<mnutt>
zarvox: it definitely shouldn’t show .jsdav, that’s a bug
<zarvox>
should I file an issue on github, or have I said enough?
<mnutt>
asheesh: it sounds like you’re trying to sync with the release folder I shared earlier, and that’s a UX fail on my part. for permissions there are currently only “view” and “edit”, and “view” doesn’t allow syncing. It should hide the clients tab when you’re in view-only mode, I jsut haven’t gotten to it yet.
<mnutt>
a github issue would be great, thanks
<mnutt>
asheesh I’d be really interested to see what the Internal Server Error is, is that inside the grain or from sandstorm? is it complaining about statvfs.blocks?
<mnutt>
zarvox: I’m actually really curious what’s in the .jsdav dir; jsDAV stores metadata in that directory but as far as I know I’m not actually using it for anything, so I wonder what is getting stored in there. I can certainly hide it from the listing though.
bb010g has joined #sandstorm
funwhilelost has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
mnutt has quit [Quit: mnutt]
jadewang has quit [Remote host closed the connection]
luckre has quit [Quit: No Ping reply in 180 seconds.]
luckre has joined #sandstorm
natea has joined #sandstorm
jadewang has joined #sandstorm
jadewang has quit [Remote host closed the connection]
paroneayea has quit [Read error: Connection reset by peer]
mnutt has joined #sandstorm
jadewang has joined #sandstorm
dwrensha has quit [Remote host closed the connection]
dwrensha has joined #sandstorm
jadewang has quit [Remote host closed the connection]
jadewang has joined #sandstorm
uppermgmt has quit [Ping timeout: 272 seconds]
uppermgmt has joined #sandstorm
rhapsodhy has quit [Remote host closed the connection]
<asheesh>
The 60 second "Session never opened" timeout just caused me some data loss (created a wekan board, added some lists, then lost 'net connectivity for ~2 min, then got a force-reload) so I'm super-duper excited about the new 15 minute timeout.
<asheesh>
I also had added some cards in that time, I'm now remembering.
<dwrensha>
that timeout got upped from 15 seconds to 60 seconds
<asheesh>
o
<asheesh>
I mis-read it, I guess, yeah.
<asheesh>
Not released yet, but yeah, maybe 60s will fix my problem. But if I really lost connectivity for 2 min, maybe not.
<asheesh>
The thing that sucks the most is that seemingly the subdomain changes, so the app can't recover. I mean, I realize that's the meaning of creating a new session, but tis sad.
<asheesh>
"Not released yet" meaning "I think this commit is not on Oasis at the moment"
<kentonv>
asheesh: did wekan refresh before or after your connectivity was restored?
<asheesh>
"refresh" meaning "the iframe's URL changed"?
<asheesh>
I am really really not sure I remember.
<asheesh>
Here's what I remember.
<asheesh>
* I had stuff in my board.
<asheesh>
* Connectivity was flaky on and off.
<asheesh>
* WeKan showed me a visual saying (approximately) "Board does not exist. Maybe it's a private board. Try signing in"
<asheesh>
* Connectivity was restored
<asheesh>
* WeKan reloaded itself entirely, and this time, had none of the data I'd inputted
<asheesh>
I then looked at the JS console
<asheesh>
I could have sworn in the JS console I saw some indication that the subdomain had changed, but now I can't clearly find any indication of that.
<asheesh>
The thing is that it may have refreshed itself before connectivity was restored, too.
<asheesh>
While I was looking at a different virtual desktop on my laptop, say.
<asheesh>
I'll make a filedrop with my JS console in case it's useful
<asheesh>
Honestly I want Richard C's filedrop alternative app! But anyway
<kentonv>
it doesn't make sense for it to have changed subdomains and then ended up with the "session never opened" error immediately after that
<asheesh>
The current IFRAME URL subdomain is the same as the only subdomain I see in the console log, but I don't know if that's just because the log is too recent.
<asheesh>
But yeah, I have no particular evidence at the moment that the subdomain changed.
<kentonv>
too bad you probably don't have a network log
<asheesh>
I would opt in to Sandstorm/Oasis tracking 100% of my traffic in a pcap dump or anything else for debugging purposes.
<asheesh>
And/or install a browser add on to log 100% of my Sandstorm traffic in plaintext on my laptop.
<asheesh>
I guess I can surf Sandstorm with (X) Preserve log enabled, within Chromium Network tab
jacksingleton has joined #sandstorm
<asheesh>
Sadly that log stops being recorded if I (X) the Chromium debug window thing
<asheesh>
Howdy jacksingleton
<kentonv>
I suspect, though, that the 60-second timeout will fix it, because I think what's happening is that after network connectivity was restored you were unlucky enough to have your next keepalive scheduled more than 15 seconds in the future
<jacksingleton>
Howdy team. 1: looking forward to tonight! 2: We hit publish on SandForms last night, I'll be hanging out here so if there are any problems let me know!
<asheesh>
1: Sweet. 2: omg whoa cool
<jacksingleton>
Hola Asheesh
<jacksingleton>
:)
<kentonv>
(because the keepalive is every minute, and since your client never actually stopped running as it would during a suspend, so it didn't move the keepalive forward...)
<kentonv>
it still seems wrong that sockjs would force a refresh on a 500, though. Maybe the error code needs to be specifically 502?
<kentonv>
pretty sure 502s will not cause a page reload, otherwise we'd see them every time Sandstorm updates.
jadewang has joined #sandstorm
<asheesh>
kentonv++
<asheesh>
jacksingleton: Maybe I should use Sandforms instead of Framadate. I'll try it once it's on the app market.
<kentonv>
In other news, I think I have confirmed that in the event loop stall, all newly-created connections are broken. The only reason the server continues to accept HTTP requests at all is because nginx is pipelining them on an existing connection, but it cannot pipeline websockets.
<jacksingleton>
asheesh: b
<kentonv>
But I spent a long time reading the libuv code last night without discovering anything interesting.
jadewang has quit [Ping timeout: 250 seconds]
<asheesh>
jacksingleton: Also, is it: sandforms, SandForms, sandForms, SandForms?
<asheesh>
(or s-and-forms just kidding)
<jacksingleton>
asheesh: SandForms
<asheesh>
+1, I will use that consistently then.
<kentonv>
jacksingleton: typo in description: "onlinewhile" -> "online while"
<jacksingleton>
kentonv: aw, thanks. anything else? if not will republish in a few minutes
<asheesh>
(make sure to bump appVersion when you do so)
<kentonv>
asheesh: no, since I haven't approved the existing one it's not really necessary
<kentonv>
(but doesn't hurt)
<asheesh>
Ah K
<kentonv>
jacksingleton: hmm, I feel like a screenshot containing more sample data would more clearly show what Sandforms is
<asheesh>
jacksingleton: Pseudo-random meetup planning thoughts for you:
<asheesh>
* We should make sure there are name tags. Do y'all have some name tags we can use for that at ThoughtWorks? If so, yay; if not, I should be sure we get some.
<asheesh>
* It'd be cool to encourage people to label themselves with "Self hoster" or "App maker" or "Sandstorm user" or "New to Sandstorm" or "Core developer". I figure I can get some stickers for that. I think it would encourage better conversations.
<asheesh>
* Anything I should know about food planning? Last time seemed to be a huge success so 'I [Jack] have it covered' is a fine answer.
<asheesh>
* I can move this to email if that's smarter.
<asheesh>
* Amy DiPlacido who's been helping with other Sandstorm things is available to help with tech checks for lightning talk speakers, and I'm happy to delegate that to her, so I did that.
<dwrensha>
i just want to say: I love saag paneer burritos.
jadewang has joined #sandstorm
bb010g has joined #sandstorm
<asheesh>
jacksingleton: Also if you always load JS/CSS assets from your app's domain, does that make SandForms a "handcrafted single-origin web app"?
<asheesh>
(I guess "no" since "origin" doesn't refer to the domain you are pulling assets from. Oh well.)
funwhilelost has joined #sandstorm
jacksingleton has quit [Ping timeout: 252 seconds]
offby1 has joined #sandstorm
<offby1>
So ... here's an odd one.
<offby1>
I've written a simple app, and have got it working nicely via "vagrant-spk dev".
<offby1>
So ... I packaged it up with "vagrant-spk pack", and tested that the package works locally.
<asheesh>
So far so good.
<offby1>
Then I uploaded that same package to oasis.standstorm.io, and it fails to start, with ``pkg_resources.DistributionNotFound: SQLAlchemy'' in the log.
<offby1>
whiskey tango foxtrot
<asheesh>
Oh my.
<offby1>
where would I start looking to troubleshoot?
<asheesh>
Are you sure you've ^C'd the vagrant-spk dev?
<asheesh>
And have you tested creating a _fresh_ grain locally via the SPK on Oasis?
jacksingleton has joined #sandstorm
<offby1>
asheesh: I did _not_ Control-C the vagrant-spk dev
<offby1>
maybe I shudda
<offby1>
and I don't understand your second question -- to me, "locally" and "on Oasis" are opposites
<asheesh>
Yeah, if you don't Ctrl-C the vagrant-spk dev, then sort of all bets are off as to if you're testing the SPK package or the "Dev package"
<offby1>
ah
<offby1>
ok, let's try that
<offby1>
pretty slow :-|
<maurer>
Is rocket chat supposed to work for anonymous users?
<maurer>
(someone just tried to use my rocket chat instance and got a blue screen)
<offby1>
maurer: I could never figure out how to let anyone but me use it :-(
<maurer>
offby1: I've got me and one other registered user using it
<offby1>
presumably not the Blue Screen Of Death
<maurer>
it's the unregistered user who's having a problem
<asheesh>
The app package should presumably ask you really hard to sign in, rather than break in half.
<maurer>
The unregistered user registered and doesn't have an issue nwo
<maurer>
*now
<maurer>
so it appears that you do need to be registered, but that the app somehow fails to tell you to ask for a server invite
<asheesh>
Turns out (and this surprises basically everyone except kentonv) that "Signing in" to an app does not require an invite.
<asheesh>
However, "Creating grains" does require an invite.
<maurer>
Huh, OK. Well, it should encourage them to sign in :P
<asheesh>
The distinction here is that "Signing in" does not confer any extra authority to you; it "just" gives you a name.
<offby1>
perhaps a change of terminology would help then
funwhilelost has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
<offby1>
asheesh: hmph, I think this time I "uploaded" my package to my local vagrant box correctly -- I don't have "vagrant-spk dev" running -0- and it still works
<offby1>
weird weird weird
<asheesh>
offby1: And did you test creating a new grain, rather than attaching to some existing grain?
<offby1>
yep
<jacksingleton>
kentonv: Just hit publish again, added a better screenshot and fixed the description
<asheesh>
...huh.
<offby1>
asheesh: I deleted all grains and installed apps first.
<offby1>
I'm pretty compulsive about resetting things back to zero if possible
<kentonv>
jacksingleton: approved
<asheesh>
Unfortunately I have to AFK but I can try to look into this more soon, later today or similar.
<offby1>
I guess I'll have to ssh in to the box on oasis, if that's possible, and do a bunch of tedious comparison of log files
<jacksingleton>
I think the typoes in the description were due to the fact that it was hard wrapped, which I wouldn't have expected to make a difference for a markdown file
<asheesh>
BTW offby1 ... you don't happen to based in SF, do you?
<asheesh>
Maybe one day you'll give a talk at some Seattle GNU/Linux Users Group event!!
<asheesh>
Cheerio for now. Must eat.
<offby1>
maybe
<offby1>
eat is good
<kentonv>
jacksingleton: hmm, if it's a bug, it's a bug specifically in my admin view and probably wouldn't have been a problem when published.
<kentonv>
jacksingleton: github is being really weird about displaying the history of that file. It doesn't want to show me the additional lines that were presumably there before your latest change...
<kentonv>
jacksingleton: I think the file may have had CR (\r) rather than LF (\n) line breaks.
<jacksingleton>
:o
<kentonv>
jacksingleton: yeah, hex dump shows 0x0d line breaks rather than 0x0a
<jacksingleton>
well then
<jacksingleton>
Yup.$ file description.md
<jacksingleton>
description.md: ASCII text, with CR line terminators
<kentonv>
what did you use to write that file?
<jacksingleton>
good catch
<jacksingleton>
I'm asking myself the same question :)
<jacksingleton>
it had to be vim...
<jacksingleton>
ooh "xsel -b > description.md"
<jacksingleton>
was me copying the text from our trello card
jadewang has quit [Remote host closed the connection]
<paroneayea>
btw, is anyone currently using sandstorm for their main site hosting?
<paroneayea>
for like, their domain of their personal site or organization?
simonv3 has quit [Quit: Connection closed for inactivity]
<offby1>
paroneayea: why does your nick sound familiar ...
<paroneayea>
offby1: I'm Chris Webber of GNU MediaGoblin
<offby1>
that doesn't sound familiar at all :-| perhaps you used to hang out on #emacs
<paroneayea>
offby1: yes I did!
<offby1>
that'd do it
<paroneayea>
offby1: learned a lot about free software by spending time in there
<paroneayea>
also accrued about a gigabyte of logs.
<paroneayea>
of mostly people arguing about things that aren't emacs related I think :)
<offby1>
unthinkable
* paroneayea
re-joins, for old times sake
<kentonv>
paroneayea: I know ocdtrekkie uses Sandstorm for his blog, FWIW. And of course our own blog.sandstorm.io is Sandstorm-hosted. (I imagine we'll switch the main site over in early 2016, once Oasis supports zero-downtime updates.)
<paroneayea>
kentonv: thanks!
<kentonv>
stats suggest that a number of people are using wordpress and ghost on oasis, but I of course don't look into what exactly they're hosting unless invited
<offby1>
kentonv: I might as well repeat an oddity I saw, that asheesh had no suggestion about: I've written a simple app, and have got it working nicely via "vagrant-spk dev". I packaged it up with "vagrant-spk pack", and tested that the package works locally. Then I uploaded that same package to oasis.standstorm.io, and it fails to start, with ``pkg_resources.DistributionNotFound: SQLAlchemy'' in the log.
<offby1>
this stumps me
<kentonv>
that's pretty strange, the environment should be almost indistinguishable
<kentonv>
one possibility is that one of your dependencies runs different code depending on the CPU properties
<kentonv>
another is that slight timing differences are somehow landing you on a different code path
<kentonv>
for a lot of the python-based apps we've ported, we've used virtualenv to gather the exact set of dependencies, then specified alwaysInclude in the pkgdef to force the entire thing to be included in the package regardless of whether the files were used at spk dev time
<kentonv>
this could help if the problem is that some dependency modules are only being used on Oasis and not locally, for some reason
rustyrazorblade has joined #sandstorm
<offby1>
aaah
<offby1>
... and yet the sqlalchemy stuff -- at least some of it -- appears in sandstorm-files.list
NOTevil has joined #sandstorm
jadewang has joined #sandstorm
<kentonv>
offby1: it's entirely possible that some stuff made it in but not everything.
<offby1>
yeah, depending on what happened to get opened when I ran it under 'vagrant-spk dev', I suppose
<maurer>
Where do I post bug reports for individual sandstorm apps?
<asheesh>
maurer: On the git repos for the individual Sandstorm apps
<maurer>
Hopefully the report bug thing actually goes to the right people, and I don't have the mainline rocketchat devs going "wtf is a sandstorm, and why are their users sending me bugs" :P
<asheesh>
Yay
<asheesh>
BTW if you file bugs against RocketChat on Sandstorm maybe I can find someone to write a patch at the upcoming meetup (-:
* offby1
pictures asheesh carrying a large mallet
<asheesh>
: D
<asheesh>
offby1: BTW you'll want to "alwaysInclude" the virtualenv directory.
<asheesh>
If the Python stack doesn't do this for you, then that is bad news and we should fix that.
<asheesh>
But also...
<asheesh>
are you _sure_ you're experiencing this crazy-seeming problem? Yow. I'd love a copy of the SPK if you'll share it with me to test too.
NOTevil has quit [Quit: +++ ATZ]
Arcalinea has joined #sandstorm
<asheesh>
brb dealing with voting
Arcalinea has quit [Client Quit]
bb010g has quit [Quit: Connection closed for inactivity]
<offby1>
I bet I have the SPK somewhere. Where should I "put" it?
<asheesh>
You could make a FileDrop!
* offby1
stares blankly
<offby1>
is that one of the apps that runs on sandstorm?
<asheesh>
Yeah (-:
<offby1>
figured
<offby1>
then I'd have to figure out how to let you at it :)
<offby1>
seriously, is that the most convenient way? I could put it on Dropbox too
<asheesh>
I can also make one and you can add it there