#dat

pfrazee: 👍

hmm i dont think discovery channel/swarm has a way to only announce, not query

would be easy to add just isnt in there yet

jhand: check out dat-doctor now

cc pfrazee

i deployed the public peer to doctor1.publicbits.org

!add 7dfee0777e34e3d469bf3cdf1b83ebf49f4aff9143d49804bb93eb2a8ae5e500

Adding 7dfee0777e34e3d469bf3cdf1b83ebf49f4aff9143d49804bb93eb2a8ae5e500

7dfee0777e34e3d469bf3cdf1b83ebf49f4aff9143d49804bb93eb2a8ae5e500 has been fully archived (35.8 MB)

[dat] clemsos opened issue #595: Can dat be used offline ? https://github.com/datproject/dat/issues/595

(marcelklehr) Can you guys try to access the following archive? I'm hosting using the dat cli, now, instead of beaker: b7777e3c824a8b8fb29c6c6187843aa4c807d87452c695c262a590f80c196e2e

substack: @marcelklehr

whoops

too many mentions

@marcelklehr still here?

(marcelklehr) yup :)

substack: tree seems to work great on peermaps data set. using the non optimized lookup i can stat any antry in around 500ms on cold boot

gonna add delete support and start integrating it

@marcelklehr you still seeding that dat? i can try fetching it

(marcelklehr) yup, dat is till running

(marcelklehr) (recently I tried hosting with beaker, which didn't work, so now I'm trying the dat cli)

yea beaker still has some weird replication issue

let me try it then

@marcelklehr got it

(marcelklehr) yay! :)

(marcelklehr) awesome, thanks!

cool website

(marcelklehr) heh, it's still kind of raw. It's supposed to be a collaborative editor

(marcelklehr) hit create to create a new doc. Share the id of the new do with people, they can fork the dat archive, give you the id of the fork, you add them as collaborators, bang: collaborate.

(marcelklehr) *new doc

@marcelklehr link? If you don't mind me checking it out :)

(marcelklehr) dat://b7777e3c824a8b8fb29c6c6187843aa4c807d87452c695c262a590f80c196e2e/ ;)

thx!!

(marcelklehr) it worked last time i checked, will most likely fail now, though heh

haha, yeah timed out.

The timeout in beaker seems pretty short though

notwes: heeey

which timezone you in?

hey!! im in us central

cool, you are one those people that start work "early" i guess

:)

haha, yeah well sometimes. the vacation around the holidays has recharged me

also, getting into the office early saves tons of commute time. Austin traffic is pretty bad sometimes

ah right

makes sense

mafintosh: is there any debug tool for figuring out why marcelklehr's dat wont load in beaker?

notwes: only dat cli related ones that i know of. jhand might know more

but i know paul is working on getting the bug fixed

(marcelklehr) wait, so you can pull it with dat, but not view it in beaker. that's weird

oh yeah, dat doctor right?

no, i havent tried pulling it yet, going to now

this is my work computer, hadnt installed dat yet, so still waiting for the npm install to finish

does dat doctor take a dat url with protocol, or just a key?

because it seemed to do stuff with the full dat:// but not with just the hash.

(marcelklehr) I think it gave me a simple key last time. I was using dat-next, though I think

ok, i'll switch to that and see

@marcelklehr yea thats the bug

that is my output from dat-next doctor

i am guessing it found beaker as a peer

but nothing else?

not sold on beaker using that port to peer, at least lsof didn't mention it

(marcelklehr) does the doctor work with normal dat urls?

(marcelklehr) I thought the keys you get from the doctor are custom ones

oh, if so then I didn't know. The discovery key?

(marcelklehr) I think, one side runs the doctor without arguments and gets a key/token then the other side runs the doctor with that token.

ohhhh....

dat doctor 079ecf8135966dee961d47ea2a1aa4ccefd2623c086cb6c66aa3e31b70303a3b

what to give that a try?

thats what mine just printed

(marcelklehr) one sec

(marcelklehr) ```

(marcelklehr) [0000] Connection established to remote peer

(marcelklehr) [0000] Remote peer echoed expected data back

(marcelklehr) ```

(marcelklehr) works :)

yep, same on my side

so its just beaker for sure

(marcelklehr) yep

(marcelklehr) :(

that is what I get when I try to use dat-next clone on it. Known issue?

and "old dat" says its an invalid link when i run `dat b7777e3c824a8b8fb29c6c6187843aa4c807d87452c695c262a590f80c196e2e test`

well technically I ran it with --temp

hmm, re-tried with old dat and it worked, no clue what I was doing wrong

notwes: oops forgot I did that breaking change in dat-next. The version on npm is kinda stale. I'll release much fresher dat-next this week. Not sure if its worth patching that old one.

notwes: didn't totally follow - but dat doctor uses links that aren't compatible with archives. so you can dat doctor with an archive hash. Its just a random hash so two peers can connect

jhand: cool, no rush. Was just playing with marcel's project. And I was just not clear what dat doctor did, but I am now

and we are working on some networking debugging tools so we can start systematically figuring out the beaker connection issues

jhand: is there a way to add usage descriptions to the subcommands? it would be nice to be able to run `$dat-next doctor --help`, and if there is I can submit a PR

notwes: no but that would be sweet.

jhand: yeah, that is one of the reasons I use commander for my cli's. Each of the sub commands can define its own full set of docs and usage instructions

notwes: ah cool. I'll check that out. If not I think I can hook up something basic with subcommand that you can PR to

jhand: cool! doesn't matter to me, but the help flag on each command would be great

and I am happy to contribute based on my noob expirence

notwes: Defnitely. I kept wanting to put more in the main usage but it was getting too big to print and read easily

jhand: yeah, man pages are nice for that as well, but harder to build and distribute. I think just breaking out all the sub commands would do the trick here

oh looks like the arg parser we are using does have a way to do help strings.

nice!

ogd: published your doctor updates + added you to npm

I like the publicbits.org domain

that's cool

^^ me too. but the web design could use some work ;)

"502 Bad Gateway" could use a google web font or something

pfrazee: is there a way to clear the dat cache in beaker? couldn't find one. If not, I can open an issue for that if you like

notwes: there's not, and I think I have an issue for it already

https://github.com/beakerbrowser/beaker/issues/144 it's a part of this

pfrazee: cool, thanks

hmmm I get a ton of random connections with dat-doctor if utp=true but not if its false. Is that expected?

Right now dat-next (and I think main dat CLI) runs doctor with utp: true but thats not the default for the module itself.

Huh. That option isn't even passed through anywhere.

jhand: there might be a bug related to the two swarm instances in client.js using the same port

jhand: cli.js

ogd: looks like it is destroyed first

Im just getting lots of random connections regardless. Options weren't causing it