<frigginglorious>
im gonna have to do some thinking… Im gonna see how much of sandstorm I can keep entirely offline
<frigginglorious>
including dev tools. It might be rough going. And since theres no way to authenticate otherwise, im pretty sure everyone has to have a passwordless dev account, right?
<dwrensha>
we've gotten requests for the preinstalled-app feature before
<zarvox>
you could set up a local LDAP server or (I don't wish this on anyone) a local SAML server, with a Sandstorm for Work key. But otherwise, local email and dev accounts are the easiest "totally self-hosted" login mechanisms.
isd2 has quit [Remote host closed the connection]
<frigginglorious>
wait a tick… if someone gets access to a shared grain, do they then have permissions to spin up a new one of the same app?
<dwrensha>
no
isd2 has joined #sandstorm
<isd2>
Aha! irc-idler is doing something not totally useless!
isd2 has quit [Client Quit]
<frigginglorious>
ohhhh. hm… i may just end up having a site with downlaods for all the .spks. but that seems silly
<isd>
(namely, proxying and repsonding to pings). Sorry for the outburst
<frigginglorious>
isd: apology accepted.
<isd>
:P
<dwrensha>
it'd be cool if there were something in the admin panel where you could select some apps, and all new accounts would automatically have those apps installed
<frigginglorious>
dwrensha: heck yes it would.
<frigginglorious>
would all that stuff be in C++?
<dwrensha>
I think the only nontrivial part would be implementing the UI in the admin panel
<dwrensha>
this would all be javascript
<frigginglorious>
i’m afraid it’s been awhile since i C++’d
<frigginglorious>
oh, well then that’s good news!
<isd>
Is there something similar to sandstorm-http-bridge but for raw network connections? I'd like to rig up my app directly to Sandstorm's network API, but it might be good to have something usable before I've wrangled all of the API gluing myself.
<zarvox>
isd: I think jparyani had a tcp bridge at one point; not sure if it's published anywhere
<zarvox>
Hmmm, I wonder how hard it would be to write a program that takes an IpNetwork, host, and port as command-line arguments and otherwise is like netcat.
<isd>
zarvox: Beyond just getting talking to the IpNetwork object in the first place, I expect not very.
<asheesh>
I haven't personally looked into the golang capnproto implementation.
<isd>
I suspect code size tends to large because go-capnpc generates more types that it otherwise needs to get type safety -- no generics, so you get a List_* for every * (though if you want nested lists you have to cast).
<isd>
but I haven't dug into it
<isd>
Go binaries tend to be big, but wow.
<asheesh>
I don't suppose now is a good time to try to convince you to learn Rust, eh? (-;
<isd>
That's not actually causing problems at hte moment. Just O.O
<asheesh>
Ah, good (-:
<isd>
12 MiB stripped, and I haven't looked at the SPK but that will be smaller (the executable is the only thing in there).
<isd>
(the whole executable is pretty much dominated by that one lib.
<isd>
)
<isd>
asheesh: rust is on my TODO list.
<isd>
but probably not this project.
<asheesh>
Nod (-:
<isd>
Right now working on the sandstorm-independent functionaltiy. Will pick at the capnp stuff again soon.
<asheesh>
Cool.
<isd>
Fun bit of awkwardness in the design: IRC doesn't actually send timestamps, so if you reconnect to the proxy, it doesn't have a clean way to convey when the messages it's accumulated were sent. I'm thinking of just inserting NOTICEs when there are significant jumps in timestamps while the user is disconnected, maybe on an optional basis
<asheesh>
Yeah - IRC is an interesting protocol.
<asheesh>
ctrlproxy iirc prefixes things with [23:32] or whatever the time was when the msg was rcv'd.
<isd>
Right now it just parses and re-serializes messages, and also auto responds to PINGs
<asheesh>
Which is admittedly a little sad.
<isd>
which is totally enough to make telnet a reasonable IRC client! :P
<isd>
only half kidding
<asheesh>
Maybe a channel NOTICE on day change is minimal and sufficient.
<isd>
asheesh: yeah, you could play with the granularity. I'll probably have experiment to figure out what makes sense.
<asheesh>
Did you still need your question from ~7h ago answered?
* isd
doesn't remember what that was, and can't look at the scrollback because his app isn't done yet
<asheesh>
00:49 < isd> Can someone more familiar with the API tell me definitively if that's something that can be worked around without the full fix?
<isd>
Ah, yes an answer to that would still be useful, if you happen to know.
<asheesh>
I definitely don't but likely kentonv does, and will maybe see this ping when he wakes up.
* isd
looks at clock
<asheesh>
Good thinking.
<isd>
O.O
<asheesh>
See you tomorrow!
<isd>
It's 4 am in Boston. Whoops.
<isd>
kentonv: good chance I won't been in the channel when you see that ^; if you have useful thoughts, I'll see them if you post to the github issue.
<isd>
night all.
isd has quit [Quit: Leaving.]
ecloud_ has quit [Ping timeout: 264 seconds]
ecloud has joined #sandstorm
_iwc has quit [Ping timeout: 250 seconds]
jadewang has joined #sandstorm
jadewang has quit [Ping timeout: 246 seconds]
jadewang has joined #sandstorm
jadewang has quit [Ping timeout: 240 seconds]
jadewang has joined #sandstorm
jadewang has quit [Ping timeout: 244 seconds]
jadewang has joined #sandstorm
jadewang has quit [Ping timeout: 276 seconds]
rgrinberg has joined #sandstorm
jadewang has joined #sandstorm
jadewang has quit [Ping timeout: 264 seconds]
frigginglorious has quit [Quit: frigginglorious]
gelnior has quit [Remote host closed the connection]
gelnior has joined #sandstorm
frigginglorious has joined #sandstorm
gelnior has quit [Remote host closed the connection]
gelnior has joined #sandstorm
jadewang has joined #sandstorm
jadewang has quit [Ping timeout: 240 seconds]
jemc has joined #sandstorm
jadewang has joined #sandstorm
jadewang has quit [Ping timeout: 240 seconds]
jadewang has joined #sandstorm
rgrinberg has quit [Ping timeout: 264 seconds]
jadewang has quit [Ping timeout: 276 seconds]
rgrinberg has joined #sandstorm
[d__d] has quit [Remote host closed the connection]
[d__d] has joined #sandstorm
amyers has joined #sandstorm
jadewang has joined #sandstorm
jadewang has quit [Ping timeout: 276 seconds]
amyers has quit [Ping timeout: 240 seconds]
jadewang has joined #sandstorm
jadewang has quit [Ping timeout: 272 seconds]
jadewang has joined #sandstorm
<frew>
so this is bizarre
<frew>
my feed reader on the oasis gets a 403 from one of the feeds
<frew>
but I don't locally
<frew>
does anyone know what the javascript stuff is about in the log for TTRSS?
<frew>
TTRSS is php so I have no idea why there'd be a JS stacktrace in the logs
<asheesh>
Seemingly we delegate to http://docs.meteor.com/api/http.html which I don't know what headers it sends, but could be no user-agent by default, let me see...
<frew>
so am I correct that if there is no browser the application cannot reach out at all?
<asheesh>
Sandstorm apps' server-side code have a concept of a "session."
<asheesh>
Today, that is a "HackSession" because we don't like the quality of the session.
<asheesh>
The server-side code of the app uses this Sandstorm session to ask questions like (a) which Sandstorm user is viewing this? and (b) hey Sandstorm make an outbound HTTP GET call
<asheesh>
The session expires over time if you're not looking at it in a web browser, but it's not literally a connection to your web browser.
<asheesh>
I hope that helps somewhat. Seems like I should write a doc at some point about what the "Session" is for Sandstorm.