IRC Logs for #cmvt Monday, 2013-05-20

DanielBairdyeah np i was just having a go while you were a.f.IRC00:00
prologicyou familiar with reStructuredText?00:00
prologicI'm planning on using Sphinx with reStructuredText for our documentation00:00
prologicwe have bare minimal docs right now00:00
prologicREADME.rst and docs/*00:00
DanielBairdhaven't done anything big in reST, i've given up on markdown and moved to textile00:00
prologicahh k00:00
prologicthey're all pretty similar ihmo00:01
DanielBairdanyway it's just thinking at the moment.. the first thing for our RESTful api is to work out what the resources actually are00:01
prologicbut yeah let's move this to docs/source/design.rst00:01
prologicrealtime editing is "cool"00:01
prologicbut I've never found it to be "productive" :)00:01
prologic*nods* agreed00:02
prologiccommit what you have so far and I'll take a look00:02
prologicthen we'll discuss :)00:02
DanielBairdokay stand by00:02
DanielBairdokay pushed up a starting doc.00:12
prologicsweet :)00:13
DanielBairdjust something to talk about, discussing by writing docs feels weird00:13
prologiclemme have a quick look00:13
DanielBairdmy editor recognised reST but doesn't seem to be helping with syntax highlighting at all, everything's all grey00:14
DanielBairdhmm i needed a blank line before my json source00:15
prologicok00:20
prologicfixed a few things00:20
prologicpip install sphinx00:20
prologicfab docs:view=yes00:20
prologicI'll probably publish this to readthedocs.org at some point when we have something worth publishing :)00:21
prologicactually I could publish itself as part of the deployment e.g: ccav.terranova.org.au/docs/00:22
prologicmy editor highlights rat okay00:22
prologicbut the code linking tool I use (which supports many many languages) doesn't recognize any of the sphinx variables or references00:23
prologicI"ll have to teach it what some of the sphinx directives are later :)00:24
prologicSphinx docs for reference: http://sphinx-doc.org/00:25
prologicmany many many many OSS projects use this :)00:25
prologichmm00:26
prologicSO00:26
prologicURI: /api/climatedata/<region_collection>/<model>/<scenario>/<year>00:27
DanielBairdyeah see you could argue a rest-pure version that said you should also have the region id in there, and the ui has to make many requests to get all the regions' info.00:28
DanielBairdbut given we never want to get a single region's info, that's not necessary00:28
prologicok00:29
prologicis there some kind of RESTFul spec on this?00:29
DanielBairdi'm proposing that a single "resource" that you can get a representation of, is actually a list of regions of the same type, and their deta for a single future climate poijnt00:30
DanielBairdon doing restful design generally?00:30
prologichttp://pve.proxmox.com/wiki/Proxmox_VE_API00:30
prologicProxmox VE API uses some00:30
prologicsee References at the bottom00:30
prologicI happen to quite like their RESTful API00:30
prologicTheir WebUI and their CLI tools all use the same identical REST API00:31
prologicyou can even query the REST API for what it can do, etc00:31
prologichmm00:33
prologicproxmox ve api refs seem to point to a book00:33
prologicdamnit :)00:33
DanielBairdyeah .. the api docs are a bit inscrutable00:33
DanielBairdbut they've got pretty complicated resources to expose00:34
DanielBairdhmm they're also doing some less-restful stuff like URLs that kick off backups etc00:36
prologichmm00:36
prologicafaik there is nothing wrong with that :)00:36
prologicanyway00:37
prologichrmm00:38
prologicare there any decent guidelines on this?00:38
DanielBairdnot very restful to have action names in your urls..00:38
prologicyes but PUT /api/backups/1?foo=bar00:38
prologicis just creating a new resource :)00:38
prologicwhich the backend picks up and starts a backup job00:39
prologicthat's restful :)00:39
DanielBairdyes that's sort of okay.  service/start is a bit less okay00:39
DanielBairdi guess the restful version would read and write the "started" property of a service object, or something00:40
DanielBairdi dont' have favourite doc, i guess the infoq post is a good intro: http://www.infoq.com/articles/rest-introduction00:41
prologicit in fact does though :)00:42
prologicI do agree though00:42
prologicRESTful web services should not perform "tasks"00:42
prologicthat's what RPC is designed for00:42
prologicbut if it creates a new resources or updates a resource00:42
prologicwhich then the backend later picks up to start a task00:42
prologicafaik that's RESTful :)00:42
prologichmm00:43
prologicthat infoq isn't anything I don't already know :)00:43
prologicguess I'm after a library / maybe00:43
prologiccircuits.web has no RESTful framework00:44
DanielBairdREST is a bit of a religion anyway, it's totally fine to do actions in urls if that's what you need to do.  it isn't RESTful but it's not necessarily failing at an API00:44
prologicbut supports all HTTP verbs00:44
prologicso you can create resources fairly easily00:44
prologic*nods*00:44
prologicalso regarding documentations tructure00:44
prologichave you seen http://docs.fabfile.org/en/1.6/index.html# ?00:45
prologicFabric's docs00:45
prologicI quite like their structure and would like to adopt it00:45
prologicit's a lot better IHMO than the structure of any of my project's docs in the past (circuits included)00:45
DanielBairdlooks okay, i can't see a way to get a full table of contents though, the ToC in the sidebar is just for the current page00:47
prologic*nods*00:51
prologicI have a ToC on the main index page right now00:51
prologicI just like their structure :)00:51
prologicbrb00:51
DanielBairdanyway i think you're doing fine with a restful API if you respect the http verbs (easy for us because we're not editing, just GETting), respect representations (also easy because we only need JSON, but that's easy to smoosh into HTML if you like) and have canonical URLs for each thing, and use the canon URLs as links to refer to things from inside other things.00:52
prologicback00:55
prologic*nods*00:56
prologicI was just curious is there was any standard or pseudo-standard on this00:56
prologicthat we can link to in the docs00:56
prologicbut anyway I'm fine with your approach thus far00:56
prologicWhat other key/value pairs does the API need to return?01:01
DanielBairdinside a region?01:02
prologicfor the "climate data" resource01:02
prologicyeah01:02
DanielBairdi presume it will be all the climate vars.. stand by i'll find a url01:02
DanielBairdhttp://www.worldclim.org/bioclim01:03
DanielBairdi thought there might be more than those 19 things.. we might invent some ourselves01:04
prologichttp://www.techrepublic.com/blog/10things/10-things-you-should-do-to-write-effective-restful-web-services/314801:04
prologichmm01:04
DanielBairdlol did you find that article by searching for "rest standard"01:05
prologicso we want the min/max/std/avg for all bioclim vars (that we have data for) across all regions?01:05
prologice.g: all ibra regions01:05
prologicin one big giant object returned to you?01:05
prologicyes :)01:05
prologicsort of01:05
prologichttps://www.google.com.au/search?sugexp=chrome,mod=14&sourceid=chrome&ie=UTF-8&q=RESTful+web+service+standard01:05
DanielBairdhere's someone else's climate data api:01:08
DanielBairdhttp://data.worldbank.org/developers/climate-data-api01:08
DanielBairdoh yeah, versioning.. cool kids have versioned apis, sometimes with a /v1/.. in the url01:09
prologicyeah01:09
prologicdo we want to support that?01:09
DanielBairdhmmm01:10
DanielBairdit's easy to.. do you expect your api to become an access point for more clients after this project?01:10
prologicI've never written a RESTful API that had more than one version before01:11
prologiconly just the one version (the latest)01:11
prologictbh - I don't know :)01:11
prologicI'm just thinking ahead01:11
prologicsome or all of this system will be used by the BCCVL project01:12
DanielBairdit's easy to drop a constant /v1/ into the urls for now01:12
prologicso in theory if they can reuse this project01:12
prologicwithout modification01:12
prologicthat would bd ideal01:12
DanielBairdno need to actually read the version number or anything, until you make a v201:12
prologicby without modification - I mean without writing from scratch01:12
prologicyeah I'm just conceptually thinking about how you'd do it in the first place01:13
prologic /api/v1/... vs. /api/v2/...01:13
prologicbasically becomes two separate endpoints01:13
prologicOR01:13
DanielBairdyes01:13
prologictreated as a parameter passed to each Resource01:13
prologiceither or method would work01:13
prologicbut IHMO01:13
DanielBairdi guess it depends on the difference between the versions01:13
prologictreating them as separate endpoints would be better01:14
DanielBairdyeah01:14
prologicless doing this in the logic:01:14
prologicif version == "v1"01:14
prologic   ...01:14
prologicelif version == "v2":01:14
prologic   ...01:14
prologicelse:01:14
prologic   ...01:14
prologicetc etc01:14
prologici.e: have a complete set of controllers that implement v1 of the API01:14
prologic-but- this does mean a lot of duplicated code though01:15
prologicI'm torn :)01:15
DanielBairdyes, if the differences are only slight you might want to add the extra logic to a resource handler, but drive space is cheap, might as well dupe the whole thing and keep them seperate.  if you're clever you could write your v2 handler to be able to invoke the v1 handler for some resources01:15
DanielBairdin ruby i'd use a Rack middleware to handle the versioning, and seperate the endpoints totally01:16
prologicthat is actually entirely possible01:16
prologiccircuits supports event handler inheritance and overrides01:16
prologicahh k01:16
prologicwell01:16
prologiclet's just stick with a static/constant v1 for now01:17
prologiclike you say01:17
prologicand worry about it later :)01:17
DanielBairdyeah that the v2 author's problem then :)01:17
DanielBaird* that's01:17
prologiclol01:17
prologicfuture James' problem :)01:17
prologicOk my original question01:17
prologicSo we want min/max/avg/std of all bioclim vars01:17
prologicfor all regions01:18
prologicin that resource01:18
DanielBairdyes, for a single point in time / modelspace01:18
DanielBairdand for just all regions of a given region type, i think01:18
DanielBairdhmm01:19
DanielBairdwe probably want region_type = "lga" etc in each region too01:19
DanielBairdthere's other region info that's general.. e.g. region boundaries01:20
DanielBairdso maybe it should be regions/ibra/climate/<model>/etc and also regions/ibra/<regionid>/boundary/ to get the region's polygon,01:21
prologichang on01:22
prologichold that thought01:22
DanielBairdbrb01:22
prologichttp://codepad.org/wT0yuEFT01:27
prologicIs this more what we're after?01:27
prologicAnd can I be as boring as name the key/value pairs as I see them in the raw data01:28
prologicAnd let you (the UI) deal with prettying them (perhaps via a GET /config.json request)01:28
DanielBairdyep whatever keynames you like01:31
DanielBairdso the summary info is "over all pixels in the region"?01:32
DanielBairdfor a given bioclim, i mean01:32
DanielBairdand maybe each region also should have a key of "bounds" that is a url01:33
prologichmm01:39
prologicyou want the bounds in the region objects?01:39
DanielBairdrest is a bit against the idea of fewer http transactions..01:39
prologicand yes this data is computed by intersecting the raster layers against all regions of the given region collection (ibra, nrm, lag, etc)01:40
DanielBairdbounds in regions would be okay but then i'll get a big polygon in every fetch, when really it doesn't change between future modelled points01:40
prologicand getting the min/max/std/avg01:40
prologicwell01:40
DanielBairdget the median as well in case that's better than the average, according to the users01:40
prologiccan we keep resources separate01:40
prologicuntil we run into performance issues with the "many http requests"01:40
prologicyeah sure I can give you the median as well01:41
prologicpossibly01:42
prologicit makes for better separation of concerns01:42
prologicand easier to write the API for01:42
prologicwe can always internally combine several or more resources together01:42
prologicinto a single view01:42
prologiclater on01:42
DanielBairdcrap i typed a long message that started with a /01:42
prologichit your up arrow key :)01:42
prologicusually most irc clients have a ^O option01:42
prologic^O/your thing that starts with a /01:43
DanielBairdseems to not work in adium01:43
prologicseems to not work with colluquay either :)01:43
prologicuseless01:43
prologicjust precede with a space :)01:43
DanielBairdanyway maybe this /ibra/<regionid>/climate/<model>/<scenario>/<year>/ gets just one of the "data" keys in your example01:44
DanielBairdthen /ibra/<regionid>/bounds/ gets the polygon01:44
DanielBairdand to be blessed by the rest angels, /ibra/<regionid>/ should get a region description full of urls to those other things01:45
DanielBairdand similarly, /ibra/<regionid>/climate/<model>/ should return a list of the scenarios and URLs to each scenario for that region and model, and /ibra/<regid>/<model>/<scenario>/ should return a list of the years and the URLs of them, etc etc etc01:47
prologicurls to where?01:50
prologicI liked your original idea though01:50
prologichere you're mixing climate data with region data01:50
DanielBairdurls to the children of the current url01:50
prologicI think we should keep those seprate01:50
prologicclimate data for a region01:50
prologicis computed by performing an intersection01:50
prologicof many raster layers (bioclims)01:51
prologicagainst the region boundary in question01:51
prologicco two resources would be good01:51
prologic /api/v1/regions/ibra/101:51
prologicfor example01:51
prologicand your original idea of01:51
prologic /api/v1/climatedata/ibra01:51
prologicbecause you want "summary climate data" of all regions of a given collection01:52
DanielBairdyeah that's pragmatically good01:54
DanielBairdjust indulging in rest-correctness for a bit..01:54
DanielBairdyou're supposed to be able to traverse a rest api without having to know any magic words01:54
DanielBairdaaaaaanyway.01:54
DanielBairdso let's have region info in a ../regions/<regionid>01:55
DanielBairdand climate info in a ../climate/<regiontype>/01:55
DanielBairdand make sure the url to the correct /regions/<regid>/ is in the ../climate/<regtype/ list for each region01:56
prologicahh okay01:57
prologicso the urls would point to children of the current level in the api01:57
prologiccool makes sense01:57
prologicin theory you should be able to also list what resources or sub-resources an endpoint has right?01:57
prologicthis is the kind of thing I was after01:57
prologic"best practices", etc :)01:57
prologiccool01:58
prologicI like this so faf01:58
DanielBairdyeah you are supposed to be able to start at one place and follow urls around that you get out of the response documents01:58
prologicshall we doco up our collected agreed upon thoughts so far?01:58
DanielBairdyep let's get it down01:58
prologicI pushed a few more changes to the docs btw (just tidying up)01:58
DanielBairdalright01:58
prologicI'll let you do that :)01:59
prologicI'm still messing around with supervisord01:59
DanielBairdokay i'll pull in your example01:59
prologicfor some reason it installs ok, but doesn't initially start up01:59
prologicbbs - lunch02:25
prologicback03:07
DanielBairdwe're including all scenarios, right, like SRES A1B etc, as well as RCPxx?03:49
DanielBairdwhat's bioclim std, is that the standard deviation?03:56
prologicyeah03:59
prologicumm03:59
prologicI believe so03:59
prologicI'm just going to deal with a small sample though04:00
prologicwhen we're develloping04:00
DanielBairdyeah just adding all the possibilities to the docs04:01
prologickk04:01
prologicwell just link to the wallace site04:01
prologickeep it terse :)04:01
prologichmm04:01
prologicwhy is r.js in fabfile/files/ ?04:02
DanielBairdonly fab uses it.  wasn't sure where it should live.04:02
DanielBairdit's possible to install requirejs as a node tool with npm, and that gets you r.js on the command line, but the require docs preferred just having the file somewhere04:03
prologickk04:03
prologicI'd prefer if we did04:03
prologicnpm install requirejs04:03
DanielBairdis possibly could just be in fabfile, but ..04:03
prologicI'll adapt the playbooks04:03
prologicit's a tool :)04:04
prologicI'll fix it up04:04
prologichaving a few issues with deploys actually04:04
prologicso in that space right now04:04
DanielBairdyes that would be fine too.  i think it's always available once it's a npm install04:04
prologicrepeated deploys are failing04:04
prologicand initial deploys aren't starting04:04
prologicgrr04:04
prologic*nods*04:04
prologicdon't worry yourself about it :)04:04
prologicI'll take care of that when I've fixed the above two issues04:04
prologicwhat's that saying04:05
DanielBairdinitial deploys aren't starting?04:05
prologic"if it ain't broke don't fix it!" :)04:05
prologicoh well04:05
prologicbut we need some kind of supervisor anyway04:05
prologicso that when the system reboots04:05
prologicthings come back up04:05
prologica simple daemon doesn't gig eyou that04:05
prologicyeah04:05
prologicfor some reason when I do a deploy04:05
prologicsupervisor dies or something04:05
prologicnot sure yet04:05
prologicinvestingating...04:05
DanielBairdoh right.. hmm well good luck04:06
prologicI'll sort it out :)04:06
prologicalways do04:06
prologic:)04:06
prologicthe other issue is a weird pip thing04:06
prologicI switched over to using circuits-dev04:06
prologicand it complains the 2nd time I try to deploy04:06
prologicthat it can't find a version spec for circuits04:06
prologic*shrugs*04:06
DanielBairdlemme know if you want me to modify the fabfile frontend thing to call global r.js, it's probaly just deleting the path but if it's fiddlier i can look into it04:06
DanielBairdmaybe we should re-write the back end in php04:07
DanielBairdLOL I'm still scarred from the stupid previous project that the guy before me had written in his own dinky php framework04:08
DanielBairdi had a month to fix up his year-long project, and ended up having to exclude almost every line of code he'd written04:08
prologiclol04:36
prologicPHP?04:36
prologicYou're joking right? :)04:36
prologicPHP == Personal Hacking Project04:36
*** DanielBaird has joined #cmvt04:40
prologicLOL04:40
prologic[root@test1 ~]# r.js --help04:40
prologicSee https://github.com/jrburke/r.js for usage.04:40
prologic[root@test1 ~]#04:40
DanielBairdcrap my Cmd-W get's away from me sometimes04:40
DanielBaird* gets04:40
DanielBairdstupid require docs :)04:40
prologicSolved one issue: https://github.com/Supervisor/supervisor/issues/23104:56
prologicActually might solve both issues I think04:56
DanielBaird"bug still exists in 3.0a12"??04:58
DanielBairdguess that guy's mistaken04:58
prologiche is very much mistaken :)04:59
prologicpeople are just generally stupid :)04:59
prologichaha04:59
prologicok that's harsh04:59
prologiclazy? :)04:59
prologicnaive :) ?04:59
prologicI installed 3.0a12 "correctly" :)04:59
prologicand rested it04:59
prologiccrash does not occur04:59
prologicI'll update to the latest when they release the next update to supervisors later04:59
DanielBairdyeah people are jerks.04:59
prologicAlso have to update to Python 2.7.5 too04:59
prologiclol05:00
prologicbut not me right? :)05:00
DanielBairdshould i say that our region polys are geojson MultiPolys?05:00
prologicyes05:00
prologicI will give you geojson05:00
DanielBairdyes of course, anyone in present company is always excluded from misanthropic statements05:00
prologicand yes they are MultiPolygon(s)05:00
DanielBairdi'm finally remembering to merge before i commit, so i don't have all these separate merge commits05:12
prologicSWEET05:12
prologicAll is well05:12
prologicand WORKING05:12
prologic:)05:12
prologicAll you need to remember is the following patern05:13
prologichg fetch05:13
prologicwhile not finished;05:13
prologic   hack;05:13
prologic   hg fetch05:13
prologichg push05:13
prologici.e: commit often05:13
prologicpull/merge (fetch) often05:13
DanielBairdbut if i commit then fetch, there's a second commit.. if I fetch/merge, then commit, that makes for a cleaner histopry05:16
DanielBairdanyway you must've pushed in the 20 secs between my fetch and my commit and push.. so you ruined it anyway05:16
DanielBairdhmm i screwed up links, apparently.05:18
prologicwell05:18
DanielBairdoh i forgot the trailing _05:18
DanielBairdhttps://bitbucket.org/ccaih/ccav/src/47fcd1e4c48563b10461e7ea274b7f3b0fc51f51/docs/source/webapi.rst?at=default05:22
DanielBairdno i screwed up the code blocks.  forgot the blank line05:23
DanielBairdfixing..05:23
DanielBairdbrb05:24
DanielBairdback. so yes the fetch will choke on uncommited changes, so I do a pull then merge, then commit changes and merge together.05:29
DanielBairdcrap I didn't fix the code blocks.05:30
DanielBairdhmmm05:30
prologicyeah no05:38
prologicwhat I'm saying is05:38
prologicyou should (I think) find it easier to follow the pattern I printed above05:38
prologicit means you always have a clem tree05:38
prologicwhen you do a fetch05:38
prologicwhich is required anyway05:39
prologicit just means less hand-merging you have to do05:39
prologicI almost never have to do a hand merge05:39
prologicunless there are merge conflicts05:39
DanielBairdah right, the example i was copying from was also broken.05:39
prologicin which case fetch bails out and lets you deal with it05:39
DanielBairdbitbucket's rst didn't like your source code block format05:40
DanielBairdtrying whatever it is that bb says, but if it doesn't work, screw them we can just use whatever sphinx uses05:41
DanielBairdyay it worked finally.  sheesh.05:41
DanielBairdum, yes but it's only manual in teh sense that i merge by typing "hg merge".05:41
prologicit's not for bitbucket05:42
prologicit's for sphinx05:42
prologicbitbucket does not use sphinx05:42
prologicit uses docutils (which is strict reStructuredText)05:42
prologicdon't' worry about it rendering on bitbucket05:42
prologicit won't05:42
prologicbut it will when it's built with sphinx and published somewhere05:42
prologicor viewed locally05:42
prologicdo:05:42
prologicfab docs:view-yes05:42
prologicto test/view the docs05:43
DanielBairddoes sphinx use sourcecode:: json as well as code-block:: json ?05:43
DanielBairdif it does, then we're fine05:43
prologicno05:43
prologicI don't think so05:43
prologicit's .. code-block:: <language>05:43
prologicbasically what I do is05:44
DanielBairdhmm that sucks.  who's idea was it to have conflicting rst formats05:44
prologicuse strict reStructured Text for the README.rst05:44
prologicso it renders properly on Pypi and Bitbucket05:44
prologicwhich does not use any rst extensions like sphinx05:44
prologicand everything in docs is sphinx built05:44
prologicand published05:44
prologicit's not conflicting05:45
DanielBairdanyway i don't actually care about bb's renderings, i was just using it for a link to send you to for easy reading05:45
prologicdifferent extensions support different directives, etc05:45
prologicahh05:45
prologicnps :)05:45
prologicI'm actually going to include the docs in future deploys by default now05:45
prologicat /docs/05:45
DanielBairdbut if bitbucket is stricter, maybe sourcecode:: will work in sphinx after all.05:45
DanielBairdbb doesn't show line numbers, but that's fine05:46
DanielBairdhmm i thought ``N``th would work, but it doesn't05:48
DanielBairdanyway it does it's job for now.  I went with .../api/v1/climatesummaries/...  and .../api/regions/... as the urls.05:49
DanielBairdactually i'll make one more change, and fix the ``n``th thing while I'm at it.05:50
prologickk05:51
prologicI guess as this has taken us nearly all day05:51
prologicI'l review in the morning and go from there05:51
prologicsorry :)05:51
DanielBairdnp there was a lot of doccing05:52
DanielBairdhad to happen some time :)05:52
prologicI know you're super excited to get backand/frontend integration happening :)05:53
prologicyeah I should write up some docs on how to install and deploy this too05:53
prologici.e: how to install ansible and run the playbooks05:54
prologicand how to run fab deploy05:54
prologic:)05:54
prologichaha05:54
prologicoh so hard!05:54
DanielBairdyour doc should be: install vagrant and vitrualbox, run vagrant up, copy that on your own server.05:55
prologiclol05:55
DanielBairdit's the classic "do as I do" instruction manual05:55
prologicyeah no :)05:55
prologica) vagrant only supports virtual box (for free)05:55
prologicb) environments are always different05:55
prologicdepending on the virtualization technology and infrastructure :)05:55
prologicI will still document it however :)05:56
prologicas an alternative installation method :)05:56
DanielBairdhence the last step, copy what happened then on your own server"05:56
prologici.e: local development/testing platform05:56
prologicos if you really do run your infrastructure on virtual box :)05:56
prologichaha05:56
DanielBairdbut anyway yes ansible / fab make docs short anyway.05:56
prologicwell05:56
prologicI was thinking05:56
DanielBairdideally your fab scripts check if each step was okay and fail nicely, but meh05:56
prologicshould the ansible playbooks do a default deploy?05:57
prologicto localhost05:57
prologici.e: the latest taggeg/relese version05:57
prologicwith a configuration domain, etc05:57
prologicor leave that as a separate step by the admins/developers05:57
prologicyeah fabric will fail if it doesn't find certain things05:57
prologicI mostly try to check for things05:58
prologicotherwise it'll fail horribly anyway :)05:58
DanielBairdshould do a deploy, if the installer won't need to un-deploy it, but can just re-config it to do what they wanty.05:58
prologichave you seen the doc string for fabfile/deploy.py ? :)05:58
DanielBairdyeah that's a very reasonable setup.  if they want soemthing different, it'll all be hand-fiddling anyway05:59
prologicwell06:00
prologicyeah06:00
prologicI think the plays could do a default deploy06:00
prologicwith some sensible defaults06:01
DanielBairdit shouldn't do a deploy if that would screw up an existing apache setup06:01
DanielBairdso if the deploy just puts stuff into a htdocs-location, that's fine any time, the admin just points their apache at it or whatever.. but we probably shouldn't automatically rewrite apache rules or anything.06:03
DanielBairddoes centos use the awesome apache setup where there's a sites-available directory of configs, that are symlinked to sites-active?06:04
DanielBairdor sites-deployed or something?06:04
DanielBairdactually no i bet it uses the shitty redhat mess of config files.06:05
prologicno it doesn't06:07
prologichmm06:07
prologicI see what you're saying06:07
prologic-however- all my ansible playbooks06:07
prologicand fabric tasks06:07
prologicassume you are deploying to a vm that has been provisioned with the playbooks06:07
prologicif not, you're kind of on your own06:07
DanielBairdhmm06:07
prologicthey'd have to follow the step-by-step instructions06:08
prologici.e:06:08
DanielBairdyes that's true, we've assumed all along a blank slate.06:08
prologicinstall these dependencies06:08
prologicdo this06:08
prologicdo that06:08
prologicetc06:08
prologicnot just a bank slate06:08
prologican automated setup06:08
prologicno hands on06:08
DanielBairdyes i'm flipping, lets continue to assume the deploy-to machine is dedicated to this only.06:08
prologic*nods*06:09
DanielBairdand there's an undocumented installation path which is basically check the stuff out yourself and do your own intergration, if you have an existing server.06:09
prologicbasically  I can't be bothered supporting a custom install06:09
prologicif anyone wants to go to that effort06:09
prologicI'll ensure enough documentation exists to help them out06:09
DanielBairdthat's when they contract you to fly in and do an install06:09
prologicbut I'm not doing anything beyond that06:09
prologicexactly :)06:10
prologiclol06:10
DanielBairdmust remember to make a humans.txt06:11
prologiccool06:11
prologicdocs deploy works perfectly :)06:11
prologiclol06:12
prologichumans.txt?06:12
DanielBairdlike a credits page.06:12
DanielBairdactually there's a whole site for it06:12
DanielBairdhttp://humanstxt.org/06:12
DanielBairdthat's new06:12
prologicyeah actually06:13
prologicyou know the footer stuff most websites/webapps include at the bottom06:13
prologicwe should do one of those for this app06:13
prologicsmall/tiny06:13
prologicwith our names06:13
prologicgriffith and jcu06:13
prologiclinks to the source code and internally hosted docs06:13
prologichaha06:14
prologicI love it06:14
prologichumans.txt06:14
DanielBairdthere will have to be somewhere to put the ANDS credits etc, it might end up as a slideup panel06:14
prologiclet's do that too :)06:14
prologicin addition06:14
prologic*nods*06:14
prologicslide up panel sounds good06:14
prologicyes ANDS funding credits should go there too06:14
prologicbut I'd also like to include this in the docs as well06:15
*** DanielBaird has joined #cmvt06:19
DanielBairdgah cnd-w'ed this chat again06:19
prologicok06:44
prologiccya tomorrow06:44
prologicwill switch to using globally installed r.js int eh morning06:44
prologicand review your api docs06:44
DanielBairdcool np06:44
DanielBairdsee you06:44
*** DanielBaird has quit IRC08:05
*** DanielBaird has joined #cmvt08:35
*** DanielBaird has quit IRC08:44
*** DanielBaird has joined #cmvt09:10
*** DanielBaird has quit IRC09:15
*** DanielBaird has joined #cmvt10:11
*** DanielBaird has quit IRC10:13
*** DanielBaird has joined #cmvt11:14
*** DanielBaird has quit IRC11:18
*** DanielBaird has joined #cmvt13:10
*** DanielBaird has quit IRC13:15
*** DanielBaird has joined #cmvt13:58
*** DanielBaird has quit IRC14:24
*** DanielBaird has joined #cmvt14:55
*** DanielBaird has quit IRC15:03
*** DanielBaird has joined #cmvt15:30
*** DanielBaird has quit IRC15:40
*** DanielBaird has joined #cmvt16:36
*** DanielBaird has quit IRC16:40
*** DanielBaird has joined #cmvt17:37
*** DanielBaird has quit IRC17:41
*** DanielBaird has joined #cmvt18:37
*** DanielBaird has quit IRC18:42
*** DanielBaird has joined #cmvt19:38
*** DanielBaird has quit IRC19:43
*** DanielBaird has joined #cmvt20:38
*** DanielBaird has quit IRC20:43
*** DanielBaird has joined #cmvt21:39
*** DanielBaird has quit IRC21:43
*** DanielBaird has joined #cmvt22:39
prologicMorn'n23:32
DanielBairdhowdy23:33
prologicFYI I'm not here tomorrow23:34
prologicHave to be at home to look after my daughter23:34
prologicbrb23:34
prologicback23:35
DanielBairdokay.  i think i get andrew and gerhard here the day after that23:38
prologicahh23:43
prologichmm23:43
prologicneed to get  a new desktop for home23:43
prologicours died on the weekend ;/23:43
DanielBairdback.  yes you said.. sounded like it's about time for a new thing anyway.  how's the budget for a replacement, got some money to spend?23:46
prologicwell23:49
prologicthat's the thing I'm tossing up on :)23:49
prologiciMac, Mac mini23:49
prologicIntel NUC23:49
prologicSomething custom23:49
prologicfitPC323:49
DanielBairdyeah the nuc looks nice but might as well buy a mac mini if you wanted that23:52
DanielBairdnot sure if i'd get an imac any more.  do you need power?  I love my 11" air, and it'll drive a pretty big monitor23:53
prologicwell I have an 11" Macbook Air at home23:55
prologiclast year's model23:56
prologicit's awesome23:56
prologicbut because of my vision impairment, using a laptop for long period of time hurts my neck :)23:56
prologicso I kinda also want to remain having a desktop at home23:56
prologicalso23:56
prologicI don't really want to get rid of Linux on the Desktop at home either :)23:56
prologicso I'm kinda learning towards CRUX (still) on some kind of replacement mini desktop (don't need/want the power)23:56

Generated by irclog2html.py 2.11.0 by Marius Gedminas - find it at mg.pov.lt!