people to spam? 🤔
  336. @platinbae:matrix.org joined the room
  337. km joined the room
  338. Kaan Mutlu changed their display name to km
  339. @planetoryd:matrix.org

    In reply to this message

    how resistant is atproto to username take downs
  340. platinbae changed their display name to Johnny
  341. @platinbae:matrix.org left the room
  342. Aaron Goldman

    In reply to this message

    Well depends on what you mean by "username take downs".
    If I control

    and want to re-assign @
    bob.example.com
    to a new user that is my right as controller of
    example.com

    If someone tries to take the domain from me ICANN has Uniform Domain-Name Dispute-Resolution Policy https://www.icann.org/resources/pages/help/dndr/udrp-en

    That said if I am following you the follow is defined by DID and not by handle so when they get the domain name and re-assign @

    to their new repo all of the followers will still be following the old user by the DID

  343. @planetoryd:matrix.org
    that's nice. so the DNS is not a single point of failure
  344. Aaron Goldman
    If my domain is taken and I am simultaneously kicked off my PDS. I just update the DID Document. keep calm and carry on to a new PDS with a new handle and none of the follow links brake.
  345. notneema
    Question - I signed up for bluesky with an email, but no where in the web/mobile app can I find which email I signed up with. Also looked over the protocol docs and saw no mention of email.
    How can I go about finding out / change my bluesky email lol. And a follow up - if I’m understanding correctly, the protocol itself doesn’t associate any emails with identity? (But if my application wanted to, I could do that on the app-level?)
  346. Aaron Goldman
    AtProto does not care about your email. My guess is that nearly all PDSs will use email as the way to login to the PDS. Email based reset is the standard of practice for account control.
  347. Matthew waves
  348. we’ve been working on how to layer matrix over other transports and protocols
  349. @planetoryd:matrix.org
    or metamask ?
  350. In reply to this message

    how will the full mesh routing be solved
  351. Matthew

    In reply to this message

    eg for MIMI (the IETF messaging interop group))
  352. so if there’d be interest in using matrix semantics for DMs or E2EE or decentralised ACLs for bluesky, lmk :)
  353. In reply to this message

    matrix has never needed full mesh routing; it’s just the default implementation
  354. @planetoryd:matrix.org
    email is a point of centralization here, increasingly. many services whitelist email providers.
    (edited)
  355. snarfed

    In reply to this message

    (this is definitely true right now and in the common case. they technically plan to support did:web's too, which don't survive losing a domain, but they're at least loud and clear about that in the docs. https://atproto.com/guides/identity#example-self-hosted )
  356. vernon (building @ yup)
    I'm interested in building an indexer for bsky server. Is there a good way to index posts, preferably something events driven so can save posts as they're coming in
  357. snarfed

    In reply to this message

    yup, lots of examples in eg https://github.com/bluesky-social/atproto-ecosystem and the Bluesky API Touchers discord
  358. @reiver (CHARLES ILIYA KREMPEAUX)
    How do I get profile information for a user? For example.... Given that I have a DID document such as: ------- BEGIN ------- { "@context": [ "https://www.w3.org/ns/did/v1", "https://w3id.org/security/suites/secp256k1-2019/v1" ], "id": "did:plc:m2jwplpernhxkzbo4ev5ljwj", "alsoKnownAs": [ "at://
    vercel.com
    " ], "verificationMethod": [ { "id": "#atproto", "type": "EcdsaSecp256k1VerificationKey2019", "controller": "did:plc:m2jwplpernhxkzbo4ev5ljwj", "publicKeyMultibase": "zQYEBzXeuTM9UR3rfvNag6L3RNAs5pQZyYPsomTsgQhsxLdEgCrPTLgFna8yqCnxPpNT7DBk6Ym3dgPKNu86vt9GR" } ], "service": [ { "id": "#atproto_pds", "type": "AtprotoPersonalDataServer", "serviceEndpoint": "https://bsky.social" } ] } -------- END -------- How do I get the profile information? My guess is this is relevant: ------- BEGIN -------- "service": [ { "id": "#atproto_pds", "type": "AtprotoPersonalDataServer", "serviceEndpoint": "https://bsky.social" } ] -------- END -------- That maybe I should be using "https://bsky.social" for XRPC requests. I.e., https://bsky.social/xrpc/{methodID} (As mentioned here: https://atproto.com/specs/xrpc ) So, maybe something like: https://bsky.social/xrpc/app.bsky.actor.getProfile https://bsky.social/xrpc/app.bsky.actor.getProfile?actor=at://did:plc:m2jwplpernhxkzbo4ev5ljwj https://bsky.social/xrpc/app.bsky.actor.getProfile?actor=at://vercel.com But all of those give me the same error: {"error":"AuthenticationRequired","message":"Authentication Required"} How do I authenticate? What am I doing wrong? How do I get profile information for a user?
  359. snarfed
    you're on the right track! first call https://bsky.social/xrpc/com.atproto.server.createSession with username and password to get a session token, then pass the token in the Authorization HTTP Header, oauth 2 style
  360. dimikalyn joined the room
  361. Anonymous
    Message deleted by Aaron Goldman
  362. skrek joined the room
  363. skrek
    I've been writing a PDS as of late, using
    staging.bsky.app
    to test against. Recently I've run into an issue where when clicking on my profile page, my server returns an invalid response. No useful output in the site's console, either. the only issue I can think of is that optional values I've set as null instead of ignoring them, am I right in thinking that that's my problem? or have I missed something somewhere else
  364. Midou (Moved to envs.net, same username. Please invite me to another room if it seems necessary, and if i do happen to be admin somewhere ping me.)
    Are bluesky users able to communicate with activitypub users? Kinda curious as to what bluesky will deliver :3
  365. Aaron Goldman

    In reply to this message

    what do you see in the dev tools network pannal?
  366. Aaron Goldman

    In reply to this message

    https://bsky.social/xrpc/com.atproto.sync.getRepo?did=<did>&earliest=<cid>&latest=<cid>

    did: did.toString(),
    earliest: commit.toString(),
    latest: commit.toString(),

    when you sync the repo for indexing give is the last commit you pulled so it will give you only the new blocks since then

  367. snarfed

    In reply to this message

    exciting! is this open source anywhere yet?
  368. someone is going to forget to say since when and download the whole thing every time and waist the servers bandwidth I don't know a good way to push people to use the sync api politely
  369. Also if your PDS pulls a repo for someone that is hosted on a difrent PDS it is still polite to give the latest commit you have if someone ask your sync API. The PDS in the DID Document is the host of last resort but not necessarily the only PDS that can sync the repo to you. It is all authenticated data.
  370. Kjartan
    If I understand it right - there is (currently?) no way to circumvent the
    plc.directory
    , is there?
  371. Chris Lace
    I’m here to see The Wizard of Bluesky?
  372. Kjartan

    In reply to this message

    Because if there isn't - what's the point? Sounds a bit drastic, but instead of relying on twitter or Facebook or whatever, one relies on the

    instead (maybe even with the downside to take care of hosting the pictures).

    I really hope I'm missing something here

  373. skrek

    In reply to this message

    not quite! it'll be there within the week hopefully, I'd like to finish up base functionality and then refactor for legibility, usually my first pass of any code is less than desirable to look at
  374. Kjartan

    In reply to this message

    What language do you use?
  375. skrek

    In reply to this message

    rust
  376. Kjartan
    Looking forward 👍
  377. skrek
  378. in theory, all that output should be correct? it's very possible i've missed something, however.
  379. AviKav joined the room
  380. salmoon set a profile picture
  381. snarfed

    In reply to this message

    they do support did:web as well as did:plc, https://atproto.com/guides/identity#example-self-hosted , and they've pretty much agreed and said did:plc is hopefully temporary, even if it's a long-lived temporary. I think there's the potential for multiple did:plc registries, Cert Transparency style, but they haven't really documented the registry part of did:plc much yet afaik
  382. In reply to this message

    minor heads up, sounds like the federation design actually won't PDS to PDS after all, it feeds through other relays/aggregators. https://bsky.app/profile/bnewbold.bsky.team/post/3jukvrq3lt226
  383. Kjartan

    In reply to this message

    Thanks. I'm somewhat disappointed. The plan to become decentralized is very different from being decentralized, imo. Something to look forward to, but not getting hyped and excited until it's there 😕
  384. snarfed
    yup, totally fair. they hope to ship (and document and turn on) federation soon, that will be a big step
  385. @dead10ck:dead10ck.com

    In reply to this message

    Could you please quote this for those that don't have access yet? 🙂
  386. @dead10ck:dead10ck.com

    In reply to this message

    In reply to
    Kjartan

    Because if there isn't - what's the point? Sounds a bit drastic, but instead of relying on twitter or Facebook or whatever, one relies on the

    instead (maybe even with the downside to take care of hosting the pictures).

    I really hope I'm missing something here

    I think it's easy to forget this, since there are people actually using it now, but Bluesky is not done. Just because it doesn't exist now doesn't mean it will stay that way forever
    (edited)
  387. aekrylov joined the room
  388. @dead10ck:dead10ck.com

    In reply to this message

    Oh neat, thanks! It doesn't seem like this works though. The "show more replies" button doesn't seem to do anything
  389. snarfed
    yeah that used to work, broke recently I guess. still useful for individual posts though
  390. jg-byte joined the room
  391. @shigop:matrix.org left the room
  392. @app1ep1e:matrix.org joined the room
  393. @app1ep1e:matrix.org set a profile picture
  394. isiah
    Hi! Anyone have an invite to Bluesky? been wanting to contribute to some open source repos but wanted to get a feel for the product first
  395. moved to @shreyan:beeper.com@shreyanjain:matrix.org
    while you wait for an invite you might want to play with some GET endpoints first, ig, but as far as trying the app itself idk
  396. isiah

    In reply to this message

    Gotcha! Will do!
  397. Isiah Hill changed their display name to isiah
  398. Mira D. Purohit (Mira R. Thakkar) joined the room
  399. jg-byte
    is indigo open to contributions?
  400. Armin Nehzat joined the room
  401. charlotte ✨ joined the room
  402. iak2iak joined the room
  403. sitiom joined the room
  404. elisepaul joined the room
  405. joelcmk joined the room
  406. Kandy (They/She, DMs Open) joined the room
  407. Kandy (They/She, DMs Open)
    wow, just learned this room was a thing
  408. hell yea, nice to meet you all!
  409. moved to @shreyan:beeper.com@shreyanjain:matrix.org
    👋
  410. Kandy (They/She, DMs Open)

    In reply to this message

    yo! ❤️
  411. been hearing a lot about bluesky and been praying for that email invite for awhile lmao
  412. im also really curious about what the vibe of the space there is gonna be there once that "new social media" energy dies down!
  413. Kandy (They/She, DMs Open)

    In reply to this message

    update, a kind person sent me an invite, so im here for real now haha!
  414. kumiko9 joined the room
  415. raggi joined the room
  416. Armin Nehzat
    I wish there was another kind person who could send me an invite
  417. I still don't understand how anyone could make money hosting a note or contributing as a creator? Is there any monetisation plan or strategy? or am i missing something?
  418. Kandy (They/She, DMs Open)

    In reply to this message

    oh yea, my username there is <
    kandyelmo.bsky.social
    >
    (edited)
  419. first impression, i like that "what's hot" is basically a kickstart with what goes on in the site
    (edited)
  420. i hope you can make more tabs based off of topics like Entertainment or something like that
  421. best example of what im mentioning!
  422. @ralph:fx45.in joined the room
  423. @planetoryd:matrix.org

    In reply to this message

    it isn't
  424. @interpol:cutefunny.art
    Hi Gays
  425. @glownigger:matrix.ralnet.org left the room
  426. S Y joined the room
  427. WubWoofWolf changed their display name to Pissed my pants in lemonade for daddy
  428. @interpol:cutefunny.art changed their profile picture
  429. @man2dev:matrix.org joined the room
  430. @albattran:matrix.org joined the room
  431. tonykent joined the room
  432. klinark

    In reply to this message

    Good morning Everyone, Hope all is well. I was hoping someone might have an invite to spare. Also wondering if anyone has found any documentation or advice on self-hosting.
  433. Compy@compy:hazenet.org
    Anyone have an invite by chance? Been tinkering with the atproto and indigo repos for a while, looking forward to federation. It'd be nice to have a perspective of the "main instance" experience
  434. Kjartan

    In reply to this message

    About selfhosting: clone the repo. make deps. make build. Go to packages/pds folder. Craft a .env file (keys and their use you can find in src/config.ts) and a "yarn start" and you should be set
  435. Probably forgot some minor steps. If you get any errors, just ask again.
  436. chunningham joined the room
  437. rockhunters08 joined the room
  438. @floragordonstar:matrix.org joined the room
  439. ordwell joined the room
  440. joshlacal
    Are any of the Bluesky client apps produced by Bluesky pbllc open source? If not, are there any plans to open source them?
  441. Compy@compy:hazenet.org
    Agreed. I've been testing my instance using the staging web client. It'd be nice to have a local instance of that too
  442. snarfed
    not yet but they hope to eventually
  443. joshlacal

    In reply to this message

    that’s awesome, thanks so much
  444. alrozzi
    Anyone can share an invite, please🥹
  445. In reply to this message

    Why people do not share 🤷‍♀️
  446. joshlacal

    In reply to this message

    Lots of people here including me don’t have an invite, please don’t spam
  447. alrozzi
    joshlacal: i am not spamming. Sorry!
  448. klinark

    In reply to this message

    Invites are limited and most people here are developers and only wish to give other developers invites to help with setting things up and that's okay.
  449. alrozzi

    In reply to this message

    Understandable. Thanks
  450. Igor Jagec joined the room
  451. goeo_

    In reply to this message

    my did:web was never retrieved despite doing both ways of handle resolution (documented, xrpc; and undocumented but de facto, dns txt), and searching for my profile in search, and trying to go to my profile with a direct link
  452. decentralisation very much not a thing just yet, it seems
  453. retr0id
  454. Pilot_51
    I was wondering the other day why an invite was needed for something that's supposed to be decentralized/federated. I intend to jump on with my own server as I'm doing for Matrix. I appreciate the conversation confirming my suspicions that it's not there yet.
  455. @interpol:cutefunny.art removed their profile picture
  456. @interpol:cutefunny.art removed their display name (Pissed my pants in lemonade for daddy)
  457. @interpol:cutefunny.art left the room
  458. Anonymous
    Because the invite is for the
    bksy.social
    PDS / client
    (edited)
  459. And federation is turned off
  460. And we still didn't get an invite, they are in very short supply it seems
  461. Deavial Drake
    this is the way i see it as i too once thought an invite code was necessary. this group is for development on the protocol which is not bluesky. bluesky is a PDS that uses the protocol, just like the rest of us. so asking us for invites just won’t happen here because we aren’t bluesky. they are a separate app running on atp
  462. Pilot_51
    Why is the limit on the client and rather than the server? The only reason I see for doing it on the client is if you don't want too many people complaining about the bugs.
  463. Deavial Drake
    it took me a bit to see the separation
  464. Anonymous
    The protocol and Bluesky are completely separate that's why
    (edited)
  465. snarfed
    somewhat, not entirely. the repo/MST data model is definitely part of the protocol, shows up in the sync XRPC methods
  466. but also again their federation design isn't published yet
  467. soon!
  468. Anonymous
    It's the lack of the federation part that's the bottleneck, right?
  469. Once that's sorted we shouldn't need to worry about invites
  470. We're currently setting up PDS etc and have a Go firehose working
  471. Pilot_51
    Invites ideally wouldn't block anyone from finding a way to try it, but I think servers should be allowed to require invites.
  472. snarfed

    In reply to this message

    interesting! will you use their TS repo for your PDS? or the indigo code? something else?
  473. Anonymous
    We have our own code
  474. Not settled on Go though, playing with rust as well
  475. https://gitlab.com/bnewbold/adenosine will look at this later in the week
  476. I think I'd prefer rust, we'll see
  477. @neeg:nitro.chat joined the room
  478. snarfed
    Anonymous: is your code open source? I'd love to look. I'm in the middle of my own PDS implementation, https://github.com/snarfed/arroba
  479. the repo commit chain and MST are nontrivial to build, significantly more work than eg an AP implementation, which is already a big project
  480. Anonymous
    It's closed at moment but will be open once we settle, it's too early to say and maybe we'll just switch to contributing to someone else's repo
  481. Will take a look at that, thanks
  482. That's worth a look as well
  483. Deavial Drake
    does anyone know if changing a handle triggers a profile change? we don’t see either DID or handle in the dataset spec (
    app.bsky.actor
    .profile)
    (edited)
  484. @dev_dude:matrix.org left the room
  485. SGOTI joined the room
  486. ryangallagher
    I’m just observing this thread as it develops. I’m curious as you all are creating the product has there been any thought put into one user that may need to fast switch between multiple accounts to manage sister brands? (Similar to how instagram does.) We are a municipal corporation that operates five large venues outside our core brand and being able to respond to customer requests via a central dashboard would be extremely helpful.
  487. snarfed

    In reply to this message

    interesting q! a couple points:

    1. none of us here work at bluesky. a few of them are here infrequently, but not often
    2. my sense is, it's way too early for them to build those kinds of tools. they've been working hard on basic admin moderation tools and user features like blocking, they're nowhere near corporate brand mgmt stuff
    (edited)
  488. on the plus side, it's very open, you could easily build those tools yourselves, depending on how much technical bandwidth you have
  489. wd
    it seems like custom clients are extremely supported (and work today, there's lots of custom clients)
  490. but yeah might be worth pulling in one of the app devs of existing custom clients -- there's a tweetdeck clone being worked on e.g. that feels like a natural fit
  491. ryangallagher
    Good to know. I wouldn’t be the builder I just have a lot of experience on the moderation side of things in out of home industry advising on what the tools need to do.
  492. Francis Chacko joined the room
  493. zonsopkomst joined the room
  494. @aaap:matrix.org joined the room
  495. @spencerthayer:matrix.org left the room
  496. cypher joined the room
  497. madiator2011 joined the room
  498. foxlet joined the room
  499. alexanderstraub joined the room
  500. foxlet left the room
  501. Alex Smola joined the room
  502. foxlet joined the room
  503. @fractal:matrix.org left the room
  504. Kjartan
  505. Can someone explain what's going on here? Never had it green and pink at the same time
  506. foxlet left the room
  507. wd
    I can't resolve that txt record?
  508. ❯ dig _
    atproto.sbane.de
    in txt ; <<>> DiG 9.10.6 <<>> _
    atproto.sbane.de
    in txt ;; global options: +cmd ;; Got answer: ;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 23289 ;; flags: qr rd ra; QUERY: 1, ANSWER: 0, AUTHORITY: 1, ADDITIONAL: 1 ;; OPT PSEUDOSECTION: ; EDNS: version: 0, flags:; udp: 1232 ;; QUESTION SECTION: ;_
    atproto.sbane.de
    . IN TXT ;; AUTHORITY SECTION:
    sbane.de
    . 900 IN SOA
    dns1.yandex.net
    .
    dns-hosting.yandex.ru
    . 16 900 90 86400 900 ;; Query time: 205 msec ;; SERVER: 192.152.0.1#53(192.152.0.1) ;; WHEN: Tue May 02 17:36:06 PDT 2023 ;; MSG SIZE rcvd: 118
  509. vs mine which works: ❯ dig _
    atproto.wd.teleri.net
    in txt ; <<>> DiG 9.10.6 <<>> _
    atproto.wd.teleri.net
    in txt ;; global options: +cmd ;; Got answer: ;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 44506 ;; flags: qr rd ra; QUERY: 1, ANSWER: 1, AUTHORITY: 0, ADDITIONAL: 1 ;; OPT PSEUDOSECTION: ; EDNS: version: 0, flags:; udp: 1232 ;; QUESTION SECTION: ;_
    atproto.wd.teleri.net
    . IN TXT ;; ANSWER SECTION: _
    atproto.wd.teleri.net
    . 86400 IN TXT "did=did:plc:3la4hhaon456xbzr3q6yd7i4" ;; Query time: 61 msec ;; SERVER: 192.152.0.1#53(192.152.0.1) ;; WHEN: Tue May 02 17:36:22 PDT 2023 ;; MSG SIZE rcvd: 100
  510. (not sure how you got "domain verified" there tbh, that's weird...)
  511. awesomepandapig joined the room
  512. awesomepandapig
    I am trying to build the repo on my apple silicon mac and it fails when trying to install the better-sqlite3 package. Does anyone know how to solve this?
    (edited)
  513. Aaron Goldman
    Did you try installing the x86 one and letting the compatibility layer do it's thing
  514. SQLite should build for apple silicon though
  515. mansuoor set a profile picture
  516. memory-system joined the room
  517. Saloframes joined the room
  518. foxlet joined the room
  519. @sanktmax:matrix.org
    Anywhere to install the Bluesky APK?
  520. @planetoryd:matrix.org

    In reply to this message

    build it yourself
  521. @sanktmax:matrix.org

    In reply to this message

    How?
  522. @planetoryd:matrix.org

    In reply to this message

    do you need an invite code
  523. @sanktmax:matrix.org

    In reply to this message

    I signed up for an invite back in October 2022, but still haven't gotten one.
  524. Kjartan

    In reply to this message

    Thanks. Mh… I think it might somewhat makes sense. Because it was able to resolve it here (I'm more surprised that it didn't work when you tried it). But that's probably the "answer" - it worked for the app (resulting in green, verified), but then didn't work for the server (like it was for you - and resulted in a pink failed)
  525. plein changed their display name to plein (DMs closed)
  526. Kjartan
    Message deleted
  527. kcchu
    Message deleted
  528. foxcode joined the room
  529. foxcode
    Hey everyone, is there API documentation that's public for Bluesky and if so, within that documentation are there any type of rate limits or other restrictions posted that I can read about?
  530. Hissan Usmani joined the room
  531. @neeg:nitro.chat changed their profile picture
  532. Quetzalcoatl joined the room
  533. mattezell joined the room
  534. mn05org
    Is anyone working on a Wordpress plugin similar to the one for activitypub that turns a Wordpress site into a node on the federation? So much activity in this space right now, looks like there will be 3 decentralized socials to choose from when the dust settles…
  535. forivall joined the room
  536. Mark Foster SSI: @mfoster.io
    I need to update
    futureinternet.io
    and I’m burned out
  537. Anyone want to help?
  538. @iancjclarke:matrix.org joined the room
  539. Mark Foster SSI: @mfoster.io
    Want to move it off Wordpress and into a more modern stack and update it with bluesky IPFs libp2p resources
  540. @planetoryd:matrix.org

    In reply to this message

    what do you want
  541. Mark Foster SSI: @mfoster.io
    @planetoryd:matrix.org: where do I start 🤦🏻‍♂️
  542. @planetoryd:matrix.org
    that's a website. why do you need bluesky for it
  543. Mark Foster SSI: @mfoster.io
    First need to pick the right publishing dev stack
  544. @planetoryd:matrix.org: No I want it to be an informative resource
  545. @planetoryd:matrix.org
    ok, then you write articles
  546. Mark Foster SSI: @mfoster.io
    Yes I do, looking for anyone that wants contribute
  547. It’s time for me
  548. @planetoryd:matrix.org

    In reply to this message

    the projects listed on your site sound like shitcoins
  549. Mark Foster SSI: @mfoster.io
    They are not
  550. @planetoryd:matrix.org
    proof
  551. Mark Foster SSI: @mfoster.io
    Solid is a specification
  552. Do you know what Solid is
  553. @planetoryd:matrix.org
    doesnt matter. there are tons of specs already
  554. Mark Foster SSI: @mfoster.io
    🤦🏻‍♂️
  555. Read my initiated comment
  556. snarfed

    In reply to this message

    pfefferle, author of the WP AP plugin, is experimenting with this right now. eg https://blue.amazingca.dev/?username=notiz.blog&postid=3jurj3zu4na2j
  557. Mark Foster SSI: @mfoster.io
    Awesome
  558. In reply to this message

    You can use the APIs to sync posts and make your profile the author
  559. snarfed
    yup. the more interesting approach eventually though will be federation, not syndication (cross-posting). eg pfefferle's AP plugin does federation
  560. Mark Foster SSI: @mfoster.io
    Yeah the whole database store would have to change to CARs for WordPress
  561. Have you looked at your repo yet
  562. snarfed
    looked at your repo?
  563. technically ATP doesn't impose any req'ts on the underlying storage. it uses CAR, IPLD encodings, etc on the wire, but you can store data locally however you want
  564. Mark Foster SSI: @mfoster.io changed their profile picture
  565. Mark Foster SSI: @mfoster.io
    What is your handle
  566. snarfed
    where?
  567. snarfed on github,
    snarfed.org
    on bsky and web
  568. Mark Foster SSI: @mfoster.io
    Atproto
  569. BlueSky handle
  570. snarfed
    lol sorry. ATP identity is via DIDs, either did:plc or did:web. https://atproto.com/guides/identity
  571. Mark Foster SSI: @mfoster.io
    Yes your did
  572. What is your did or handle
  573. Either one
  574. snarfed
    mine personally in bluesky right now is
    snarfed.org
  575. Mark Foster SSI: @mfoster.io
    Ok one sec
    (edited)
  576. Jeff Rhie joined the room
  577. Jeff Rhie
  578. Mark Foster SSI: @mfoster.io
    This is your CAR https://bsky.social/xrpc/com.atproto.sync.getRepo?did=did:plc:fdme4gb7mu7zrie7peay7tst In order for WordPress to truly be on the ATProto the store would need to be a CAR so it would have to be completely rebuilt and moved away from a MySQL database and the underlying store would need to be the following Content Addressable aRchives (CAR / .car) https://ipld.io/specs/transport/car/
  579. In reply to this message

    See below
  580. snarfed
    no, as I mentioned, you can separate actual storage in WordPress from format on the wire
  581. you can load data from MySQL, reformat and re-encode it into DAG-CBOR CAR in memory, and then serve that
  582. Mark Foster SSI: @mfoster.io
  583. Sounds like an interesting concept I originally built
    futureinternet.io
    on WordPress because I didn't have time to allocate to it so that sounds like an interesting concept to me
  584. Jeff Rhie
    Hey guys :)
  585. Jeff Rhie

    One question, I was reading the blog post on Bluesky and convinced by the idea of account portability using DID.

    But if dns becomes an handle of your id, won’t it be no longer ‘portable’ since dns can only be acquired through centralized registrar?

    I’m a newbie so I may be asking a stupid question, but please help me :)

  586. Mark Foster SSI: @mfoster.io

    In reply to this message

    Right now the did:plc is being used which stands for placeholder the goal is to start creating unique identifiers currently on technology DNS that people currently understand then eventually iterate the protocol as it moves forward https://atproto.com/specs/did-plc
  587. Right now you can interact with repository with https://atproto.com/specs/xrpc
  588. Jeff Rhie

    In reply to this message

    Aba, thanks for the info! I wonder eventually it will be on the Bitcoin network, (sth like Microsoft ION?)

    Also, if anyone has a remaining invite code, it would be so thankful if you can share me one via DM. Dying to use bluesky😇

  589. Mark Foster SSI: @mfoster.io
    It's a little confusing at first but you have to think of it in fashion of an Orthogonal Web
  590. Top down with DNS and server to server messaging Sync
  591. Jeff Rhie

    In reply to this message

    Sounds like an compromise for scale tho.
  592. Mark Foster SSI: @mfoster.io

    In reply to this message

    Yes it could be bridged with did:ion eventually... if you are interested in did:ion you should start an open source project around it
  593. Jeff Rhie

    In reply to this message

    I thought Microsoft abandoned the project, wasn’t it? Never heard anything significant about ION since 2021.
  594. Mark Foster SSI: @mfoster.io
    They did create the project at the DIF https://identity.foundation/
  595. They used OP_RETURN for commitments
  596. yablak joined the room
  597. Jawa joined the room
  598. yablak
    Hi everyone; qq. Does bluesky (main server) have a content moderation policy?
  599. bnewbold: sounds like you may have thought about this in the past?
  600. Scott Beamer joined the room
  601. 🌎⭐️🌟Flora Gordon joined the room
  602. Sam joined the room
  603. @iancjclarke:matrix.org left the room
  604. Mark Foster SSI: @mfoster.io
    Yes it does
  605. yablak
    This is more a technical answer (good) but I was thinking something along the lines of this: https://transparency.fb.com/policies/community-standards/
  606. snarfed
    there's https://staging.bsky.app/support/tos and https://staging.bsky.app/support/privacy , but otherwise afaik no, their policies for the main server are still very young and evolving fast, I don't think they have anything anywhere near that formalized yet
  607. Chris Lace
    Message deleted
  608. bignessie

    In reply to this message

    the TOS page also includes a link to their community guidelines, which seems to be what you're looking for
  609. moved to @shreyan:beeper.com@shreyanjain:matrix.org

    In reply to this message

    Solid is awesome
  610. Mark Foster SSI: @mfoster.io
    Yes it is I originally started in Solid and expanded out
  611. moved to @shreyan:beeper.com@shreyanjain:matrix.org

    In reply to this message

    that is not really possible right now
  612. but you could totally use something like snarfed's bridgy
  613. it doesnt have bluesky integration yet but that lands soon from what i hear
  614. Mark Foster SSI: @mfoster.io
    Yes we talked about it
  615. moved to @shreyan:beeper.com@shreyanjain:matrix.org
    yeah cool
  616. Mark Foster SSI: @mfoster.io
    I might need to add a little more context it’s built on WordPress and I’m thinking about moving to a more modern CMS to stay in the modern frameworks I’m working in now
  617. I’m working on a NextJS/Expo mix and every time I want to go update it I’m like meh I’ll do it later
    (edited)
  618. @reiver (CHARLES ILIYA KREMPEAUX)
  619. flawedworld joined the room
  620. @reiver (CHARLES ILIYA KREMPEAUX)

    In reply to this message

    Someone "took over" the "

    " domain from the point of view of BlueSky / AT-Protocol.

    If I understand what he did accurately —

    He made the S3 URL:

    https://s3.amazonaws.com/xrpc/com.atproto.identity.resolveHandle​

    Return:

    {"did":"did:plc:imkvi5glxfcpaqcinktnbpwt"}

    .

  621. moved to @shreyan:beeper.com@shreyanjain:matrix.org
    yeah.
  622. that's what I assumed too
  623. the specific url you sent resolves to an error though
  624. 123456<Error>
    <Code>AccessDenied</Code>
    <Message>Access Denied</Message>
    <RequestId>E59TYEEEYYJZ04N7</RequestId>
    <HostId>Io+4JkFKqPr9kkW3ZflcgfeQfwq1cdRALqxvo43vAzNaPtDus1mQXQJ0qu3+cAH0dCvstnpt6ok=</HostId>
    </Error>
    
  625. @reiver (CHARLES ILIYA KREMPEAUX)

    In reply to this message

    It was working minutes ago.
  626. moved to @shreyan:beeper.com@shreyanjain:matrix.org
    they probably took it down then
  627. or he did
  628. @reiver (CHARLES ILIYA KREMPEAUX)
  629. moved to @shreyan:beeper.com@shreyanjain:matrix.org
    hmm i can't get it to work on my browser
  630. that's still impressive
    (edited)
  631. remove %E2%80%8B from address, and it will show above
  632. moved to @shreyan:beeper.com@shreyanjain:matrix.org
    yep thanks
  633. Mark Foster SSI: @mfoster.io
    Url encoding 🤦🏻‍♂️
  634. moved to @shreyan:beeper.com@shreyanjain:matrix.org
    reminds me of webfinger
  635. @dead10ck:dead10ck.com

    In reply to this message

    Lol see this is exactly what I was thinking could happen when you have handle resolution in a separate protocol stack from DNS. You allow implementations that aren't DNS to cause a split brain. You can't allow it to be implementation agnostic and have a consistent delegation of authority.

    If we want DNS handles, then DNS needs to actually be specified in the spec and not just be an option

  636. vernon (building @ yup)
    what's the best way to decode binary data received from subscribe repos ws api
  637. trying to do this in nodejs
    (edited)
  638. Mark Foster SSI: @mfoster.io

    In reply to this message

    It’s part of the did:web method did resolvers should follow the scheme of its method uri right now did:plc is not official while in beta it is a placeholder but similar to did:web temporary https://atproto.com/guides/identity

    https://www.w3.org/TR/did-core/#a-simple-example

    https://dev.uniresolver.io/

  639. lamrongol
    Is Bluesky planning to migrate from Typescript to Go? Bluesky is looking for Go programmers most ( https://blueskyweb.xyz/join ) and also requires experience of Go for Typescript Developer( https://blueskyweb.xyz/join/ts-developer ).
  640. mark set a profile picture
  641. @dead10ck:dead10ck.com

    In reply to this message

    Yeah, the problem with did:web though is that your DID is forever tied to that domain name. If you lose the domain name because you wanted to change it or because it unintentionally lapsed and someone else grabbed it, then you're SOL and you've lost your handle
  642. I can only presume this was at least part of the reason bsky didn't consider it suitable, given the goal of portability
  643. Mark Foster SSI: @mfoster.io

    In reply to this message

    Since you control the did document with the encryption and authentication methods described in it you can update the did to point to your new domain. This is part of the Self Sovereign Identity methodology of thinking
  644. @dead10ck:dead10ck.com

    In reply to this message

    Are you referring to the
    service endpoint
    ?
  645. Mark Foster SSI: @mfoster.io
    So imagine this thought process and Zero Knowledge Proofs on a Orthogonal web of trust
  646. @dead10ck:dead10ck.com
    Doesn't the id itself remain the same forever though?
  647. Mark Foster SSI: @mfoster.io
    I am talking about the handle token in the did:plc did document here is mine https://bsky.social/xrpc/com.atproto.repo.describeRepo?repo=mfoster.io
  648. @dead10ck:dead10ck.com

    In reply to this message

    Right, that's did:plc though, not did:web. Identity in plc is not tied to a domain name
  649. Aaron Goldman

    In reply to this message

    There is nothing wrong with having both a DNS and https domain binding methods as long as the priority is well defined
  650. Mark Foster SSI: @mfoster.io
    You can replace the get value in that xrpc call with you did and it returns the same document
  651. @dead10ck:dead10ck.com

    In reply to this message

    I bet Amazon will have a different opinion about this 👀
  652. Aaron Goldman
    I would prefer DNS overrides https but as long as order is defined meh
  653. That was the argument for using .well-known/ rather than xrpc 🤷‍♂️
  654. Mark Foster SSI: @mfoster.io

    In reply to this message

    Are you talking about the the _aroroto TXT record
  655. @dead10ck:dead10ck.com
    You get the same problem if a atproto DNS record doesn't exist, it will just fall back to https and you get the same split brain
  656. Aaron Goldman
    But maybe DNS > well-known > xrpc is the way to go
  657. Mark Foster SSI: @mfoster.io

    In reply to this message

    You could even do WebFinger
  658. @dead10ck:dead10ck.com
    DNS doesn't exist → well known doesn't exist → rando s3 bucket hijacks identity
  659. Aaron Goldman
    I think that is not a bug
  660. If you control the bucket you control the handle but for a better claim
  661. @dead10ck:dead10ck.com
    Amazon shouldn't have to have a bsky account to avoid having random people impersonate them
  662. Aaron Goldman
    And if Amazon wants to make a better claim they can
  663. Mark Foster SSI: @mfoster.io
    Utilize every fallback possible but ultimately I think VCs linked to real world brick and mortar third parties of trust could fix this issue
  664. Aaron Goldman
    If you let me set your robots.txt I can choose your crawling preference. It's a reasonable way to handle it
  665. Mark Foster SSI: @mfoster.io
    Aaron Goldman: I’ve been thinking of Marks webs of trust with an s theory vs web of trust
  666. If we have Semantic Webs of trust we are moving towards Peter Wang’s Orthogonal Distributed Web theory
    (edited)
  667. When I say semantic I’m not referring to RDF specifically
  668. Any kind of semantics as long as there’s some presentation exchange like schema negotiations
  669. Only Paul could make schema negotiations and intent popular with shit posting a dirty sock 🧦 😂
  670. @dead10ck:dead10ck.com

    This has the potential to happen with any web service that lets users host their own content. It's actually interesting to note that this person claims to be an Amazon security engineer. I imagine they will be reaching out about this.

    Personally I think the protocol should do everything it can to guard against this kind of identity theft. If it doesn't, I don't really see what the point of domain name handles is. Making everyone everywhere have to opt into identity theft protection sure does sound like a bug to me.

  671. Aaron Goldman
    That was the whole point of .well-known/* Protocols can use it and every server only has to protect /.well-known/* not 10,000 different robots.txt like things
  672. Mark Foster SSI: @mfoster.io
    This is where the moderation lexicons can help and social claims hence a mesh network of trust https://github.com/bluesky-social/atproto/blob/main/lexicons/com/atproto/moderation/defs.json
  673. Aaron Goldman
  674. Mark Foster SSI: @mfoster.io
    There are several ways to tell the network including humans looking out for humans in the open
  675. @dead10ck:dead10ck.com

    In reply to this message

    In my opinion, no one should have to participate in the network to avoid being impersonated on that network
  676. In reply to this message

    Yeah but they doesn't mean anything when you can have /xrpc/*, which I bet will be a lot easier to pull off on domains you don't own than /.well-known/*
  677. Mark Foster SSI: @mfoster.io
    When you combine those concepts together the issue you are concerned can easily be resolved when the people that work at the companies you mention are part of the social network
  678. When the claim is out in the open there is accountability
  679. @dead10ck:dead10ck.com
    That's my point: you have several vectors of attack and allow the weakest link to win
  680. Mark Foster SSI: @mfoster.io
    AI is part of this network as well 😉
  681. Aaron Goldman
    I argued for well known I lost
  682. moved to @shreyan:beeper.com@shreyanjain:matrix.org
    wish you'd won
  683. everyone else uses it for good reason
  684. webfinger, nostr...
  685. Aaron Goldman
    Yeah xrpc asks each domain with user generated content to know about and block that path
  686. Shame it's too late for robots.txt but we have it for future protocols
  687. @dead10ck:dead10ck.com

    In reply to this message

    Curious what the argument against it was?
  688. If you're at liberty to share
  689. @caramelpen:matrix.org left the room
  690. mimizo joined the room
  691. Aaron Goldman
    It makes the routing more complicated. A PDS can't just route xrpc but also a well-known which often requires a security team sign off
  692. @dead10ck:dead10ck.com

    In reply to this message

    In reply to
    Aaron Goldman
    It makes the routing more complicated.
    A PDS can't just route xrpc but also a well-known which often requires a security team sign off
    Not sure I understand. In a scenario where you need security team sign off, how is .well-known more difficult than /xrpc? And why can't a PDS route /xrpc? Anyone can write a web server to do whatever they want, can't they?
    (edited)
  693. Mark Foster SSI: @mfoster.io
    Is the issue that if one reverse look up values is off it fails to resolve? If webfinger and robots.txt is added could a propagation stream be initiated with something like big endian little endian data packing with all types or maybe something like a vector 🤔
  694. In reply to this message

    Yes I hacked my SoLid pod to have a webfinger
  695. moved to @shreyan:beeper.com@shreyanjain:matrix.org
    i thought solid did its own thing with activitypub
  696. @dead10ck:dead10ck.com
    AP also uses well known delegation
  697. moved to @shreyan:beeper.com@shreyanjain:matrix.org
    .well-known/webfinger or something like that, i forget
  698. I was able to cross post from Solid to Mastodon this way
  699. Xinyu Ma joined the room
  700. Mark Foster SSI: @mfoster.io

    In reply to this message

    SoLiD is just a node.js Personal Online Data Store that can store anything… wrapped with linked data to query it with a decentralized query language like SPARQL or GraphQL-LD it has WebSockets caked in and a SPARQL endpoint to send data back and forth to other pods to create a mesh network it has granular control for ACL privacy and WAC

    https://github.com/nodeSolidServer/node-solid-server

  701. moved to @shreyan:beeper.com@shreyanjain:matrix.org
    in theory I guess you could use Solid as a BlueSky PDS, but it would take some tweaking to both
  702. Mark Foster SSI: @mfoster.io
    Yeah I thought about that
  703. moved to @shreyan:beeper.com@shreyanjain:matrix.org
    If that happens it would be awesome, but I don't know if there's much interest
  704. Mark Foster SSI: @mfoster.io
    It would be interesting
  705. To nerds like me
  706. Yiko Song joined the room
  707. Dominick Rangel joined the room
  708. @costanza99:matrix.org removed their profile picture
  709. @costanza99:matrix.org removed their display name (costanza99)
  710. @costanza99:matrix.org left the room
  711. Yusuf joined the room
  712. nathannever changed their display name to Yusuf
  713. Arthur Lonsdale joined the room
  714. @julian:nope.chat joined the room
  715. @julian:nope.chat left the room
  716. @neilalexander:matrix.org

    _atproto.microsoft.com. 3600 IN TXT "did=did:plc:fivojrvylkim4nuo3pfqcf3k"

    alright which one of you was responsible for this

  717. @sarangbaloch:matrix.org left the room
  718. jesopo
  719. lamrongol
    https://staging.bsky.app/profile/did:plc:fivojrvylkim4nuo3pfqcf3k has "
    microsoft.com
    " name but no icon and description and posts. Is this real? Or Is there a hall in Bluesky?
  720. jesopo
    it's real
  721. lamrongol
    Hmmm... Why don't they announce or even set up a profile?
  722. jesopo
    be patient?
  723. @pathumego:matrix.org left the room
  724. @dead10ck:dead10ck.com
    It sounds like there are a lot of companies experimenting with it right now and big ones are starting to stake their claims
  725. Mark Foster SSI: @mfoster.io

    In reply to this message

    The future 🚀
  726. Aaron Goldman

    In reply to this message

    id 13061 opcode QUERY rcode NOERROR flags QR RD RA ;QUESTION _
    atproto.microsoft.com
    . IN TXT ;ANSWER _
    atproto.microsoft.com
    . 3600 IN TXT "did=did:plc:fivojrvylkim4nuo3pfqcf3k" ;AUTHORITY ;ADDITIONAL
  727. xnf0k

    Also:

    12❯ dig @1.1.1.1 _atproto.figma.com TXT +short
    "did=did:plc:z7mxys3ms5slfb26r5vmtve2"
    
    (edited)
  728. @sudousrcode:matrix.org left the room
  729. Anonymous
    You might find this handy if you're keeping an eye on your favorite domains https://pastebin.com/NC1YaZZT
  730. @neeg:nitro.chat
    Is there any protection from DNS hijacking in bsky/atproto?
  731. snarfed
    probably not. we believe it supports DNSSEC, implicitly, but nothing more than that
  732. feel free to try to hijack a domain as a proof of concept if you want!
  733. (I expect we'd be curious to hear specifically how you would)
  734. Gustavo Garcia De Andrade joined the room
  735. Kjartan

    In reply to this message

    Thats actually me (not kidding) but its fake
  736. It also has a couple of users, if you continue searching :)
  737. Also it has posts, just not on bsky (although the instance claims to be the main instance)
    (edited)
  738. snarfed
    federation still isn't turned on yet
  739. regardless, interesting! mind telling us how you did the DNS hijacking?
  740. xnf0k
    Did you submit to PLC directly?
  741. But no the TXT record really exists hm
  742. snarfed
    right, hence my question. it'd be nice to see some proof that Kjartan really does own that DID and did create that TXT record, but if so, hell of an accomplishment
  743. Kjartan

    In reply to this message

    Happened here 👆
  744. snarfed
    oh so then you're saying a different thing. so the current _
    atproto.microsoft.com
    DNS TXT record isn't yours?
    (edited)
  745. Kjartan

    In reply to this message

    There has never been a txt record for it. Create a pds, configure it to use
    microsoft.com
    as an available handle prefix and the pds to use
    bsky.app
    as its own address - and there you have it
  746. snarfed
    there's an _
    atproto.microsoft.com
    TXT record right now. run dig txt _atproto.microsoft.com to see it
    (edited)
  747. Kjartan
    (Would be actually easier to just send the request to the plc registry though, if the only thing one would want to do is to create such a record in plc registry)
  748. snarfed
    sounds like that's separate from yours
  749. Kjartan

    In reply to this message

    Ohhhh. I didnt now. That wasn't me though :D
  750. snarfed
    they haven't finished or published the federation design yet, or how DNS handles will work with it, so it doesn't really matter what you do on your own PDS right now that's not connected to anything
  751. so yeah, you can register any domain with
    plc.directory
    , but that doesn't mean anything until you also create the corresponding DNS record
    (edited)
  752. Kjartan

    In reply to this message

    "Not to anything" it's connected to the plc.registry - doesn't matter though
  753. snarfed
    right, doesn't matter
    (edited)
  754. Kjartan

    In reply to this message

    True true
  755. Kjartan
    I only saw it on the phone, and missed the "there is a txt record" part (which definitely isn't of me) 👍
  756. code. joined the room
  757. Shol K (Kshol) joined the room
  758. Kjartan
    wouldn't know where else to report it, so I'll mention it here: if you have single use invite codes, and multi use invite codes, the client (at staging…) tells you there would be all single use (and it seems also not to be visible which ones might be the multiuse codes - at least to me they look all the same in the ui)
  759. Aaron Goldman
    I wonder what the UI should do with the half-bound names. If it is just the DNS or just the DID Document but not the other. Probably just not show anything?
  760. Kjartan
    Aaand schould someone know where to report client issues. The address for What's hot is hard coded. And worse: if therefore the request fails is wrong handled (as in "hey, update your language settings bla bla" although the request never was successful because it didn't went where it was supposed to go, because of the first mentioned bug)
  761. Soooo. Sorry for the delay. And first of all, I have to mention: I think something isn't quite right. Should you run somewhere into an issue, but can't figure out what you are doing wrong, it could be because of something being wrong with the server (wouldn't want you to spend hours on something which was actually kinda my fault).
  762. And: please don't upload a lot of images or so, ideally less than 50MB of blob data each of you, as there is only about 500MB left :-(
    (edited)
  763. Feel free to create directly up to 5 users, so you can test things better (like, liking each others posts, etc)
  764. Kjartan
    "schnitzel-mit-pommes-de-l5u5ti7", "schnitzel-mit-pommes-de-4lnpyep", "schnitzel-mit-pommes-de-jnotpci", "schnitzel-mit-pommes-de-uduat54", "schnitzel-mit-pommes-de-7nu5zor", "schnitzel-mit-pommes-de-uiovcii", "schnitzel-mit-pommes-de-usvexhs", "schnitzel-mit-pommes-de-krhbixd", "schnitzel-mit-pommes-de-fb4eury", "schnitzel-mit-pommes-de-thlhoax", "schnitzel-mit-pommes-de-cazscw5", "schnitzel-mit-pommes-de-sivug6s", "schnitzel-mit-pommes-de-hfhridk", "schnitzel-mit-pommes-de-rgqq7r3", "schnitzel-mit-pommes-de-spsrpa6", "schnitzel-mit-pommes-de-tmcinnr", "schnitzel-mit-pommes-de-bj2gcog", "schnitzel-mit-pommes-de-zhzyltu", "schnitzel-mit-pommes-de-sqliou2", "schnitzel-mit-pommes-de-4zjm76q", "schnitzel-mit-pommes-de-gvfeesa", "schnitzel-mit-pommes-de-noqrpf2", "schnitzel-mit-pommes-de-lwseer5", "schnitzel-mit-pommes-de-ulxjwcp", "schnitzel-mit-pommes-de-gdkuzar", "schnitzel-mit-pommes-de-v6cqboz", "schnitzel-mit-pommes-de-qgjgusv", "schnitzel-mit-pommes-de-u7aagqa", "schnitzel-mit-pommes-de-oqd5ijb", "schnitzel-mit-pommes-de-4tgz7vm", "schnitzel-mit-pommes-de-7wzdq3p", "schnitzel-mit-pommes-de-timzla2", "schnitzel-mit-pommes-de-6pl4jzq", "schnitzel-mit-pommes-de-agw54t6", "schnitzel-mit-pommes-de-unv5ifr", "schnitzel-mit-pommes-de-mwwn2vb", "schnitzel-mit-pommes-de-lwx56v7", "schnitzel-mit-pommes-de-llsswwy", "schnitzel-mit-pommes-de-bkbeemk", "schnitzel-mit-pommes-de-erqfxlt", "schnitzel-mit-pommes-de-jzriiqu", "schnitzel-mit-pommes-de-maawdx4", "schnitzel-mit-pommes-de-2u6sxo4", "schnitzel-mit-pommes-de-z6sbxqq", "schnitzel-mit-pommes-de-ovees7c", "schnitzel-mit-pommes-de-e2jsysp", "schnitzel-mit-pommes-de-5ytpb6o", "schnitzel-mit-pommes-de-g73gsxn", "schnitzel-mit-pommes-de-33hqp6g", "schnitzel-mit-pommes-de-yvh7kgw", "schnitzel-mit-pommes-de-zesyn47", "schnitzel-mit-pommes-de-axys5e3", "schnitzel-mit-pommes-de-ynv3dtr", "schnitzel-mit-pommes-de-usbimi2", "schnitzel-mit-pommes-de-gfsnght", "schnitzel-mit-pommes-de-pjxeouw", "schnitzel-mit-pommes-de-vzebxce", "schnitzel-mit-pommes-de-w5ldnxr"]
  765. have fun!
  766. (if you use only 2 or 5mb for pictures, I won't be sad either ;-) Or jsut delete pictures once you are done with testing
  767. Flora joined the room
  768. Kjartan
    Also: "what's hot" works in general, but not with the client at staging.bsky… use "
    kite.black
    " or something (or write your own :P)
  769. Oh, and I didn't connect it to an email server (I hope so at least), so "I forgot my password"-thingy won't work
  770. In reply to this message

    ping
  771. Should you run out of codes - let me know
  772. snarfed
    you can submit feedback directly in the app
  773. Kjartan

    In reply to this message

    They completely ignore me (for whatever reason) same when I ping them on bsky (while they are online and active). But would be still nice if things get noticed and eventually fixed :D
  774. snarfed
    the team has been pretty slammed and overwhelmed for a while now, definitely don't take it personally 😀
  775. Kjartan
    I kinda do (while I have no doubts that they have been busy a lot). Anyway, just thought I mention those (=bugs) so they can be found somewhere (and meanwhile people can be aware of them)
  776. @alexdeltax:genix.chat joined the room
  777. Miguel Aieo joined the room
  778. mst4uth joined the room
  779. mikuhl
    app.bsky.unspecced.getPopular should be allowed without a session
Next group of messages