#dat

mafintosh: 2 messages? why the second one?

pfrazee: one for each hypercore feed in the dat

was the easiest eay impl it right now

mafintosh: why not group them together? Combine the byte count, spit out the meta hash

just cant imagine needing the content core's key, but I can imagine having to explain to people multiple times why there was 2 messages

pfrazee: yes good idea. thats why i said that this was just the easiest way to impl it right now

mafintosh: ok cool :)

going to bed now; but if someone feels like debugging something - I still can't figure out why this line contains legit DOM elements (if files are added) but they never show up in the UI https://github.com/yoshuawuyts/dat-desktop/blob...

so far my best guess has been: chrome bug - but a fresh pair of eyes would do miracles I reckon

yoshuawuyts: you try putting it in a tbody?

yoshuawuyts: also, try wrapping the <th> elements in a tr, so that it's thead > tr > ths

pfrazee: lol, no - let's give it a shot

pfrazee: I def like the markup better now, but it's not the solution

yoshuawuyts: :(

pfrazee: thanks for the thinking tho - for a moment I was like: "ohhh, wish I would've known about tbody sooner!"; at least I learned a thing now haha :D

yoshuawuyts: yeah :) did you try the tr in the head?

pfrazee: yeah, that's what you suggested right?

yoshuawuyts: that and the tbody

(pushing new markup now)

HTML is generally forgiving for stuff like this, but yoyo can be weird about markup things

(markup pushed)

yeah perhaps the solution would be to not use a table?

you might try it, if just to see whether it renders

yeah might just do that - have spent a whole lot of time debugging this today; feel it might like actually be a chrome bug - not using tables could be the best / least hacky solution

pfrazee: thanks for thinking along!

yoshuawuyts: no prob! always happy to help out with a tough debug sitch

ok now actually off to bed haha :D - night!

night

yoshuawuyts: seems like it may be something with rendering + data updating. If you give it data hardcoded it works okay (e.g. https://gist.github.com/joehand/52adba4dc7cf122...)

mafintosh: Do you know anything about sodium.crypto_stream. I'm very confused as to what it does

emilbayes: nope

mafintosh: There's noting on google about it either. I'm thinking it might be something to do with one-time-pads but there's no mention of it's purpose anywhere

"Unless you specifically need unauthenticated encryption, \Sodium\crypto_secretbox() is the operation you should use instead"

emilbayes: o/ from docs

so secret key enc without a mac?

mafintosh: Yeah, read that, I am going to use the crypto_stream_xor

mafintosh: Oh, but look at it. It will give you a buffer(length) encrypted with nounce and key. But what's encrypted? Garbage data?

mafintosh: Maybe it's for overwriting something that was previously encrypted with crypto_stream_xor with plausible data

yea dunno

secretbox sounds "safer"

mafintosh: Yeah, but I wanted to try and do some random access stuff ^^

what does that mean?

mafintosh: With secret box you have to decrypt the whole thing. With crypto_stream_xor you can decrypt at any point of the buffer (I think)

mafintosh: But for it to be secure there would still have to be some kind of authentication somewhere

ah. i've been using varint prefixed secretboxes

varint containing the length

over a stream

mafintosh: I was thinking more of a abstract-random-access interface to an encrypted file

basically just dominic's secret-handsake procotol but without the handshake part lol

ah cool!!

mafintosh: But may be moving into stupid dangerous territory here

yea dunno, i'm no expert

make something. lable it secure. announce on twitter. wait for the world to tell how wrong you are

mafintosh: Haha, that's the spirit!

and ask dominic

mafintosh: Yep! He's not on irc so it's not as convenient as just googling :p

oh so its a util for generating pseudorandom bytes

yoshuawuyts: it seems like an array of objects instead of a single HTML element

yoshuawuyts: yeah jhand has a point make sure updateArchives reducer is updating the archives

yoshuawuyts: i read that choo article. good stuff. i understood most of it

[dat] joehand closed pull request #558: fix usage print out on empty stdarg. fixes #557 (master...fix-empty-stdarg) https://git.io/vXJvI

[dat] joehand pushed 1 new commit to master: https://git.io/vXkCQ

dat/master 86b7e47 Joe Hand: 11.5.3

oh

jhand: anything to test in dat-cli?

TheLink: nah just doing minor bug fixes

TheLink: Thanks though!

ok

[dat] joehand closed issue #551: Question: does dat support branching and merging between different forks https://git.io/vPQzD

mafintosh: yayy

jhand: yeah I experienced that too; I'll dig down that path a bit further

yoshuawuyts: you solve that bug yet? (Im curious about it)

Hi! I read about PublicBits at http://dat-data.com/blog/2016-02-01-announcing-... and I'm wondering what relationship it has with https://github.com/karissa/dps , if any.

Hi pdurbin! That was a prototype version done for the grant application I think. We are still pursing the same idea (registry for data) but UI will be a bit different.

jhand: ok, thanks. I think you're going to be on my floor at 2. :)

pdurbin: I'll actually be at Harvard this afternoon (been corresponding with Danny) and happy to answer more questions.

Yes =)

jhand: can you please take a look at https://github.com/IQSS/dataverse/issues/1838#i... ?

pdurbin: Sure!

thanks

jhand: any thoughts? Is requiring an API Token a show stopper for including Dataverse installations in PublicBits?

pdurbin: ya, sorry my internet died. I think it's not a complete blocker now. That original idea was more user driven requests closer to one time http requests. But now I think we'll be doing something closer to centralized harvesting.

So easier to manage tokens

jhand: so something more like https://share.osf.io/discover?q=* which uses a Dataverse API token?

pdurbin: yep, that seems right.

cool

jhand: while I'm making noise, can you please look at https://twitter.com/denormalize/status/73154237... too? I want to let you and ogd know about something new.

pdurbin: you have my curiosity peaked

jhand: the new thing is http://dataverse.org/metrics which is powered by https://github.com/IQSS/miniverse and there are API docs at https://services.dataverse.harvard.edu/static/s...

pdurbin: cool! Still on phone but I'll check it out more when internet is back.

thanks :)

Ah much more interesting on my computer! (there was just one graph on mobile)

pdurbin: a few ?'s: who determines the affiliation? The monthly stats are across all the dataverses? Kinda surprised to see those not go up with a much higher number of installations.

Did y'all do something for the big jump in installations over the summer?

jhand: for the affiliation thing, could you please ask at https://github.com/IQSS/miniverse/issues ?

sure

I guess the total numbers are from the dataverses on Harvard's portal (since the 1.8k dataverse number matches), not groups that install it themselves.

jhand: the "miniverse" app sits on top of the postgres database of a single Dataverse installation. So what you're seeing is only metrics from Harvard's installation of Dataverse. You don't see other installations.

jhand: we can bring this to #dataverse on freenode if you want :)

pdurbin: okay, that makes sense! I'll hop on there but I think I just needed to browse around a bit more and got most my ?s answered.

awesome, hope to see you at 2

context is scholarly libraries

pfrazee: not yet; gonna instrument choo now; feel u might have missed something

S/u/I/

cool

ogd: if that isn't a resounding ad for dat

ogd: wheres that pic from?

dat cli is acting kind of fubar for me

Im on latest

I host in one cli tab, I download in the other, and it just doesnt complete

the file and byte counts are all incorrect

as is the progress bar

downloading will just stop at random points, but files will be missing

might be caused by downloading a dat created in an old version...

pfrazee: it's broken since quite some time

TheLink: the counts have been a little bit off before, but usually things worked overall