01:35:13canuckyMe again - still unclear on what I should do to resolve this issue ... we used to have a landing page at 'oet.tru.ca' - this landing page had a link pointing to oet.sandcats.io:6080 ... made the HTTPS changes, which results in former URL oet.tru.ca giving the DNS TXT message
01:37:19asheeshOtherwise I'd "just" suggest you add some DNS TXT records.
01:37:51asheeshIf this server runs nginx, then yeah, I'd set up something on port 6080 that redirects people to a different URL. I'd use nginx for that, and I can try to provide such a config if it helps.
01:37:58asheeshAlso good morning, great to hear from you canucky
01:38:31canuckySure - looking for whatever approach may be best
01:38:58canuckyGood afternoon :) ( logged in from the Atlantic Timezone )
01:39:34asheeshYeah, I'd go with the nginx (or Apache) redirect approach. If you need help with it, just ask, and I am happy to provide.
01:39:57asheeshCrucially it sounds like you'll be happiest if Sandstorm does *not* listen on port 6080.
01:41:54asheesh(You can move Sandstorm to port 80 and port 443!)
01:46:59asheeshOK, give me ~ 5 min. Just making sure I know how names<=>usernames are mapping, you're Grant, right? If so then I have your video chat info from last time we talked.
01:56:31mnutt__I don’t suppose anyone has looked into making sandstorm’s web UI work in offline mode? It seems like it may conflict with the iframes implementation?
01:59:22asheeshI would *love* to see that, and I think it's an architectural question to ask kentonv about.
02:10:45zarvoxYeah, that sounds like it'd need ServiceWorkers, which would need stable hostnames, and also a way to sandbox ServiceWorkers.
02:11:41zarvoxdwrensha: cool to hear that we no longer need AlignedData with the dummy field!
02:26:57asheeshcanuky: Also if you're going to go the static publishing route for doing the redirect, you probably do want port 6080 bound, but you do have the two options anyway.
03:49:56asheeshDo you have other programs listening on port 80?
03:50:44asheeshzarvox/kentonv: Is there any good advice for going from "stack: 0x4f1312 0x4f12ca 0x473c1e" into actual line numbers? Should I trust line numbers from a run-bundle crash?
03:52:04kentonvasheesh: addr2line, but for non-debug builds it won't be able to give you line numbers, but there's a flag (-f?) that will at least give you function names
03:59:00asheeshcanucky: I have to handle something unrelated over the next 30 minutes, but I do want to make sure we get to the bottom of this. So I'll ping you in ~30 minutes and we'll see where we are. I'm also busy 1p-1:30p but can make time outside of that however is needed.
05:35:20canuckyQuestion about DNS TXT entries ... I have a Ghost site at 4zeu8cyfcqyju4xwggw0.oet.sandcats.io
05:35:50asheeshchezmichel: Hi! OK so are you willing to share your sandcats hostname with me, first things first?
05:35:57asheeshcanucky: Great, then what's the question? : P
05:36:16asheesh(but also the Davros app has really really good directions on setting up DNS TXT record stuff, so I would actually try to get a Davros grain working first)
05:36:22asheesh(props to mnutt for making something massively usable IMHO)
05:41:47asheeshSo it looks to me like you have another web server installed as well. Maybe we can figure out how to have that server play nicely with Sandstorm.
05:42:05asheeshWhat is the value of PORT=... in /opt/sandstorm/sandstorm.conf ?
05:42:18asheeshAnd what is the value of BASE_URL=... in /opt/sandstorm/sandstorm.conf ?
05:42:44asheeshTypically Sandstorm will set itself up on port 6080 when there's another web server, but it looks like port 6080 is firewalled on that computer, so no one can reach Sandstorm.
05:45:37asheeshGreat - so then probably it thinks it has port 443 & port 80 available, but isn't reachable over the Internet, and needs some ports forwarded
05:46:00asheeshFWIW in your case I would start by disabling HTTPS and making HTTP work OK, since that's easier to debug.
06:02:12canuckyBut still getting the DNS TXT message ghost.networkeffects.ca ...
06:02:56asheeshcanucky: That should do the trick, but your local DNS resolver might have cached the absence of the entry if you ever missed it.
06:03:07asheeshGiving it ~30 minutes should typically be enough.
06:05:31canuckyok - will give it some time ... one more question - have a blank index.html file in Davros, but still not getting the contents listed
06:05:35asheeshHaving said that, it is kind of sad that people run into this all the time.
06:06:03asheeshA blank index.html really should do the trick, so long as it's non-empty. Would you be willing to share the grain with me and I can take a look? You can type @asheesh in the share box and it should autocomplete.
06:06:07canuckystill trying to resolve that oet.tru.ca redirect issue
06:39:29asheesh1. All API tokens get used on the same hostname (api.yoursandstorm.example.com) at the moment.
06:39:46asheesh2. The bearer token ("Authorization: Bearer asdfasdfasdfasdfa") is what routes, as well as authorizes the request.
06:40:20asheesh2.a. This strikes nearly everyone as crazy but one subtle upside is that it means there's really just one thing, not two things, so there's less room for error.
06:40:51asheesh3. CalDAV and CardDAV are treated as "the app exporting an API" to HTTP clients that are not web browsers.
06:41:13asheeshReally the definition of API in Sandstorm is, "HTTP requests whose response is not going to be rendered by a browser's normal text/html response processing"
06:41:22asheeshso far soo good synchrone ? If not, then ask questions now because it gets crazier soon.
06:43:40asheeshIf someone knows that you're using CardDAV on api.oasis.sandstorm.io and you're unlucky enough to have entered the basic auth thing into your browser, and therefore it remembers it,
06:45:31synchronebut not see the response... but post might be modifying\deleting
06:45:56asheeshRight, yeah, modifying & deleting is already pretty bad.
06:46:21synchroneyou'd have to be unlucky enough to auth your browser for Basic... which is wrong but you think it should disallow users by design from harming their own security
06:56:34asheeshThat and approximately no one thinks "Hey, what HTTP needs is special features that we will carefully refuse to implement in web browsers" (-:
06:57:41synchronewebbrowsers wanna be first-class platforms, and they have the right to
06:57:54asheesh"well then they can fix their confused deputy problems and then we'll talk"
06:58:01asheeshBut yeah, indeed, HTTP is the web browser protocol.
08:39:37zarvoxfrigginglorious: "same as any other grain on Oasis" AFAIK, which I think is basically "however much a worker machine has". I don't think we currently cap memory usage, though we do plan to account/bill for it.
08:41:18digitalcircuitzarvox: in context of the original question, does that include billing for memory usage of app demos?
08:45:14friggingloriouszarvox: Thanks for that. I have a feeling my digitalocean server could go down if a bunch of peeps log in during a live demo ;)
08:47:06zarvoxdigitalcircuit: we don't have anyone to bill in particular for app demos, so doubtful
08:47:12friggingloriousI wanna see how many people i can get logged into a Groove Basin session to simultaneously stream that song that Darude about the software
08:48:43zarvoxIf you're likely to write a very memory-hungry app and post an appdemo link expecting a lot of traffic, a heads-up would not go unappreciated ;)
09:19:39jadewangcan I quote excerpts from your HN comments in a blog post?
09:35:00mnutt__I keep on meaning to work on the etag / PATCH changes, but have been procrastinating hoping to get some test coverage in place for sandstorm *dav support. I'm thinking of packaging a tiny webdav server spk and then hit it with webdav commands via selenium. Any reason this wouldn't work, or better approaches?
09:36:05zarvoxThis is pretty good, and approximately what we do in our integration tests.
09:36:56zarvoxIt would be amazingly great if we could integrate it into sandstorm/tests/ , so we can make sure we avoid regressing, since we run those tests for every PR. :)
09:40:52mnutt__that would be my plan. it's a little odd that the tests just pull spk's down off apps.sandstorm.io, but I guess there isn't anything wrong with it
09:41:57zarvoxSome tests even install packages from a Davros grain's static publishing folder on Alpha!
09:42:37zarvoxActually, hmm. Were you planning to make the webdav command requests from the browser, or from the node-that-drives-the-browser?
09:42:57zarvoxand would the tests themselves live in the spk, or in sandstorm/tests?
10:18:51friggingloriousI just followed the tutorial and packaged a simple diceroller app I wrote for D&D.
10:19:41friggingloriousIts incredible how easy working with sandstorm is. Seriously amazing.
10:41:44mnutt__zarvox: the tests would be in sandstorm/tests, I guess if it could be done from node-that-drives-the-browser it'd be a bit more efficient
13:01:10friggingloriousI'm titling my talk: Utilizing Open Source Web Apps Painlessly with Sandstorm.
14:50:11friggingloriousneynah: Thank you, I will be grabbing some of those background images. I was already a fan of Reveal js, but Hacker Slides is far too nice of a tool. Couldnt think of using Google Docs, especially for this!
14:52:16neynahNp frigginglorious. Yeah, I really should move the slide BGs to a Lychee grain or something more reasonable. :)
19:25:50synchroneso in this code, both actions launch .caldavCommand, but if I switch their positions in the actions = [] array - both options start to launch .carddavCommand
![https://vyb4th03506cdh13m3bw.oasis.sandstorm.io/abstractlogo-black.png"](https://vyb4th03506cdh13m3bw.oasis.sandstorm.io/abstractlogo-black.png"){kind=link}