<mnutt_> it’s specific to my setup, as I just tried vagrant-spk with paperwork and everything worked
<mnutt_> it turns out I had modified vagrant to forward 6080 on the host to 6081, I have no idea why that would break anything that way but updating it fixed it.
eternaleye has quit [Ping timeout: 265 seconds]
neynah has quit [Quit: http://www.kiwiirc.com/ - A hand crafted IRC client]
neynah has joined #sandstorm
rustyrazorblade has quit [Quit: rustyrazorblade]
mnutt_ has quit [Quit: mnutt_]
yuvipanda_ has joined #sandstorm
yuvipanda_ has quit [Client Quit]
neynah has quit [Quit: http://www.kiwiirc.com/ - A hand crafted IRC client]
neynah has joined #sandstorm
mnutt_ has joined #sandstorm
neynah has quit [Quit: http://www.kiwiirc.com/ - A hand crafted IRC client]
bb010g has joined #sandstorm
todayman has quit [Ping timeout: 240 seconds]
todayman has joined #sandstorm
Aric has joined #sandstorm
Aric has left #sandstorm [#sandstorm]
neynah has joined #sandstorm
neynah has quit [Quit: http://www.kiwiirc.com/ - A hand crafted IRC client]
bb010g has quit [Quit: Connection closed for inactivity]
mnutt_ has quit [Quit: mnutt_]
neynah has joined #sandstorm
neynah has quit [Quit: http://www.kiwiirc.com/ - A hand crafted IRC client]
uppermgmt has quit [Ping timeout: 268 seconds]
uppermgmt has joined #sandstorm
larjona has joined #sandstorm
jjb is now known as oao
larjona has quit [Quit: Konversation terminated!]
_iwc has joined #sandstorm
xet7 has joined #sandstorm
_iwc has quit [Killed (Sigyn (Spam is off topic on freenode.))]
mnutt_ has joined #sandstorm
<paulproteus> Howdy mnutt_
<paulproteus> I saw your remark here. FWIW, port 6081 is the mongodb port in Sandstorm, so it's semi-understandable that things would go haywire if you try to speak HTTP to that.
<paulproteus> Aw ckocagil I get invalid token at your /shared/ link above.
<ckocagil> paulproteus: lemme reupload
<paulproteus> Yay
<paulproteus> Ooh NodeBB!
<ckocagil> yep!
<dwrensha> ckocagil: what databse does it use?
<dwrensha> *database
<ckocagil> redis :/
<mnutt_> paulproteus: makes sense, thanks!
<dwrensha> only redis?
<ckocagil> yep.
<ckocagil> it can use mongo if you configure
<dwrensha> I'm wondering why it requires 10MB on initialization
<ckocagil> dwrensha: I copy some directories that the app needs to be writeable
<ckocagil> and symlink the rest
larjona has joined #sandstorm
<dwrensha> why do they need to be writable?
<paulproteus> Howdy larjona
<paulproteus> Saw your email; was going to hop on debconf-team and reply
<dwrensha> are they for plugins or logs or something like that?
<larjona> hi
<ckocagil> dwrensha: it modifies the public/ directory for I guess theme stuff
<dwrensha> themes, ah
<dwrensha> so you can cumstomize the theme?
<ckocagil> yeah.
<paulproteus> larjona: If you'd prefer to reply, that's OK with me.
<ckocagil> first person to log in is the admin
<ckocagil> who can change the theme and colors
<larjona> no, in fact, paulproteus, I'm not sure what else to say
<paulproteus> It seems to me that someone (e.g. me) should fix NoScript so it can whitelist *.storm.debian.net.
<paulproteus> ( :
<dwrensha> ckocagil: whoa, I just clicked through to my admin dashboard (on my own grain)
<dwrensha> it opened in a new tab
<ckocagil> yeah I gotta fix that too
<larjona> I don't know very well NoScript nor similar, I thought it was weird that I cannot whitelist *.whatever, but I thought maybe *I* don't know how to do it
<dwrensha> or, rather, it busted out of the iframe
<paulproteus> i,i what we need is the opposite of framebusting
<dwrensha> is "framebusting" the appropriate term here?
<paulproteus> I believe that what it did is framebusting but I'm not sure, yeah.
* ckocagil knows who to call when there are frames in the neighborhood
* paulproteus literally lols.
<dwrensha> I think target="_parent" is what causes it
<ckocagil> can it not be controlled by iframe sandboxing?
<dwrensha> seems like it ought to be!
<mnutt_> is there any way to get ionotify events into the filesystem namespace within a grain? with vagrant-spk, ideally I’d be able to update a file on the host system and the grain would pick it up and recompile the app. or is this in direct conflict with the way namespaces work?
<paulproteus> It's not in direct conflict, but it is does use Linux syscalls that are (I believe) blocked within the Sandstorm sandbox.
<paulproteus> But we could create a shim to make it work again.
<mnutt_> it’s somewhat of an edge case since it’s only useful for development
<paulproteus> 'spk dev' should already be allowing files to be edited hot, so I'm not quite sure what you're trying to do!
<paulproteus> So maybe I should ask: "What are you really trying to do?"
<mnutt_> ember, in dev mode, watches the filesystem and recompiles the app when files change. but I’m not seeing the changes come through
<paulproteus> Oh, interesting.
<mnutt_> it’s complicated slightly by going vagrant -> debian -> grain, but even when I touch files in the debian host I don’t see the updates.
<paulproteus> Yeah, we're probably blocking the inotify syscalls that ember attempts.
<paulproteus> Let me dou Let me double check that by reading the syscall blacklist.
<paulproteus> I don't _see_ us blocking inotify directly or via fcntl.
<mnutt_> I’m going to see if I can add some debugging to try to see what’s actually coming through
<paulproteus> Great. But also if you can wait until Monday I imagine kentonv already knows exactly what's going on.
<paulproteus> I'd love to see your experimentation, too, so feel mega free to do that.
<mnutt_> sounds good. at this point I’ve spent more time messing around with my dev environment trying to get it just right than actually working on the app :)
<paulproteus> Also BTW it seems that Ghost uses some Ember plugin for auth, so if you end up doing something Ember-y Sandstorm-y auth-y thing, then possibly we can make the Sandstorm package of Ghost use that, which would probably save us a huge amount of time refreshing the package periodically.
<paulproteus> Which also means I'm very incentivized to help you with Ember-y things.
<paulproteus> mnutt_: It could also be the case that VirtualBox file sharing is absorbing the inotify events. https://serverfault.com/questions/453826/vagrant-shared-folder-and-file-change-events suggests it's doin thag that.
<paulproteus> It also offers this arguably-useful answer: "I know this is an older question, but here is a more up-to-date answer:
<paulproteus> "
<paulproteus> http://serverfault.com/a/643857 is probably a useful link.
<mnutt_> oh that’s pretty cool. but the events seem not to go through even when I fire them from within the VM.
<paulproteus> Yeah, so I think then VirtualBox file sharing ("vboxsf") is absorbing the events.
<paulproteus> You could switch to NFS-based file sharing in your .sandstorm/Vagrantfile if you're up for it. I think that should improve things.
<paulproteus> Let me double-check if that's true.
<mnutt_> ok, I’ll give that a try. I may also just see if I can switch ember’s watcher to polling.
<paulproteus> So yeah, "for now", switching ember's watcher to polling seems reasonable. I wonder how Meteor seems to handle this fine.
<paulproteus> Presumably polling. If so, I'm impressed it isn't pegging the CPU. Go it.
<mnutt_> polling works! (other than the constant 60% cpu)
<mnutt_> but given it’s for development only, it’s really not a big deal
<paulproteus> Presumably if you can decrease the polling interval, CPU usage will decrease proportionally.
<paulproteus> But glad your immediate issue is addressed.
neynah has joined #sandstorm
* paulproteus waves
<paulproteus> ckocagil: My personal profile page in NodeBB shows an email address that is @example.com, which is therefore unlikely to reach anyone.
<paulproteus> Might as well remove that from the UI, I guess.
<ckocagil> yep. I'll make a kanban board to keep track of todos
<paulproteus> Yay
<paulproteus> larjona: I wrote http://thread.gmane.org/gmane.linux.debian.conference.team/13426/focus=13435 , hope that helps at all.
<paulproteus> Also I realizedl ike six ways I could have written it better since posting it. Oh well.
<larjona> paulproteus i whitelisted storm.debian.net but I still cannot do anything useful with for example http://deb.li/testwekan
<larjona> no, no, sorry, that's not related to noscript
<paulproteus> Interesting
<larjona> for wekan I think I should have a look at filing a bug, because when I share the link (edit-capable), people cannot edit unless they sign in and I grant them "invited user" permissions
<larjona> (but scrumblr for example works well, even with non-logged-in users)
<larjona> paulproteus thanks for your documented reply, in any case
<paulproteus> Sure, happy to reply.
<paulproteus> I think nothing should hinge on "Invited User" permissions, so if you can reproduce that, that could be a Sandstorm bug.
<paulproteus> So long as you shared the grain with the people, (maybe and they sign in, depending on the app) it should work for them.
<paulproteus> Although I admit I'm +1 on there being more "Invited Users".
neynah has quit [Quit: http://www.kiwiirc.com/ - A hand crafted IRC client]
<larjona> ok paulproteus, I'll reproduce it, and file a bug (I'll have a look at wekan-sandstorm issues before, maybe it's already reported)
<paulproteus> Awesome. Great, thank you.!
neynah has joined #sandstorm
neynah has quit [Client Quit]
<ckocagil> Q: has anyone considered the ability to give pretty URLs to grains?
<larjona> ckocagil I'm actually user an url shortener when sharing links (in my case, deb.li/prettyURL when I share with the debian community)
<paulproteus> Yeah -- we were discussing this a little bit yesterday around this time; the archives have some of that.
<larjona> user -> using
<paulproteus> The simplest thing I'd like to see is sandstorm.example.com/{{shortName}} as a redirect to a specific sharing link.
<larjona> btw my preferred url shortener ( pili.la ) died, and I'm considering to selfhost one.
<paulproteus> It'd be "easy" to create a Sandstorm-hosted URL shortener that uses the web publishing features, if you're OK with each shortened URL being a HTML page with a META REFRESH.
<paulproteus> Presumably we could improve the web publishing features to permit redirects, but you could hack it in the above way.
<paulproteus> Then you could bind a domain to the web publishing content, like how WordPress in Snadstorandstorm works, and you could make https:/.
<paulproteus> ...
<paulproteus> Then you could bind a domain to the web publishing content, like how WordPress in Sandstorm works.
<ckocagil> paulproteus: serving proper short urls would be better though
Takumi has joined #sandstorm
<Takumi> Hi. Are sandcats recovery emails working right now?
<larjona> FTR the wekan not-logged-in-users-cannot-edit bug is already reported, I just added some more info (https://github.com/wekan/wekan/issues/349 )
<kentonv> mnutt_, paulproteus: the problem is that the FUSE layer doesn't support inotify.
<mnutt_> kentonv: ah, that makes sense. polling works well enough for my use case.
<kentonv> we could in theory support it, it's just work. And any work relating to inotify tends to be very complicated frustrating work. :P
<kentonv> (FUSE has the hooks, but the FUSE layer in spk dev was implemented assuming the underlying data was immutable)
<mnutt_> yeah, as I started looking into it I realized it’s probably odd to start up a grain and then start modifying its contents from the outside. I’m now thinking it may make more sense to run the watcher / asset compiler on the host system and integrate a reload of spk dev when files change.
<mnutt_> (the only challenge there is that npm is really unhappy with native modules compiled for different archs. I really need to switch my dev environment to linux one of these days…)
<Takumi> So about sandcats recovery emails - are they sending or is it just me?
<dwrensha> Takumi: I suppose that paulproteus would know
<kentonv> Takumi: what's your email address? (Or just a prefix if you don't want to reveal it here.)
larjona has quit [Quit: Konversation terminated!]
larjona has joined #sandstorm
rysiek|pl has quit [Remote host closed the connection]
rysiek|pl has joined #sandstorm
larjona has quit [Quit: Konversation terminated!]
larjona has joined #sandstorm
rustyrazorblade has joined #sandstorm
<Takumi> My url appears to be working now, I just hadn't retried it with the right configuration for a while.
<Takumi> Thanks for the help.
Takumi has quit [Quit: Leaving]
rustyrazorblade has quit [Quit: rustyrazorblade]
dwrensha has quit [Ping timeout: 260 seconds]
larjona has quit [Ping timeout: 240 seconds]
bb010g has joined #sandstorm
rustyrazorblade has joined #sandstorm
micahd is now known as artygeek
artygeek is now known as micahd
rustyrazorblade has quit [Quit: rustyrazorblade]
rustyrazorblade has joined #sandstorm
rustyrazorblade has quit [Quit: rustyrazorblade]