Perl 6 - the future is here, just unevenly distributed

IRC log for #opentreeoflife, 2014-06-05

| Channels | #opentreeoflife index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:38 mtholder joined #opentreeoflife
01:24 josephwb joined #opentreeoflife
01:52 kcranstn joined #opentreeoflife
02:05 mtholder joined #opentreeoflife
02:11 towodo joined #opentreeoflife
02:41 josephwb joined #opentreeoflife
03:44 mtholder joined #opentreeoflife
04:03 mtholder left #opentreeoflife
04:22 josephwb joined #opentreeoflife
05:17 mtholder joined #opentreeoflife
05:58 travis-ci joined #opentreeoflife
05:58 travis-ci [travis-ci] OpenTreeOfLife/phylesystem-api#544 (treesource - 3410837 : Mark T. Holder): The build passed.
05:58 travis-ci [travis-ci] Change view : https://github.com/OpenTreeOfLife/phylesystem-api/compare/6cf5a668d2ba...34108375c403
05:58 travis-ci [travis-ci] Build details : http://travis-ci.org/OpenTreeOfLife/phylesystem-api/builds/26824701
05:58 travis-ci left #opentreeoflife
10:50 josephwb joined #opentreeoflife
11:13 josephwb joined #opentreeoflife
11:57 kcranstn joined #opentreeoflife
12:01 towodo joined #opentreeoflife
12:11 josephwb joined #opentreeoflife
12:41 josephwb joined #opentreeoflife
12:44 travis-ci joined #opentreeoflife
12:44 travis-ci [travis-ci] OpenTreeOfLife/phylesystem-api#545 (master - 1392ca6 : Jonathan A Rees): The build passed.
12:44 travis-ci [travis-ci] Change view : https://github.com/OpenTreeOfLife/phylesystem-api/compare/8d7fd577e404...1392ca66defa
12:44 travis-ci [travis-ci] Build details : http://travis-ci.org/OpenTreeOfLife/phylesystem-api/builds/26849632
12:44 travis-ci left #opentreeoflife
12:46 blackrim joined #opentreeoflife
13:23 lcoghill joined #opentreeoflife
13:28 sckott joined #opentreeoflife
13:59 josephwb joined #opentreeoflife
15:15 kcranstn joined #opentreeoflife
15:34 mtholder joined #opentreeoflife
16:06 mtholder joined #opentreeoflife
16:22 mtholder joined #opentreeoflife
17:04 pmidford joined #opentreeoflife
17:56 kcranstn joined #opentreeoflife
18:04 towodo kcranstn, I hope you aren't waiting for me for anything on the api doc… I just tried to gather enough info to get you started
18:05 kcranstn nope, not waiting for you. Just me. ;)
18:06 towodo to answer your earlier question the URL names are all up for grabs - can be mapped using apache redirects or rewrites
18:06 kcranstn did that link for activating the SSL cert work?
18:07 towodo no. i'm waiting to hear from support. will get on chat with them if i don't hear soon. they told me it could be done
18:07 mtholder towodo: what is up with the routing on http://api.opentreeoflife.org/phylesystem/v0/study/pg_10 are we intentionally routing v0 to v1 ?
18:08 mtholder seems confusing
18:09 mtholder fwiw, I'd prefer to keep the routing tricks to a minimum - it will be a barrier for other developers to grok the code if the deployed routes are usually different from the routes used on a local, out-of-the-box installation.
18:10 towodo The question is what will the API documentation look like. We haven't really discussed this. I was saying all of the APIs exposed to outside users are v0. We could say v1.  Right now phylesystem/v0 -> api/v1 for certain URLs, api/ for others (api is not consistent).
18:11 towodo Can we change the phylesystem-api application name in URLs to 'phylesystem'?  And add /v1/ where needed?
18:11 mtholder sure that is not hard.
18:11 mtholder the name is just the link name on installation.
18:12 mtholder Let me merge another branch so, I don't get conflicts when I move the functions inside v1.
18:12 towodo right, i think that's isolated to the config files and maybe some of the install scripts
18:12 mtholder but are we going with v0 instead of v1
18:13 mtholder ?
18:14 towodo Right now Karen has review authority on the URL choices. I sort of prefer v0 since this is all so preliminary but I see your point.
18:14 jimallman joined #opentreeoflife
18:14 kcranstn I don’t really care about v0 vs v1, as long as we are consistent
18:15 jimallman towodo: sorry i’m late, lots going on here
18:15 towodo hmm. consistency would mean v1 I think since the phylesystem-api code wires it in...
18:16 towodo I think I need to set up a branch for us all to use so that master doesn't get trashed during the mss review period.
18:16 mtholder It is wired in now, but about 7 character replacements to change it... ( on the server side - probably a few config edits on for the curator atpp).
18:17 towodo it's all over the documentation I think
18:17 mtholder well the new docs say v0 and the api docs say v1, so we're in for a lot of editing regardless.
18:17 * jimallman is catching up via IRC logs…
18:18 towodo there's only one new doc.  find&replace
18:18 towodo and I'm happy to change the redirects, they're isolated to the apache config, and no one uses the new urls
18:18 jimallman to me, v1 makes more sense if we’re planning to “stand pat” with the current APIs for the foreseeable future.
18:18 mtholder sticking with v1 is fine with me.
18:19 jimallman otherwise it’s like google products that stay in “infinite beta”…
18:19 towodo v1 is ok but there are so many rough edges I think we should expect a v2 in about 3 months.  a matter of setting expectations.
18:20 mtholder s/months/years/ and I agree
18:20 jimallman i see. if we’re planning a follow-up surge on API, v0 would communicate that this is preliminary (Not Ready for Prime Time)
18:20 towodo e.g. we've got inconsistent case and naming conventions
18:20 jimallman i see what you mean. cleanup wouldn’t be too hard, but it’ll break lots of existing calls.
18:21 mtholder and we'll always have them, I'm afraid. in a few months (hopefully) we'll have outside client code.
18:21 towodo you can't design a good api in a week.  i was defining v0 = what we tell hackathon participants tomorrow, pre-release, then api for a wider announcement
18:21 towodo s/api/v1/
18:21 mtholder outsiders will be more annoyed by changes to casing than inconsistency that works a month later.
18:21 towodo well, then two weeks.  I don't think we'll have something by tomorrow that we'll want to live with for a year.
18:23 kcranstn let’s not agonize over v0 vs v1
18:23 mtholder true, but we don't want to be in the last of the grant either. Some things we can't change. We can spend lots of dev time making api more consistent, or we can add features. I have a feeling that I know which will seem more pressing...
18:23 kcranstn consistency! ;)
18:24 * jimallman is a known agonist
18:24 kcranstn there are a few in this project - you are not alone
18:28 towodo there are 2 issues, naming and expectations. re expectations… the doc for the hackathon has to say this is preliminary and a vn+1 may come along soon.
18:28 towodo re naming, there are 2 notes for v1 and 1 vote for v0, and one abstention.  the v1s have it
18:29 towodo s/notes/votes/
18:29 kcranstn I will manage expectations
18:29 kcranstn Manager of Expectations
18:30 towodo I'm concerned about managing and naming github branches now.  until I've been the only one tweaking the post-'dev' machines. now there will be many. there needs to be a branch for post-'dev' (ot11 + ot12).
18:32 mtholder I thought the deployed-sytems repo let us decouple the deployed branc from the machine (so we would not have to have a branch in each repo for each machine).
18:32 mtholder we just record what branch is deployed in the config files.
18:32 towodo yes, but since we're doing collaborative development we need to agree with one another which branches we will use for communication.
18:33 mtholder so one "production" branch in each repo? or do you think we need more than that?
18:35 towodo that's fine but it's very odd to call out active development branch 'production'.  or we could continue using master, and set up a set of branches to freeze our currently deployed production system (dev.opentreeoflife.org …)?
18:36 kcranstn I thought the active dev branch was master
18:36 kcranstn and production = deployed
18:37 mtholder yeah. that is what I meant (master= dev. production lags behind, but reflects what is deployed)
18:38 towodo right now active dev = deployed = master.  we are talking about switching our testing from ot3/ot9/ot10 to ot11/ot12. we need one set of branches corresponding to each configuration.  later when we have a 'production' release we'll need yet another set of branches.
18:39 towodo this is because as stephen points out we need stability for dev.opentreeoflife.org  until the production system is ready
18:41 towodo maybe create a set of branches for the current dev setup, that will be stable, changing only to repair bugs ?  could be called 'production'
18:42 towodo because really it's dev.opentreeoflife.org that's our production setup right now
18:42 mtholder that sounds reasonable to me.
18:42 kcranstn agreed
18:43 towodo ok.  jimallman is short on time so I will do this after I talk to him
18:43 jimallman towodo: i’ve removed my devtree changes from ‘prod-setup’, for safety’s sake, and created a new ‘dev-to-devtree’ branch:  https://github.com/OpenTreeOfLife/deployed-systems/compare/dev-to-devtree...prod-setup
18:44 jimallman the latter isolates change for the new devtree subdomain, and the related app (and client ID) on GitHub
18:45 towodo now i have to figure out how to sync my prod-setup branch from github
18:45 towodo git merge origin/prod-setup   I think
18:46 jimallman yes, or just git pull origin prod-setup (should fetch recent changes without damaging local tweaks)
18:46 towodo ok. I registered two apps at github
18:46 towodo I need to do the config and create the secret files
18:47 towodo it's confusing for the directories to be called 'development' and 'production' now...
18:50 jimallman ? this was intended to describe a set of working, multi-server systems in common terms (dev, staging, production) …
18:51 towodo I should have called my directory something other than production
18:51 towodo I was sort of optimistic...
18:52 jimallman towodo: i can transfer the new (devtree) curator app to OpenTreeOfLife, if that would help. but yeah, i haven’t registered its companion tree-viewer app for devtree
18:52 jimallman s/transfer/transfer ownership of/
18:53 towodo I'm staying away from devtree for now, I just want tree/api to work.  And I already did github setup for those apps
18:53 jimallman i see, the production apps. of course.
18:54 towodo ok. so first I need to set up the TREEVIEW, so the main page will work (that's needed for comments, right?… would be nice if the app worked without the secrets setup)
18:54 towodo http://tree.opentreeoflife.org/
18:55 towodo => 500
19:00 jimallman do you need the graphic “badge” for these? i’ve added it here (upload this during GitHub app registration): https://github.com/OpenTreeOfLife/opentree/blob/master/deploy/github-registration/OTOL-badge-140px.png
19:01 towodo hmm. I guess so
19:01 jimallman ..but back to the bug… just a sec
19:04 towodo i'm assuming that installing the secret will make the 500 go away
19:04 towodo which is what i'm doing now
19:05 towodo we need to be careful not to let the API on ot12 push to github yet.
19:06 jimallman noted
19:08 towodo that didn't work.
19:08 jimallman you’ll also want to update both CLIENT_ID vars in ot11.config, to match the client IDs of your newly-registered apps on GitHub
19:09 towodo Isn't the github bot user 'opentreeapi' used only by the api?  Its token file is pushed in push.sh in push_opentree, which seems like the wrong place
19:10 jimallman i believe we’re also using this for “anonymous” comments in the synth-tree explorer
19:11 towodo oh… foo… will separate out the setup as a subroutine and call it both places
19:13 towodo so, any thoughts about the 500?
19:13 jimallman not sure, the curator app loads properly: http://tree.opentreeoflife.org/curator
19:13 towodo not even sure how I'd start to track that down.  not a web2py error, so won't be in errors/ directory
19:14 jimallman i don’t see anything in web2py logs, nor any errors/ dir here. which is weird.
19:14 jimallman oops, i take that back.. found webapp/errors
19:15 mtholder joined #opentreeoflife
19:15 jimallman with lots of goodies
19:16 jimallman it’s an Unauthorized (401) error from the local-comments plugin. so yeah, this suggests a problem with OAuth login on GitHub
19:16 jimallman looking at config now
19:18 jimallman towodo: github_client_id = 5600970c523d057ae6ff … does this match the Client ID of the registered tree-explorer app for ‘tree.opentreeoflife.org’? we’ll want to check that, and for a matching secret in webapp/private/GITHUB_CLIENT_SECRET
19:19 towodo I just installed that.
19:19 towodo was going to do curator after treeview worked.
19:20 towodo oh… hang on… let me double check
19:20 jimallman right, i’m looking at the treeview config (opentree/webapp/private/)
19:21 towodo client id should be 5600970c523d057ae6ff.  that's what the config file says (local checkout)
19:21 towodo hadn't installed yet.  doing so now
19:24 towodo ok… new id installed and apache restarted… still getting 500s
19:25 towodo checked secret in private/, looks OK
19:26 jimallman i’m seeing the same Unauthorized error, checking localcomments plugin code now...
19:27 towodo the bot user isn't set up yet
19:27 jimallman for both client-id and secret, make sure they match the TREEVIEW app and not CURATION app
19:29 jimallman hm, looks like we’re missing the OPENTREEAPI_OAUTH_TOKEN file, should be in /home/opentree/.ssh
19:30 towodo right, that's what I meant by "the bot user isn't set up yet"
19:30 jimallman yeah, this should have gone up during push_opentree
19:30 towodo was looking at doc to figure out how to get token
19:31 towodo don't know what user name to use, doc doesn't say
19:31 jimallman i’m guessing this is the problem. we should be able to use the existing bot user for all hosts
19:31 jimallman https://github.com/opentreeapi
19:31 towodo I'm a broken record about these dependencies,you know what i want to say
19:31 jimallman the token is one of those secure files that’s just shared among deployment users… yes, more docs!
19:32 towodo no, I mean the site should work (milidly degraded) without privileges or secrets
19:32 towodo we want people to be able to fork & hack easily
19:32 jimallman ah yes. a worthy goal.
19:33 towodo hmm.  so i should copy the token from another server, rather than request it anew using curl
19:33 towodo does that account have a password?
19:33 towodo does anyone other than duke know what it is?
19:35 jimallman yes, we (kcranston and i) got the credentials from duke.
19:35 jimallman meanwhile, here’s how to get the current token file:
19:36 jimallman $ cd ~/.ssh/opentree
19:36 towodo copied oath token from ot3
19:36 jimallman $ scp ot7:/home/opentree/.ssh/opentree/OPENTREEAPI_OAUTH_TOKEN .
19:36 sckott left #opentreeoflife
19:36 jimallman towodo: i’m sure that’s fine then
19:36 towodo pushing
19:37 jimallman ah, that’s better already
19:37 towodo It would be nice to be able to display a settable news banner when the app starts up
19:38 towodo yes.  spinning
19:38 jimallman i’ll add a ticket for settable news banner..
19:39 jimallman meanwhile, spinning because of an unnamed node (not caught by argus’s alpha-sorting logic)
19:39 towodo running 'top' on ot11 and ot12 shows no activity
19:39 towodo search box works!  'Rana' goes to isolated node
19:40 towodo 'Homo' leads to spinning
19:40 jimallman yes, there’s something very weird about the arguson data here
19:40 towodo cool.  'about' page works
19:41 jimallman it’s a JS error.. here’s what i’m seeing for a typical value in node.children:
19:41 jimallman "org.neo4j.server.rest.repr.ArgusonRepresentationConverter$1@3dc3740a"
19:41 towodo bibliographic references works!
19:41 towodo humph.
19:41 jimallman it looks like we’re getting a callable instead of the result of a call to it  :-/
19:42 towodo looking at https://github.com/OpenTreeOfLife/treemachine/commits/master
19:43 jimallman same here.. maybe a change in this commit?  https://github.com/OpenTreeOfLife/treemachine/commit/c43eb9fc5ce6df9e4514e373d1cca6f82c78a94e
19:44 towodo the offending class is mentioned here https://github.com/OpenTreeOfLife/treemachine/commit/aa21cb6cf42b87eefbb169553de748f803ffaa58
19:45 towodo If we can isolate a POST request we can hand this off to Cody to debug… the server is just running the latest version of treemachine
19:45 towodo do you have that curl call handy?  (the first one the webapp does to get arguson)
19:46 towodo s/first one/first POST/
19:47 jimallman just a sec, i can retrieve from the browser...
19:48 jimallman building an email now…
19:48 jimallman no, a GitHub issue
19:48 towodo on treemachine?  with reference to ot12 for reproduction.  great
19:50 towodo OK, I think we're in pretty good shape.  I just need to set up the webhook and the curator git identity
19:52 jimallman https://github.com/OpenTreeOfLife/treemachine/issues/94
19:52 towodo how did you get it from the browser? oh i see, debugger, network, params
19:53 towodo hey i even see the CORS preflight request ! I've read about those but never seen one
19:55 towodo jimallman, thanks for your help. i'm going to get something to eat now
19:55 jimallman Chrome dev tools > Network tab… this has an excellent option to “Copy as CURL”
19:55 towodo omg.
19:55 jimallman great. i’m heading to the hospital, will try to get back online there
19:55 jimallman ??
19:55 towodo omg = wow, re copy as curl.
19:56 jimallman *phew*
19:56 towodo hope all's well there, although i guess it's not, by definition
19:56 jimallman note that its curl is super verbose and (strangely) sometimes omits the -X POST option
19:57 jimallman not a crisis, mostly moral support. i appreciate everyone’s flexibility! i should be back to normal routine by Fri morning.
20:31 pmidford joined #opentreeoflife
20:41 josephwb joined #opentreeoflife
20:50 mtholder joined #opentreeoflife
21:06 travis-ci joined #opentreeoflife
21:06 travis-ci [travis-ci] OpenTreeOfLife/phylesystem-api#547 (production - 1392ca6 : Jonathan A Rees): The build passed.
21:06 travis-ci [travis-ci] Change view : https://github.com/OpenTreeOfLife/phylesystem-api/compare/production
21:06 travis-ci [travis-ci] Build details : http://travis-ci.org/OpenTreeOfLife/phylesystem-api/builds/26891448
21:06 travis-ci left #opentreeoflife
21:07 travis-ci joined #opentreeoflife
21:07 travis-ci [travis-ci] OpenTreeOfLife/phylesystem-api#546 (prod-setup - 726f119 : Jonathan A Rees): The build passed.
21:07 travis-ci [travis-ci] Change view : https://github.com/OpenTreeOfLife/phylesystem-api/commit/726f1197fea2
21:07 travis-ci [travis-ci] Build details : http://travis-ci.org/OpenTreeOfLife/phylesystem-api/builds/26891425
21:07 travis-ci left #opentreeoflife
21:09 josephwb joined #opentreeoflife
21:36 mtholder joined #opentreeoflife
22:03 jimallman joined #opentreeoflife
22:05 josephwb joined #opentreeoflife
22:23 mtholder joined #opentreeoflife
22:31 travis-ci joined #opentreeoflife
22:31 travis-ci [travis-ci] OpenTreeOfLife/phylesystem-api#549 (master - ec30ccd : Mark T. Holder): The build passed.
22:31 travis-ci [travis-ci] Change view : https://github.com/OpenTreeOfLife/phylesystem-api/compare/1392ca66defa...ec30ccd5ad0b
22:31 travis-ci [travis-ci] Build details : http://travis-ci.org/OpenTreeOfLife/phylesystem-api/builds/26897010
22:31 travis-ci left #opentreeoflife
22:40 mtholder towodo: I think the code currently deployed on ot3, ot9, and ot10 is a bit behind the "master" (I only checked phylesystem-api on ot10) because people don't necessarily deploy every time they commit, or merge a pull request.
22:41 mtholder we may want to redeploy, or note that they are not exactly working with the "production" branches yet.
23:06 josephwb joined #opentreeoflife
23:11 josephwb joined #opentreeoflife
23:25 josephwb joined #opentreeoflife
23:54 josephwb joined #opentreeoflife
23:58 josephwb joined #opentreeoflife

| Channels | #opentreeoflife index | Today | | Search | Google Search | Plain-Text | summary