gopar has quit [Remote host closed the connection]
jadewang has quit [Remote host closed the connection]
prosody is now known as prosodyC
larjona has quit [Quit: Konversation terminated!]
ocdtrekkie has quit [Remote host closed the connection]
aldeka has quit [Remote host closed the connection]
aldeka has joined #sandstorm
ocdtrekkie has joined #sandstorm
isd has quit [Ping timeout: 246 seconds]
isd has joined #sandstorm
jadewang has joined #sandstorm
gopar has joined #sandstorm
simonv3 has joined #sandstorm
jadewang has quit [Remote host closed the connection]
jacksingleton has joined #sandstorm
jacksingleton has quit [Ping timeout: 244 seconds]
mnutt__ has joined #sandstorm
jadewang has joined #sandstorm
home has joined #sandstorm
mnutt__ has quit [Quit: mnutt__]
jadewang has quit [Remote host closed the connection]
jacksingleton has joined #sandstorm
losvedir has joined #sandstorm
isd has quit [Quit: Leaving.]
gopar has quit [Remote host closed the connection]
losvedir has quit [Quit: losvedir]
bb010g has joined #sandstorm
losvedir has joined #sandstorm
gopar has joined #sandstorm
simonv3 has quit [Quit: Connection closed for inactivity]
jadewang has joined #sandstorm
jadewang_ has joined #sandstorm
jadewang has quit [Read error: Connection reset by peer]
losvedir has quit [Quit: losvedir]
home has quit [Ping timeout: 252 seconds]
losvedir has joined #sandstorm
losvedir has quit [Ping timeout: 264 seconds]
jacksingleton has quit [Ping timeout: 255 seconds]
larjona has joined #sandstorm
jadewang_ has quit [Remote host closed the connection]
gopar has quit [Quit: Leaving]
jadewang has joined #sandstorm
jadewang has quit [Ping timeout: 252 seconds]
KaZeR has quit [Ping timeout: 256 seconds]
mort___ has joined #sandstorm
mort___ has left #sandstorm [#sandstorm]
KaZeR has joined #sandstorm
mrflos has joined #sandstorm
<mrflos> hi everybody, i m trying to get sandstorm work on an nginx server, with forced https (autosigned), previously configured with a sandcats dyndns, i can connect and see the interface and docs, but i had to authorize acces for static.mysandstorm.domain to see the icons again, and now when i click on any grain, i have error 500... Can anyone help me
<mrflos> to find out what is wrong?
jadewang has joined #sandstorm
jadewang has quit [Ping timeout: 252 seconds]
<dwrensha> mrflos: do you see any errors in /opt/sandstorm/var/log/sandstorm.log on the server, or in your browser console?
<mrflos> hi dwrensha yes i have something long :
<mrflos> openSession error Error: Internal server error [500]
<mrflos> Trace de la pile :
<mrflos> sorry for the inconvenience..
<mrflos> and the log file
<mrflos> Exception while invoking method 'openSession' Error: connect(): Permission denied; toString() = unix:/var/sandstorm/socket/backend
<mrflos> at Object.Future.wait (/programs/server/node_modules/fibers/future.js:398:15)
<mrflos> at waitPromise (app/server/proxy.js:79:35)
<mrflos> at startGrainInternal (app/server/proxy.js:526:10)
<mrflos> at continueGrain (app/server/proxy.js:489:10)
<mrflos> at openSessionInternal (app/server/proxy.js:320:17)
<mrflos> at [object Object].Meteor.methods.openSession (app/server/proxy.js:154:12)
<mrflos> at maybeAuditArgumentChecks (packages/ddp/livedata_server.js:1617:1)
<mrflos> at packages/ddp/livedata_server.js:648:1
<mrflos> at [object Object]._.extend.withValue (packages/meteor/dynamics_nodejs.js:56:1)
<mrflos> at packages/ddp/livedata_server.js:647:1
<mrflos> - - - - -
<dwrensha> what do you see when you do `ls -l /opt/sandstorm/var/sandstorm/`?
<mrflos> the apps and grains were created while still on sandcats dyndns, but i don t know if it's relevant..
<mrflos> :q
<dwrensha> did you run re-install on top of an existing install?
<mrflos> drwxrwx--- 32 root sandstorm 4096 sept. 17 18:43 apps
<mrflos> drwxrwx--- 2 sandstorm sandstorm 4096 juil. 3 16:17 backups
<mrflos> drwxrwx--- 2 root sandstorm 4096 juil. 3 10:33 downloads
<mrflos> drwxrwx--- 32 root sandstorm 4096 sept. 17 19:07 grains
<mrflos> drwxrwx--- 2 root root 4096 sept. 18 13:41 socket
<mrflos> drwxrwx--T 2 root sandstorm 4096 sept. 18 13:41 tmp
<mrflos> yes i think so
<dwrensha> it sounds like you're hitting https://github.com/sandstorm-io/sandstorm/issues/660
<dwrensha> this might fix it for you: `sudo chown root:sandstorm /opt/sandstorm/var/sandstorm/backend`
<mrflos> yes just read it, it sounds similar, i m trying your permission changes, thx a lot
<mrflos> nope..
<mrflos> changed sudo chown root:sandstorm /opt/sandstorm/var/sandstorm/SOCKET/backend
<mrflos> and then
<dwrensha> oops, I meant `sudo chown root:sandstorm /opt/sandstorm/var/sandstorm/socket`
<mrflos> sudo chown root:sandstorm /opt/sandstorm/var/sandstorm/socket -R
<dwrensha> it's important to change the directory
<dwrensha> then make sure you do `sudo sandstorm restart`
<mrflos> i restarted, not better, shall i massivelly change everything to root:sandstorm in /opt/sandstorm and restart?
<dwrensha> are you seeing the same error as before in sandstorm.log?
<mrflos> no, in fact there is no more error 500 but
<mrflos> ** Shutting down due to signal
<mrflos> ** Restarting
<mrflos> ** Starting back-end...
<mrflos> ** Starting MongoDB...
<mrflos> note: noprealloc may hurt performance in many applications
<mrflos> about to fork child process, waiting until server is ready for connections.
<mrflos> forked process: 530
<mrflos> child process started successfully, parent exiting
<mrflos> ** Back-end and Mongo started; now starting front-end...
<mrflos> Migrations already applied: 9/9
<mrflos> No such grain: 3fmRcSmrzSXLE9QJ36rJgM
<mrflos> capnp/rpc.c++:122: info: returning failure over rpc; exception = kj/async-io.c++:120: failed: expected result >= minBytes; Premature EOF
<mrflos> stack: 0x50fe52 0x413b30 0x4136b0 0x412ac0 0x4118a0 0x4bbd40
<mrflos> Exception while invoking method 'openSession' Error: remote exception: expected result >= minBytes; Premature EOF
<dwrensha> does that happen only when you try to open an existing grain, or also when you try to create a new grain?
<mrflos> that was a new one, just tried an old one, there is still 500 error, also when trying to install new app from store
<mrflos> connect(): Permission denied; toString() = unix:/var/sandstorm/socket/backend
<dwrensha> still!
<mrflos> yes..
<mrflos> but
<mrflos> ls -l /opt/sandstorm/var/sandstorm
<mrflos> total 24
<mrflos> drwxrwx--- 32 root sandstorm 4096 sept. 17 18:43 apps
<mrflos> drwxrwx--- 2 sandstorm sandstorm 4096 juil. 3 16:17 backups
<mrflos> drwxrwx--- 2 root sandstorm 4096 juil. 3 10:33 downloads
<mrflos> drwxrwx--- 33 root sandstorm 4096 sept. 18 14:15 grains
<mrflos> drwxrwx--- 2 root sandstorm 4096 sept. 18 14:13 socket
<mrflos> drwxrwx--T 2 root sandstorm 4096 sept. 18 14:13 tmp
<mrflos> root@vps156534:/opt/sandstorm# ls -l /opt/sandstorm/var/sandstorm/socket/
<mrflos> total 0
<mrflos> srwxrwx--- 1 root sandstorm 0 sept. 18 14:13 backend
<mrflos> srwxrwx--- 1 root sandstorm 0 sept. 18 14:13 devmode
<mrflos> seems normal
<mrflos> i have no vital data, maybe i should do a fresh install, i don t want you to loose precious time on shitty bugs, unless its usefull
<dwrensha> I suppose I would also try `sudo sandstrom stop` and then `sudo sandstorm start`, instead of just `sandstorm restart`
<mrflos> ok
<dwrensha> a fresh install should probably make things better
<mrflos> not better, is it ok that the sockets have 0 bite?
<mrflos> not better, is it ok that the sockets have 0 bit?
<dwrensha> yeah, the sockets are special files
<mrflos> yes lets try this option! i will erase all /opt/sandstorm and do another install
<dwrensha> and I think 0 bytes is the usual size for them
<dwrensha> `sudo sandstorm stop`
<dwrensha> `sudo rm -rf /opt/sandstorm/`
<mrflos> i'm logged as root, can i do it without sudo ?\
<dwrensha> ... and if you want to change you configuration in the future, you can edit /opt/sandstorm/sandstorm.conf and then restart sandstorm
<mrflos> yeah.. my mistake
<mrflos> i did it
<mrflos> but was not carefull about https
<dwrensha> it's not your mistake, it's our bug.
<mrflos> so i first used the nginx.conf in github
<mrflos> but with http:// in the config file
<mrflos> after the oauth was not working
<mrflos> i decided to do install again
<mrflos> i should have checked in the config file before
<mrflos> i m not a sys admin, my will is to port some php programs to sandstorm, it s such a gamechanger
<mrflos> one click secure install of all kinds of programs! this is the end of google drive supremacy
<mrflos> ;
<mrflos> ;)
<dwrensha> it would be awesome if you could port some apps!
<dwrensha> do you have any particular apps in mind?
<dwrensha> ooh. I've not heard of YesWiki
<mrflos> yeah it s small but it s a wiki with kind of wiki-style databases
<mrflos> so it can be use to replace google forms
<dwrensha> I should warn you that zarvox is working on packaging Framadate for Sandstorm
<dwrensha> so you may want to sync up with him if you want to avoid duplicate work.
<dwrensha> (He's on California time. I'd guess he'll be up approximately three hours from now.)
<mrflos> ok, i will try to get in touch! i read from him in github!
<mrflos> thanks a lot, and i will contact the community if something is worth testing!
<mrflos> do you have any idea which php project is better integrated for now with sandstorm?
<dwrensha> I'm sorry, I don't understand the question. Are you asking about already-packaged apps? What do you mean by "php project"?
<mrflos> yes which packaged app as php is worth looking at to integrate with sandstorm
<mrflos> (learn from the bests)
mrflos has quit [Quit: http://www.kiwiirc.com/ - A hand crafted IRC client]
mrflos has joined #sandstorm
<dwrensha> mrflos: https://github.com/JamborJan/paperwork might be a good place to start, as it uses vagrant-spk
<mrflos> thx and lot bookmarked
<mrflos> see you, thamks again and have a nice end of week
<dwrensha> Man, I wish the app icons in our app market were actually <a> elements. Then I could control-click on a bunch of them to open them in browser tabs.
mrflos has quit [Quit: http://www.kiwiirc.com/ - A hand crafted IRC client]
jadewang has joined #sandstorm
<dwrensha> This is something that has bothered me about https://capnproto.org/ too. The menu there works by detecting click() events, apparently. If I control-click a menu item, it always navigates the current tab to that section, and depending on where I click sometimes opens a new tab to that section.
jadewang has quit [Ping timeout: 244 seconds]
larjona has quit [Quit: Konversation terminated!]
xet7 has joined #sandstorm
jadewang has joined #sandstorm
jadewang has quit [Ping timeout: 244 seconds]
larjona has joined #sandstorm
mnutt_ has joined #sandstorm
jadewang has joined #sandstorm
jadewang has quit [Ping timeout: 256 seconds]
losvedir has joined #sandstorm
larjona has quit [Ping timeout: 240 seconds]
KaZeR has quit [Ping timeout: 250 seconds]
KaZeR has joined #sandstorm
bb010g has quit [Quit: Connection closed for inactivity]
mnutt_ has quit [Quit: mnutt_]
jadewang has joined #sandstorm
jadewang has quit [Ping timeout: 246 seconds]
maurer is now known as maurer_
maurer_ is now known as maurer
gopar has joined #sandstorm
NOTevil has joined #sandstorm
<ckocagil> porting party?
<paulproteus> I'd be up for it ckocagil
<paulproteus> But why do you mention? : D
<paulproteus> Do you mention because you want to have one?
<paulproteus> I'd show up. I'm based in SF, fwiw.
<ckocagil> I wish I could attent an irl porting party :) SF is on the other side of the globe
<paulproteus> OK! I could do a remote-is-OK porting party too.
<paulproteus> What are you porting & where in the world are you?
<paulproteus> & hi, I'm Asheesh, nice to meet you.
<ckocagil> I'm Cem, currently porting Gogs, from Turkey
<ckocagil> nice to meet you!
<paulproteus> Oh that's SO COOL re: Gogs
<dwrensha> Gogs!
<paulproteus> dwrensha is the person who's packaged the other two git apps we have now, fwiw.
<dwrensha> Syste from GitLab actually recommended that I look into Gogs
<dwrensha> as it might be more lightweight than GitLab
<paulproteus> Oh right, hilarious.
<paulproteus> I guess GitLab is pretty non-light.
<dwrensha> *Sytse
<ckocagil> Gogs is quite lightweight. perfect for a self-hosted server.
<ckocagil> maybe Cgit would be even better, especially for large repos
jadewang has joined #sandstorm
<dwrensha> ckocagil: looks like Gogs has an issue tracker and maybe so other things
<dwrensha> so I guess it has some kind of database aside from just the Git repo itself?
<ckocagil> dwrensha: yes. I'm using embedded sqlite for now (/var/gogs/db)
<dwrensha> nice! Does Gogs officially support sqlite? Because if so, that's super awesome.
<ckocagil> yes! that made things much smoother for me
<ckocagil> does this look familiar to anyone? sandstorm/util.c++:257: failed: rmdir(path.cStr()): Permission denied; path = ../tmp
<paulproteus> ckocagil: It doesn't super-duper look familiar, but I wonder then if the permissions in your Sandstorm install are set up weird.
<ckocagil> paulproteus: 'vagrant-spk up' isn't running sandstorm, so I'm trying to run it manually. which user should I run it as?
<paulproteus> Oh my
<paulproteus> It totally _should_ run Sandstorm.
<paulproteus> Maybe we broke vagrant-spk in an update or something.
<paulproteus> Let me test it.
<ckocagil> it actually did run sandstorm the first few times.
<paulproteus> Huh.
<ckocagil> correction: it ran once. then I halted and powered the vm, and it wasn't running anymore
<paulproteus> That seems like it should work.
<paulproteus> ckocagil: I'm baffled. Would you be willing to let me SSH into your vagrant-spk box?
<paulproteus> It looks like if you do
<paulproteus> cd .sandstorm
<paulproteus> vagrant share # hmm, requires an account at https://atlas.hashicorp.com/ I guess, maybe that's annoying
<ckocagil> well, it's also inside another vm...
<paulproteus> Whoa. That works? Fascinating.
<paulproteus> What VM is around it?
<ckocagil> vmware player
<paulproteus> That's amazing.
<paulproteus> Is there anything more in the log?
<paulproteus> Does it do that at Sandstorm start time, or does it do it at 'vagrant-spk dev' time?
<paulproteus> BTW also you can run vagrant-spk without the VMWare player.
mnutt_ has joined #sandstorm
<paulproteus> vagrant-spk works on Mac OS (& Windows mostly) too.
<paulproteus> yay hi mnutt_
<mnutt_> hey paulproteus
losvedir has quit [Quit: losvedir]
<ckocagil> I prefer working in a linux environment
<ckocagil> I digged that line from the logs -- it might be due to me running './sandstorm start' as the wrong user though
losvedir has joined #sandstorm
losvedir has quit [Client Quit]
<paulproteus> Part of the point of vagrant-spk is that you don't have to run 'sandstorm start', so I'm surprised you're running that directly at all!
<mnutt_> do most people on sandstorm core develop on linux? I’ve been working in vagrant/virtualbox but it’s a bit of a pain because fs events don’t come through properly so I end up having to do a bunch of `make update`s
<paulproteus> Yeah, most of the Sandstorm core devs definitely dev on Linux.
<zarvox> ckocagil: for what it's worth, if you prefer Linux, you might still find it easier to run vagrant-spk on the host, and just "vagrant-spk ssh" to get a shell on the Linux VM.
<paulproteus> For better or for worse, but there you go.
<paulproteus> What I have been doing lately mnutt_ is cheating by doing 'make continuous' in Linux and doing 'sudo cp bin/sandstorm /opt/sandstorm/latest/sandstorm'
<mnutt_> I have a linux box, but the graphics card fan makes a terrible noise so I leave it off most of the time
<paulproteus> Aw.
<mnutt_> cool, makes sense. I sometimes just touch the files I’m working on from within vagrant, I could probably automate that
<paulproteus> Someone needs to set up a Vagrant setup that is optimized for hacking on the C++.
<paulproteus> No one has yet that I know, and someone requested it in ~February or March.
<zarvox> mnutt_: out of curiosity, is your host OSX, Windows, or Linux?
<mnutt_> zarvox: it’s osx
<mnutt_> I can use virtualbox’s regular sharing and get fs events but terrible performance, or use nfs and get decent performance but no fs events
<dwrensha> for what it's worth, I'm happy on OSX with virtualbox and no filesystem sharing
<mnutt_> I’m interested to see where xhyve goes, if they end up with better sharing
<paulproteus> ++
<paulproteus> mnutt_: If you want to get tricky you can invert the paradigm and do sshfs
<paulproteus> sshfs by mounting /home/vagrant/sandstorm-dev (or what have you) on the Mac side.
<mnutt_> ah, yeah that’s something to try
<paulproteus> Wait ckocagil tell me more about why you're running 'sandstorm start', if you are still around and willing!
<ckocagil> paulproteus: I decided to destroy that vm and start from scratch. I will report if the same issues comes up.
<paulproteus> Okay!
mort___ has joined #sandstorm
xet7_ has joined #sandstorm
prettyvanilla has joined #sandstorm
<paulproteus> Howdy new people. Nice to have you here.
xet7_ has quit [Read error: Connection reset by peer]
xet7_ has joined #sandstorm
xet7 has quit [Ping timeout: 246 seconds]
jadewang has quit [Remote host closed the connection]
jadewang has joined #sandstorm
mort___ has quit [Quit: Leaving.]
jadewang has quit [Remote host closed the connection]
jadewang has joined #sandstorm
jadewang has quit [Ping timeout: 240 seconds]
<ckocagil> non-unique_usernames_make_me_sad
<ckocagil> (so_does_the_fact_that_my_spacebar_is_broken...)
<dwrensha> ckocagil: does Gogs use a username/handle as a primary key?
<ckocagil> dwrensha: yes,like_github
<ckocagil> or_wait_i'm_not_sure_about_github
<dwrensha> I think github requires that handles are globally unique
<dwrensha> GitLab is similar
<dwrensha> but in the database, the primary key is an ID that is never displayed
<ckocagil> yeah._sandstorm_display_names_are_non-unique_though,_right?
<dwrensha> "display names" and "preferred handles" are not globally unique
<ckocagil> it_would_be_nice_if_their_uniqueness_were_enforced,_but_they_could_be_changed_at_anytime
<ckocagil> like_twitter_handles
<dwrensha> the identity ID, i.e. X-Sandstorm-User-Id, is globally unique
<ckocagil> true,_but_using_that_id_is_ugly
<ckocagil> and_if_I_used_display_names,_user_would_be_very_confused
<dwrensha> right, I was saying--- for GitLab I put the ugly identity ID in the database as the primary key
<dwrensha> and the database also has other fields, like handle, that have uniqueness constraints
<ckocagil> it_was_amazing_how_I_didn't_even_have_to_touch_one_source_file_to_get_that_on_Gogs
<ckocagil> it_already_had_reverse_proxy_login_and_registration_support!
<ckocagil> dwrensha: are_the_display_names_ever_used_on_Gitlab?
<dwrensha> so when I'm registering a new user in a GitLab instance, I first check the uniqueness constraints, and append some numbers to the handle if necessary
<dwrensha> and I just use the display name as-is (if I'm remembering correctly) because it doesn't have uniqueness constraints
<ckocagil> thatd_work,_but_then_you_would_have_to_change_your_display_name_separately_for_each_sandstorm_app
<dwrensha> how so?
<dwrensha> you update your display name in your Sandstorm profile
<dwrensha> and Gitlab can notice that next time you open it
<dwrensha> (I don't think it actually does that right now, but it wouldn't be hard)
<ckocagil> well,_you_can_do_that_but_that's_some_extra_work
<dwrensha> a bit of extra work might be unavoidable here
<dwrensha> what we're buying is the ability to transfer a grain between different Sandstorm servers, and have the identity stuff continue to work
<dwrensha> and there's no way we could globally guarantee that a display name or handle is not in use on another Sandstorm server
<dwrensha> or, rather, the way that you get such a guarantee to have an
<paulproteus> ckocagil:_now_I_want_to_type_like_this_in_moral_support_with_you
<dwrensha> "identity provider"
<ckocagil> paulproteus: choo_choo_ima_train
<paulproteus> : D
<paulproteus> :_D
<paulproteus> typo
<ckocagil> dwrensha: sandstorm_logins_are_global?
<ckocagil> :O
<dwrensha> no, "identities" are global
<dwrensha> they are you authenticate with a Sandstorm Server
<dwrensha> Github, Google, email token
<dwrensha> we're going to add more
<ckocagil> I_see
<dwrensha> each identity provider might have its notion of handle and display name
<dwrensha> but we definitely want to support many different identity providers
<dwrensha> so, as far as I can see, the best
jadewang has joined #sandstorm
<ckocagil> dwrensha: maybe_this_display_name_uniqueness_thing_could_be_made_into_a_library
<dwrensha> ... globally unique thing we can get out of them is a hash
<ckocagil> dwrensha: I'm_talking_about_your_solution_that_enforces_uniqueness
<dwrensha> the specific thing I did for GitLab?
<ckocagil> dwrensha: yes,_and_it_should_also_support_changing_display_names
mort___ has joined #sandstorm
<dwrensha> ckocagil: I'm not exactly sure what you're suggesting
<dwrensha> it's true that display names and handles should be able to be changed
mort___ has quit [Quit: Leaving.]
gopar has quit [Read error: Connection reset by peer]
<ckocagil> sorry, I was repairing my keyboard
<ckocagil> dwrensha: I mean a library that maps (ID, display name) -> username
<ckocagil> so we don't have to implement the same functionality for each app we port
gopar has joined #sandstorm
<ckocagil> another idea: sandstorm could make you pick a unique name for that server after login
mnutt_ has quit [Quit: mnutt_]
mnutt_ has joined #sandstorm
mcpherrin has quit [Read error: Connection reset by peer]
mcpherrin has joined #sandstorm
ocdtrekkie has quit [Remote host closed the connection]
aldeka has quit [Remote host closed the connection]
ocdtrekkie has joined #sandstorm
aldeka has joined #sandstorm
NOTevil has quit [Quit: Leaving]
_jax_ has joined #sandstorm
<_jax_> anyone around that can help me with something
<_jax_> sandstorm is installed, but apps are forever stuck at loading/pinwheel of death
<_jax_> i'm seeing nothing in the mongo or sandstorm.log
<dwrensha> _jax_: what about your browser console?
<_jax_> admin works fine
<dwrensha> sounds like it could be a wildcard DNS issue
<_jax_> that would make sense
<_jax_> i'm having to use the ip address of the box in the absence of dns
<dwrensha> what are you values for BASE_URL and WILDCARD_HOST ?
<paulproteus> _jax_: In that case you can use xip.io!
<dwrensha> ... in /opt/sandstorm/sandstorm.conf
<paulproteus> BASE_URL=http://10.0.0.1.xip.io:6080/
<paulproteus> WILDCARD_HOST=*.10.0.0.1.xip.io:6080
<_jax_> BIND_IP=0.0.0.0
<_jax_> BASE_URL=http://192.168.12.129:9001
<_jax_> WILDCARD_HOST=*.local.sandstorm.io:9001
<_jax_> thank you paulproteus
<paulproteus> er remove trailing slash from my proposed BASE_URL!
<paulproteus> BASE_URL=http://10.0.0.1.xip.io:6080
<_jax_> ah yeah i can't use that
<_jax_> has to be 100% internal
<_jax_> i'm cutting off network access to this box as soon as the apps are properly config'ed
<paulproteus> Nod.
<paulproteus> Does your internal network have any internal DNS at all?
<_jax_> they didn't release the xip.io code publicly?
<_jax_> unfortunately not that i can use for this purpose... long story
<paulproteus> You can apt-get install powerdns, and use the xip stuff on top of it.
<paulproteus> Hmm... if you don't get the ability to have internal clients do DNS at all, then I don't know if this can work. Which is disappointing as I think about it.
<paulproteus> You could theoretically have them DNS to an internal DNS server, but you said that wouldn't work for reasons.
<paulproteus> (Which I totally grant you are real reasons.)
<_jax_> i just need to be able to point directly to the server ip.
<_jax_> in the absence of that, i'll figure something else out.
<paulproteus> The problem is that we use subdomains of the server host kind of like half of a user's session ID cookie, and there's no current way to disable that.
<paulproteus> So I don't think it can work for you today.
<paulproteus> Sorry about that.
<paulproteus> Two questions then.
<paulproteus> 1. If Sandstorm itself could operate as a DNS server for this internal mapping stuff, would that work for you in the future?
mnutt_ has quit [Quit: mnutt_]
<_jax_> ohhh
<_jax_> i see...
<paulproteus> 2. Can I get you to file a bug on GitHub about your needs, so that we don't forget about it?
<_jax_> is this really a bug?
<_jax_> who doesn't have dns? lol
<paulproteus> : D you don't
<_jax_> sounds like a unique use case that just needs documentation
<_jax_> "hey btw you need dns"
<geofft> if I'm reading right, you're cutting off external network access to the Sandstorm box, but not to clients? or to both?
<paulproteus> But also in general the use of Sandstorm on entirely private networks is something we should make feasible.
<_jax_> geofft: to the box, yes, but not to clients.
<_jax_> it's a large unverified code base on the box and can't afford anything getting out
<paulproteus> We could have a docs page called "How to use Sandstorm with limited/absent Internet access".
<_jax_> agreed paulproteus
<paulproteus> And the docs live in the main repo, so see, that's a good reason to file a bug (-:
<paulproteus> I can file it if you'd prefer; that's fine too.
<_jax_> another point, i had trouble finding the application packages to install without hitting the web
<geofft> then you can install xip-pdns on the Sandstorm box, point it at itself as authoritative for xip.io, and let clients look up xip.io against the public xip.io servers :)
mnutt_ has joined #sandstorm
<_jax_> geofft: can't touch public internet
<_jax_> paulproteus: thank you
<geofft> the clients can't touch public internet either?
<geofft> then I have no good suggestions, sorry :)
<_jax_> clients can, but that's an info leak of its own
<geofft> true
<paulproteus> Sadly seemingly mDNS doesn't seem to do wildcards for hostnames.
<_jax_> sounds stupid but it's partly due to where i work
<paulproteus> ...oh hi you're the person I was Tweeting at!
<paulproteus> Cool. I'm Asheesh, nice to e-meet you.
<paulproteus> At least maybe. Maybe that's an identifying information leak and I should hush.
<paulproteus> Anyway. I can write a docs page. I think the short answer is, "We don't support that yet" but I'd really like to. Is your interest in Sandstorm sort of now-or-never, or would you be interested to hear about when we do?
<paulproteus> The UI does assume both the server and the clients can reach the Internet.
<paulproteus> As you mentioned, for app installs.
<paulproteus> But that's not something that needs to be the case forever.
<_jax_> oh hey asheesh :D
<_jax_> I can submit something on Github if it'll help for tracking purposes
<paulproteus> Yeah, that would be great.
<paulproteus> Also I'm curious how you heard about Sandstorm.
<_jax_> i was looking for alternatives to etherpad
<_jax_> google rabbit hole...
<paulproteus> Also I wish we could solve your problem today. Unfortunately my current priority is "Focus on giving end users easily configured HTTPS" which turns out to be slightly more complicated than I thought.
<_jax_> agreed, https is a major issue
<_jax_> i'm having to look at wrapping with nginx
<_jax_> would be nice to have that as an option during install
xet7_ has quit [Read error: Connection reset by peer]
xet7_ has joined #sandstorm
<paulproteus> HTTPS without Internet access is also interesting because how would you look up the hostname that the HTTPS cert is issued to?
<paulproteus> I guess you could use an internal CA pretty easily though.
<_jax_> yep ^ that.
<paulproteus> Thumbs up.
<_jax_> submitted on github, thanks again paulproteus
<_jax_> i think this model is very interesting and i'll be following the project.
<paulproteus> Awesome.
<_jax_> re: SSL cert, i was more interested in protecting interception, validating via CA won't be necessary at this early stage
<paulproteus> A few requests which you're very welcome to ignore at your option.
<_jax_> but PKI sucks anyway, and if it's not configured 100% properly it's sortof not even worth setting up
<paulproteus> Consider subscribing to the announcements email list (at most, once a month) at https://sandstorm.io/ -- the form up top.
<paulproteus> Consider subscribing to https://groups.google.com/forum/#!forum/sandstorm-dev if you want more insight into what we're up to and want to shape that by commenting.
<paulproteus> Stay in this channel and make requests for apps and features you need.
<paulproteus> </requests>
<paulproteus> _jax_: All very fair points re: SSL.
<_jax_> done and done
<_jax_> i wonder if something like this could work paulproteus http://blog.pivotal.io/labs/labs/sslip-io-a-valid-ssl-certificate-for-every-ip-address
<_jax_> ugh. pki sucks.
<paulproteus> My plan for now is to make a deal with a CA and just buy a bunch of certs and then give them out for free. Yay being a startup.
<paulproteus> Since we need wildcards I'm hoping Let's Encrypt can be an avenue for the future.
<paulproteus> https://letsencrypt.org/ fwiw.
<paulproteus> (aka they don't do wildcards yet so only in the future will it be a good fit)
<paulproteus> _jax_: BTW -- your clients are separate from your server, right?
<_jax_> yep.
<paulproteus> One silly proposal for you is:
<paulproteus> If you can port-forward localhost:6080 to the server, on all the clients, then you can take advantage of the fact that localhost has an automatic wildcard that browsers & operating systems implement.
<paulproteus> But you'd have to do this silly port forward thing from the clients which is presumably a substantial configuration hurdle.
<_jax_> hmm
<_jax_> don't you think local powerdns with xip script is a better idea?
<paulproteus> Oh, I guess similarly you could do something like http://daniel.hahler.de/easy-dns-wildcard-setup-for-local-domains-using-dnsmasq on the clients which is far more reasonable.
<_jax_> i saw that too.
<paulproteus> I think powerdns with xip script is *FAR* more reasonable, yeah, than my port forwarding jokes.
<paulproteus> Great.
<_jax_> i'm working on it but i have my daughter tugging at my hem so it may be tomorrow before it's worked out
<_jax_> i'll let you know what i end up with
<paulproteus> Cool.
<paulproteus> Excited to hear about it.
<_jax_> small security changes break all kinds of things.
<_jax_> powerdns doesn't give you the option to set the admin user for mysql upon initial config, so that breaks the install
<_jax_> irritating.
<_jax_> http://7a3781c08da0795bf5abe05b5b52da32.local.sandstorm.io:9001/_sandstorm-…id=53578f06eab4d24837788713366d7964394cf83457545678bf71842e9de1abe0&path=/ Failed to load resource: net::ERR_CONNECTION_REFUSED
<_jax_> why would this be happening
mnutt_ has quit [Quit: mnutt_]
<paulproteus> That hostname presumably resolves to 127.0.0.1. Is that what you're looking for?
<paulproteus> https://docs.sandstorm.io/en/latest/administering/wildcard/ talks about local.sandstorm.io if that helps at all.
<paulproteus> Unrelated but: the name Atmosphere for a Meteor package site is _so good_.
<paulproteus> It makes me feel so futuristic.
<dwrensha> "crater" is a good name for this: https://crater.io/ , but it's an even better name for this: https://github.com/brson/taskcluster-crater
<paulproteus> I'm hoping there's also a forkbomb named crater.
<paulproteus> HTTP request sent, awaiting response... 302 Moved Temporarily
<paulproteus> Location: [object Object] [following]
<paulproteus> Warning: wildcards not supported in HTTP.
<paulproteus> --2015-09-18 15:58:43-- http://localhost:6080/[object%20Object]
<paulproteus> That's... not right.
<paulproteus> [fixing]
<ckocagil> Gogs repo is here: https://github.com/cem/gogs-sandstorm
<ckocagil> in case anyone wants to try out
<ckocagil> you can't push or pull yet. I'll have to steal some of Derek's code for that.
<paulproteus> Nice!
<zarvox> I totally want to play with this! Gitweb is a bit on the minimal side and GitLab a bit on the heavyweight side...
<zarvox> oh, they use gravatars by default; we could add sandstorm avatar integration!
<geofft> dwrensha: does GJ io handle the case of being in the middle of an AsyncWrite::write() and the connection dying?
<geofft> I don't see write_internal doing anything to check for that
<dwrensha> sure, it's return in the Err() branch then
<dwrensha> *it'll
<geofft> oh hey try!. ok.
<dwrensha> geofft: what version are you using?
<dwrensha> I don't see a try!() in the latest write_internal()
<geofft> whatever I last git pulled. I guess I should pull.
<geofft> yeah I was right at 0.0.3. OK, this version also seems to match the prototype of the function on the web docs better :)
<dwrensha> note that now you don't lose the buffer and the stream in the error case
<dwrensha> ... made possible by adding an error type parameter to Promise. So now it's Promise<T, E>.
<geofft> yeah
<dwrensha> a Promise<T,E> is basically a deferred Result<T,E> , with some nice functions for chaining