--- Log opened Thu Aug 02 00:00:07 2012
00:18 < Raynos> SubStack: where do you hide the uptime
00:18 < Raynos> of these things
00:23 < Raynos> SubStack: https://github.com/substack/propagit/blob/master/index.js#L398 this is the thing I want to read from in `fleet ps`
00:23 < SubStack> uptime?
00:25 < Raynos> yes
00:25 < tblobaum_> uptime could be kind of nice
00:28 < tblobaum_> but with that, maybe just make some kind of http endpoints that will stream monitor or stream process meta data as json
00:29 < Raynos> or store the spawning time for the process
00:29 < Raynos> or get the ctime of the pid file for the process
00:33 < SubStack> aha that would make sense
00:34 < SubStack> I think the start time is already recorded in the metadata too
00:34 < Raynos> where is it?
00:34 < SubStack> propagit/index.js:392
00:34 < SubStack> oh I'm mistaken, doesn't store the start time
00:35 < SubStack> but that's where I'd put it
00:35 < SubStack> I think a separate service should handle historical process uptime
00:35 < SubStack> since that would involve saving things to disk
00:36 < SubStack> but that system could just hook into the same events that `fleet monitor` does
00:36 < tblobaum_> maybe the hub could emit some changes events
00:36 < tblobaum_> with some meta data
00:36 < tblobaum_> and build on that whatever ui or history
00:36 < SubStack> it already does!
00:37 < SubStack> it's just a matter of adding a field to the process metadata
00:37 < tblobaum_> nice it makes sense it would have to, but not sure if its easily tapped in to
00:38 < SubStack> look at how fleet monitor hooks into the event stream
00:38 < tblobaum_> is the hub coupled to the fleet commands very closely? or could i conne t to it with dnode and some things would work
00:38 < tblobaum_> ill take a look, just easier to ask :)
00:38 < SubStack> the fleet project itself is just a collection of scripts
00:39 < SubStack> all the service logic is in propagit
00:39 < maxogden> the internet is just a collection of scripts
00:39 < SubStack> you can directly connect over dnode just fine
00:39 < SubStack> maxogden: fact
00:57 < SubStack> Raynos: sounds like what you wanted to do with seaport: https://twitter.com/jedschmidt/status/230823555980484608
00:58 < Raynos> SubStack: he uses buzzwords
00:58 < Raynos> all I want to do is talk to arbitarily named web hooks over seaport
00:58 < Raynos> ugh
00:58 < Raynos> I used his silly web hook word
00:59 < Raynos> a browser is simply a process. seaport makes processes talk to processes. end.
00:59 < SubStack> here's what it would look like https://twitter.com/jedschmidt/status/230825146049511424
01:00 < SubStack> something like browser-side mountie I think
01:01 < Raynos> browser side mountie
01:01 < Raynos> is needed at some point
01:01 < Raynos> to aggregate many small pieces into a public facing whole
01:03 < maxogden> web hooks a) arent silly and b) arent jeds word :)
01:04 < tblobaum_> Raynos: you dont want to expose browsers to browsers though, right?
01:04 < Raynos> maxogden whats a web hook
01:04 < Raynos> tblobaum: in theory yes
01:04 < tblobaum_> i think he just means service
01:05 < maxogden> webhooks are like http callbacks
01:05 < Raynos> a browser is simply a named service on a seaport
01:05 < Raynos> and anything can query for a named service on seaport
01:07 < tblobaum_> Raynos: oh
01:07 < tblobaum_> Raynos: I thought you just wanted to access seaport services directly from the browser
01:07 < Raynos> seaport is bidirectional
01:07 < Raynos> firstly
01:07 < Raynos> i want to access seaport services directly from the browser
01:08 < Raynos> but there is no reason why a browser cant register a service on the seaport proxy
01:08 < tblobaum_> like, building semver mostly, and the service aggregation aspect
01:08 < tblobaum_> so you can get the seaport service for queueing up emails, then fro the seaport metadata get the location of that service, make an xhr request there, and so on
01:09 < tblobaum_> s/building/building on/
01:09 < Raynos> yeah
01:10 < Raynos> i was thinking you can ask seaport and get back a stream that talks directly to that thing
01:10 < Raynos> what I really want is airport in the browser
01:10 < Raynos> so I can just assume we all talk over upnode
01:10 < Raynos> and fire left & right
01:12 < Raynos> what does substack use for his logging purposes
01:12 < SubStack> console.log
01:12 < tblobaum_> yeah airport might be better
01:13 < Raynos> SubStack: but console.log does not show up in fleet monitor
01:13 < Raynos> and if I dont have fleet monitor up the console.log just goes nowhere
01:13 < SubStack> it should!
01:13 < SubStack>     em.on('stdout', function (buf, proc) {
01:13 < SubStack> monitor listens for stdout events
01:13 < Raynos> i know it should
01:13 < SubStack> and stderr events
01:14 < Raynos> im tired ._.
01:14 < SubStack> could be a bug
01:14 < SubStack> lots of bugs still
01:14 < Raynos> i think ill just go home
01:14 < Raynos> and sleep for 14 hours
01:14 < Raynos> ._.
01:14 < Raynos> im going to oversleep
01:14 < Raynos> and feel tired again
01:14 < SubStack> DO EEET
01:14 < LOUDBOT> YOU CAN READ THE SOURCE IT WON'T BITE
01:14 < Raynos> this is a vicious cycle
01:15 < Raynos> how do you nullify tired
01:15 < SubStack> naps
01:15 < Raynos> tired | /dev/null
01:15 < SubStack> and strategic coffee consumption
01:15 < maxogden> require('coffee-beans').pipe('brewer').pipe(process.stdin)
01:16 < SubStack> exercise helps too
01:16 < Raynos> maxogden: where are you?
01:16 < Raynos> oh wait
01:16 < Raynos> your near the thing
01:16 < tblobaum_> yes some mild exercise
01:16 < maxogden> wat
01:16 < tblobaum_> and waking up quickly, rather than slowly
01:16 < maxogden> tblobaum_: whoa really
01:16 < SubStack> Raynos: go bike over to the sfjs thing to listen to max's cat presentation!
01:16 < maxogden> isaacs: i wish i could tag other peoples node modules
01:17 < maxogden> isaacs: instead of just star them
01:17 < SubStack> taghub
01:17 < Raynos> tag:shit
01:17 < Raynos> tag:win
01:17 < maxogden> haha
01:17 < Raynos> tag:solve-my-problems
01:17 < isaacs> maxogden: that's a good idea.
01:17 < Raynos> tag:WARNING-COFFEESCRIPT
01:17 < maxogden> isaacs: i was just thinking it would be fun to tag all the streaming parsers with #streaming-parser
01:17 < isaacs> maxogden: the "star" data structure is set up to enable that
01:17 < Raynos> then I can write a twitter bot
01:18 < maxogden> isaacs: you use a database that supports sparse documents?!?
01:18 < isaacs> maxogden: it's actually "users":{"maxogden":true,...} on the package doc
01:18 < Raynos> that listens on twitter hash tags and replies with github repo links
01:18 < maxogden> isaacs: haha
01:18 < Raynos> twitter: I need a #streaming-parser!
01:18 < isaacs> maxogden: with the intention that you could make tht whatever
01:18 < SubStack> a string instead of true?
01:18 < isaacs> "users":{"maxogden":["maxos","tagsos"]}
01:18 < isaacs> SubStack: or an object!
01:18 < isaacs> anything truthy!
01:18 < SubStack> right-o
01:19 < SubStack> haha I'm going to start calling mac osx maxos
01:19 < maxogden> YES
01:19 < isaacs> SubStack: the proper name for it is "Fruity Cat"
01:20 < maxogden> Mac OS X 10.9 Max Ogden
01:20 < maxogden> isaacs: i broke my own personal stream record earlier
01:20 < maxogden> isaacs: most pipes in a single program
01:21 < maxogden> in the browser no less
01:21 < maxogden> https://github.com/maxogden/dataroller/blob/master/attachments/drophandler.js#L32
01:21 < isaacs> nice :)
01:21 < maxogden> and on the server https://github.com/maxogden/dataroller/blob/master/index.js#L29
01:21 < maxogden> the first link pipes to the second
01:21 < maxogden> 8 streams in a row wooooooo
01:22 < SubStack> unixifing the browser
01:22 < SubStack> with some hot piping node point javascript ideology
01:22 < maxogden> haha
01:23 < maxogden> isaacs: so you liked that stream constructor side effect pattern?
01:23 < isaacs> maxogden: yeah, i mean, i use it a lot
01:23 < maxogden> isaacs: coolio. seemed too convenient not to use
01:23 < isaacs> ok i'm heading over to some sf js thing
01:23 < maxogden> isaacs: hahah im talking at that
01:23 < isaacs> i heard there's a cat there talking about databases with slides featuring bearded men
01:23 < maxogden> isaacs: your cat is in it like 4 times
01:24 < maxogden> isaacs: bring aristotle for bonus points
01:24 < maxogden> isaacs: and also so the presentaiton can be delivered by him
01:24 < isaacs> maxogden: hahhaah, no, he's even more of a homebody than i am
01:24  * isaacs away!
01:32 < tblobaum_> maxogden: nice
05:23 < Raynos> isaacs: When checking dependencies into git. Whats the best way to handle compiled dependencies, cc mikeal, SubStack
05:24 < SubStack> Raynos: npm rebuild
05:24 < SubStack> fleet exec --drone=* -- npm rebuild
05:24 < SubStack> after you deploy
05:25 < Raynos> i see
05:26 < Raynos> npm rebuild without package name rebuilds all the ones in package.json ?
05:26 < Raynos> is there a post git checkout hook or something that can run npm rebuild for me?
05:26 < Raynos> does fleet have any post deploy hooks ?
05:28 < Raynos> are there any traps with accidentally committing compiled binaries to git in the node_modules folder?
05:31 < SubStack> just add entries to your .gitignore for those
05:32 < SubStack> .git\n*.so\n*.node\*.swp\n*.swo
05:51 < Raynos> sweet
05:51 < Raynos> SubStack: again some post deploy hook in fleet where I can run npm rebuild would be sweet
05:54 < SubStack> maybe
05:54 < SubStack> you could also just make a deploy.sh and put those commands in it
05:55 < SubStack> fleet deploy\nfleet exec --drone=* -- npm rebuild
06:03 < Raynos> i have a make deploy target already
06:03 < Raynos> that does that
06:03 < Raynos> it feel it would be nicer if this was a git based hook thing
06:09 < SubStack> git hooks are actually pretty lame
06:09 < SubStack> I wouldn't use them for anything important
06:09 < SubStack> http://www.youtube.com/watch?v=4VHaORv6GYo
06:09 < SubStack> wups
06:09 < SubStack> https://github.com/substack/node-git-emit
06:10 < SubStack> git hooks are just shell scripts anyways
06:10 < SubStack> so just tell it what to do explicitly from the start
06:11 < SubStack> or make a deployinator module that wraps fleet deploy
06:11 < SubStack> hmm actually
06:11 < SubStack> putting a "scripts" field into fleet.json might work
06:11 < SubStack> but it should not use git hooks because those suck
06:12 < SubStack> a "scripts" field and a global ~/fleet.json are not compatible features
06:12 < Raynos> note the global fleet.json
06:13 < Raynos> is simply a default thing you inherit from
06:13 < Raynos> node-confuse style
06:13 < Raynos> mainly for the purpose of setting up the self remote that points at localhost:7000
06:13 < Raynos> so that I can locally test fleet deployment on all my modules
06:14 < SubStack> could be if it worked in the merge way
06:14 < Raynos> yes
06:14 < Raynos> thats what I intended
06:14 < Raynos> in the PR
06:14 < SubStack> ah
06:16  * SubStack is fixing browserify watch issues
06:17 < SubStack> blarg, no modules on npm for showing exactly where a syntax error occured
06:18 < SubStack> going to have browserify do a quick syntax check with Function() and generate a nice error message if that throws
06:21 < SubStack> this will also be helpful for the testling overhaul
06:21 < SubStack> so I can check the syntax with node first
06:39 < guybrush> SubStack: `var foo = './foo.js'; require(foo);` works with browserify right?
06:40 < SubStack> browserify will show a warning and skip that
06:40 < guybrush> oh i just saw it
06:49 < SubStack> pow ok have a sweet little lib using esprima and Function() now for checking and formatting syntax errors
06:51 < Raynos> substack: gh link
06:52 < SubStack> not up yet, writing docs, examples, tests, etc
06:53 < SubStack> post production
06:54 < Raynos> gist it?
06:54 < Raynos> also why write docs afterwards
06:54 < Raynos> I write docs/examples then (tests if TDD) the code
07:02 < SubStack> because I like exploring first
07:02 < SubStack> usually I don't know what I'm trying to build until I've built it already
07:02 < SubStack> designing things ahead of time seems to waterfall for me
07:17 < Raynos> i kind of agree
07:18 < Raynos> however I like writing down how the module is used
07:18 < Raynos> so I have some direction of what its supposed to do and what the API should be like
07:19 < Raynos> it also helps a lot to have a starting point for TDD
07:20 < SubStack> mostly I mess around on the repl and move the snippets over into an index.js
07:21 < SubStack> https://github.com/substack/node-syntax-error
09:11 -!- Topic for #stackvm: http://browserling.com | http://testling.com | We put a browser in your browser!
09:11 -!- Topic set by SubStack [~substack@ec2-50-17-202-46.compute-1.amazonaws.com] [Thu Oct 27 21:10:07 2011]
09:11 [Users #stackvm]
09:11 [ Altreus       ] [ dhoss        ] [ ircretary    ] [ Nexxy       ] [ Raynos    ] [ SubStack ] 
09:11 [ antix         ] [ dools        ] [ isaacs       ] [ niftylettuce] [ rook2paw1 ] [ sveisvei ] 
09:11 [ blakmatrix    ] [ duncanbeevers] [ ITpro        ] [ paul_irish  ] [ rowbit    ] [ tanepiper] 
09:11 [ CAPSLOCKBOT   ] [ egadcable    ] [ jesusabdullah] [ perlbot     ] [ ryanseddon] [ tblobaum_] 
09:11 [ captain_morgan] [ elliottcable ] [ LOUDBOT      ] [ pikpik      ] [ saijanai_ ] [ timoxley ] 
09:11 [ chapel        ] [ farnsworth   ] [ Madars       ] [ pkrumins    ] [ simcop2387] [ yorick   ] 
09:11 [ coffeecup     ] [ framlin      ] [ maxogden     ] [ purr        ] [ slurp     ] [ zu       ] 
09:11 [ cubert        ] [ guybrush     ] [ mikeal       ] [ py1hon      ] [ sorensen  ] [ zz_shykes] 
09:11 [ devaholic     ] [ ik           ] [ munro        ] [ rannmann    ] [ stlsaint  ] 
09:11 -!- Irssi: #stackvm: Total of 53 nicks [0 ops, 0 halfops, 0 voices, 53 normal]
09:11 -!- Channel #stackvm created Thu Mar 18 00:02:18 2010
09:11 -!- Irssi: Join to #stackvm was synced in 18 secs
09:27 < SubStack> https://gist.github.com/3235797
09:28 < SubStack> watch mode fixed, nice syntax errors
17:55 < st_luke> inefficiency
18:10 < yhuang> Hey guys, seems like it's not working again. it'll be helpful if you guys have a page on the site to check status of the servers
18:22 < maxogden> +1
18:27 < chapel> yhuang: http://browserling.com/nexus/encoders.html
18:28 < yhuang> chapel: Thanks!
18:28 < chapel> just remember substack linking to it
18:28 < chapel> not sure if its useful really
18:30 < yhuang> it's actually very useful, those are the list of browsers that are currently working
18:30 < yhuang> I was trying to run via chrome and was hanging , and chrome isn't on that list
18:30 < yhuang> so I assume whatever was down isn't list
18:30 < yhuang> listed*
18:44 < Raynos> SubStack: seaport.get("foo@~0.0.2"), that will always try to return the highest number, correct?
19:02 < jesusabdullah> SubStack: http://imgur.com/adoxq DC Tycoon irl
19:24 < SubStack> Raynos: I'm unsure
19:24 < Raynos> SubStack: is it a bug if it does not do so?
19:25 < Raynos> the same with get(foo@0.0.x)
19:25 < Raynos> I want to use that syntax so I can spawn up newer versions of my code
19:25 < Raynos> and have all the other things in the seaport cluster always get the latest code
19:26 < Raynos> that is the main selling point semver in seaport
19:39 < yhuang> Hmm, seems like down again
20:15 < Raynos> SubStack: do you have some thing that notifies me if one of my fleet processes is down or if seaport can't find a service
20:15 < Raynos> basically rowbit
20:23 < Raynos> SubStack: we are sending the fleet secret over the wire in HTTP and not in HTTPS
20:23 < Raynos> whats the easiest way to punch propagit to encrypt my secret
20:47 < mikeal> hey SubStack
20:48 < mikeal> did you wanna come to SummerCamp
20:55 < maxogden> what kind of question is that
22:53 < SubStack> Raynos: I agree that the random hex string names are a bit crap
22:53 < Raynos> well `fleet drone --name=beepboop`
22:53 < Raynos> would fix this
22:53 < SubStack> that already should work
22:54 < Raynos> it should?
22:54 < SubStack> yep
22:55 < SubStack> oh hmmm
22:55 < SubStack> it did work that way at one point
22:55 < Raynos> bring it back :)
22:55 < SubStack> on it
22:56 < Raynos> this reminds me It would also be useful if I can see env data in fleet ps
22:57 < Raynos> like thats a process I spawned with evn blarg
22:57 < SubStack> with an --env flag maybe
22:57 < Raynos> yeah
22:57 < Raynos> fleet ps --env
23:04 < Raynos> SubStack: https://github.com/substack/node-marx#readme that needs a picture
23:08 < SubStack> Raynos: ok the latest propagit now lets you set an opts.name and auto-increments to avoid conflicts
23:08 < Raynos> sweet :)
23:08 < SubStack> and fleet passes argv directly to propagit
23:08 < SubStack> so just fleet drone --name=rawr
23:11 < Raynos> o/
23:11 < Raynos> --name=cheesefactory
23:12 < Raynos> do you use fleet stop?
23:18 < SubStack> yep
23:18 < SubStack> it's pretty handy when a process is crashing in a loop or when you want to take down an old version of a service
23:18 < Raynos> what did you use instead before mikeal wrote it
23:19 < Raynos> do you have a badass secret deploy-new-version workflow?
23:19 < SubStack> nothing, it fleet was just a proof of concept
23:19 < SubStack> s/ it/
23:19 < mikeal> SubStack: did you get my message
23:19 < SubStack> mikeal: yeah sounds fun
23:19 < mikeal> did you get the message with the link?
23:19 < SubStack> I have a tent
23:20 < mikeal> i've got a bed for you, just fill out the damn form :)
23:20 < SubStack> done!
23:20 < SubStack> awesome thanks
23:21 < mikeal> np :)
23:22 < SubStack> also fleet isn't totally broken now like it was previously
23:22 < Raynos> i should sign up for summer camp
23:23 < SubStack> probs
23:23 < mikeal> today is probably the last day you can
23:23 < Raynos> oh i see
23:23 < Raynos> then I should do that now
23:25 < Raynos> ok
23:25 < Raynos> i paid o/
23:25 < mikeal> awesome :)
23:26 < Raynos> I never thanked you for the beer
23:26 < Raynos> so I'll buy you some at summer camp
23:33 < mikeal> i buy all the beer at my conference :)
23:38 < Raynos> mikeal: I will buy you beer at all your conferences then
23:51 < SubStack> Raynos: fleet@0.1.4 now takes the default drone names from os.hostname()
23:51 < mikeal> shit
23:51 < SubStack> and propagit does all the -$n stuff to ensure uniqueness
23:58 < SubStack> hmmm tests are failing...
--- Log closed Fri Aug 03 00:00:13 2012