Perl 6 - the future is here, just unevenly distributed

IRC log for #opentreeoflife, 2014-06-12

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

All times shown according to UTC.

Time Nick Message
00:16 josephwb joined #opentreeoflife
00:59 josephwb joined #opentreeoflife
01:15 mtholder joined #opentreeoflife
02:09 kcranstn joined #opentreeoflife
02:13 jimallman joined #opentreeoflife
02:22 mtholder joined #opentreeoflife
03:01 josephwb joined #opentreeoflife
04:29 mtholder joined #opentreeoflife
04:29 mtholder jimallman: I'm running a script to correct the errors in the index caused by having non-namespaced study Ids (the ids have been fixed in the Nexsons)
04:30 mtholder script will probably take an hour or so.
04:30 jimallman cool. i’ve been testing apache config, but will ignore oti weirdness for now.
04:31 mtholder sounds good. good night...
04:31 jimallman yes, good night
05:25 jimallman_ joined #opentreeoflife
09:10 jimallman_ joined #opentreeoflife
10:53 mtholder joined #opentreeoflife
11:06 kcranstn joined #opentreeoflife
11:31 josephwb joined #opentreeoflife
12:16 josephwb joined #opentreeoflife
12:20 josephwb joined #opentreeoflife
12:43 towodo joined #opentreeoflife
13:00 kcranstn joined #opentreeoflife
13:01 towodo dangit. someone tightened these screws way too tight. (on hotswap caddy for RAID)
13:02 towodo Attention: Whenever you screw in a machine screw, back off a quarter turn after it's fully tightened. Make life easier for the fool who has to unscrew it later
13:29 mtholder joined #opentreeoflife
14:04 mtholder joined #opentreeoflife
14:26 towodo kcranstn, files.opentreeoflife.org is now under version control https://github.com/OpenTreeOfLife/opentree/pull/309
14:37 kcranstn should I merge that, or wait for jim?
14:38 kcranstn jimallman, that is
14:45 mtholder joined #opentreeoflife
14:57 jimallman kcranstn: hi, looking over these changes now...
14:58 jimallman i feel funny about putting all this in the deploy directory, but i guess it gives us a backup
14:59 towodo we could put it in deployed-systems/
15:00 towodo like I said I'm reluctant to create a new repo for every little thing
15:00 jimallman no, that has a very different purpose IMO (not content, but configuration)
15:00 jimallman agreed that a new repo is overkill. i think it’s the name ‘files.opentree.org’ that’s throwing me.
15:01 towodo couldn't think of a better name.  would have called it 'ftp.opentreeoflife.org' but didn't want to run an ftp service.
15:01 jimallman in the nav, don’t we refer to these as “developer resources”?
15:02 towodo no, developer resources is a much broader category.  the files are a subset.
15:02 towodo e.g. developer resources include the git repos, trello boards, etc.
15:02 jimallman gotcha. so the name is just legacy cruft… but we don’t plan to actually use the domain name files.opentree.org, right?
15:03 towodo we are using the name, it's published, prety much
15:03 towodo what name would you use?
15:03 * jimallman is thinking…
15:03 towodo (that's opentreeoflife, not opentree)
15:03 towodo actually not much point in changing the name, not sure why you'd want to
15:04 jimallman if it’s an arbitrary subset, there might not be a “natural” name. ie, this is the collection of stuff formerly found at files.opentree.org…
15:04 towodo there are purls for anything important
15:05 jimallman sorry for the derail, let’s keep this for now. if it’s a pebble in our shoe, we can rename it later.
15:05 towodo I wouldn't say it's an arbitrary subset at all.  It's the developer resources that are static files provided by the project.
15:06 jimallman ah, this is progress (toward a name)!
15:06 towodo I have an errand to do… back around 12:30
15:22 mtholder joined #opentreeoflife
15:28 mtholder joined #opentreeoflife
15:29 jimallman kcranstn: i’ve got a couple of open questions on jar’s pull request, but no show-stoppers. do we need this deployed ASAP? if so, i can push from the new branch.
15:30 mtholder https://github.com/OpenTreeOfLife/opentree/pull/309
15:32 jimallman mtholder: i see the current development system installs most/all services on ot10 ('api oti treemachine taxomachine’), but ot11 still looks for neo4j services on ot12 (ie, $OPENTREE_NEO4_HOST = ot12). is this intentional, or just stale cruft in the config?
15:32 mtholder cruft, I think.
15:32 jimallman s/NEO4/NEO4J
15:33 mtholder it probably does not matter a whole lot (because the taxomachine/treemachine services are read-only operations), but I think that we probably want devtree to use devapi
15:34 mtholder errr. ot10
15:34 jimallman right. so it’s a happy coincidence that ot12 is up and running with treemachine and taxomachine… i’ll confirm with jar when he returns, then probably re-deploy.
15:35 mtholder as I recall, last week he set up ot11 and ot12 as a pair. Then we decided that we wanted a freeze on ot12 because the hackathon announcement points at it.
15:37 josephwb joined #opentreeoflife
16:08 mtholder so...
16:08 mtholder curl -L http://api.opentreeoflife.org/phylesystem/v1/study_list
16:08 mtholder works (because of the -L)
16:09 mtholder which allows redirection.
16:09 mtholder but curl http://api.opentreeoflife.org/phylesystem/v1/study_list does not
16:09 mtholder can we fix the routing on ot12, so that this does not require redirection?
16:10 mtholder would it be dangerous to add a second link to the phylesystem-api in the web2py/applications on the that machine?
16:11 mtholder towodo and/or jimallman ^
16:11 mtholder the redirection seems to be a pain for ajax too (or maybe that is my dumb JS skillset)
16:16 josephwb study curator app is still temporarily disabled? just checking.
16:16 mtholder hmmm.
16:17 mtholder jimallman^
16:18 jimallman pondering this now..
16:18 mtholder I'm getting the Blocked loading mixed active content "http://ot10.opentreeoflife.org/oti/ext/QueryServices/graphdb/findAllStudies"
16:18 mtholder message in firebug
16:19 jimallman yes, switch to HTTP for now. i’m testing a fix for this, using scheme-relative URLs like //ot10.opentreeolife.org/oti…
16:19 mtholder josephwb, if you know the study # the http://devtree.opentreeoflife.org/curator/study/view/pg_### urls seem to be working
16:20 mtholder works for me now
16:22 josephwb mtholder: this is not an existing study
16:22 mtholder jim's http rather that https seems to work
16:22 mtholder (for me)
16:24 jimallman apologies, the mixed-content problem is at the top of my list
16:27 josephwb what I get is: http://dev.opentreeoflife.org/opentree/static/bounce-to-devtree-curator.html
16:27 mtholder yes just follow that link.
16:27 josephwb so, data entry is fine there?
16:27 mtholder yup
16:27 josephwb ok, thanks
16:37 josephwb jimallman: name mapping is very slick!
16:38 jimallman ah, glad you like it!
16:38 josephwb is there a way to map everything in one go?
16:38 josephwb i have a tree of ~6000 taxa
16:39 josephwb it is a lot of "Approve all and continue"
16:39 josephwb i have 269 pages of taxa to go through
16:39 jimallman sorry, currently not since we don’t want to encourage reckless mapping… :)
16:39 mtholder I hope you're checking each carefully ;-)
16:39 josephwb grr
16:40 jimallman we’ve had notions of async mapping (a separate service), but i think that’s on hold.
16:40 josephwb ok
16:40 jimallman i suppose we could add an “I’m feeling lucky” checkbox here, and let the AJAX mapping just plow through everything.. perhaps a question for the call Monday.
16:41 mtholder It'd be nice to support sorting the list by the score of the match. look at the weird ones, and then click and "OK to the rest..."
16:42 josephwb yeah, that would be nice
16:42 jimallman interesting!
16:43 mtholder I guess autoboxComplete just returns a boolean "exact" field
16:44 mtholder though it returns > 1 entry for (some of ) the tough matches
16:44 josephwb i think the context selector might be better placed in the mapping box.
16:44 josephwb i missed it the first time through, as it was below off the page
16:46 josephwb and maybe even make it non-optional: a context must be selected. should help a lot with mis-mappings.
16:47 jimallman noted. thanks!
16:52 josephwb on the study curator main page, how hard would it be to be able to click on the column headers to sort studies?
17:00 josephwb very nice how you can click on an unmapped taxon, and it comes up highlighted in the tree
17:06 josephwb which version of OTT is being used for name-mapping?
17:12 jimallman study sorting would be easy. it’s currently done in the filter control..
17:12 towodo we're still at OTT 2.6 for everything
17:13 towodo I built a 2.8 draft today to make sure I still could, but have done no QC (haven't even reviewed log or deprecated list)
17:13 towodo (almost said WC instead of QC)
17:17 jimallman towodo: i was talking with mtholder earlier about the current ot11 config. it seems we’re installing treemachine and taxomachine on ot10, but ot11 still looks to ot12 for these services. is that intentional, or a bug in configuration?
17:17 mtholder joined #opentreeoflife
17:17 josephwb ok, thanks.
17:18 towodo jimallman, that's a bug, I thought it got fixed yesterday.
17:18 towodo thanks for catching that
17:18 jimallman hm, not pushed to master perhaps
17:18 josephwb for some reason the name mapping is not finding a taxon that is definitely in OTT2.6
17:18 jimallman i’ll pull again and make the change to ot10 if needed, then push to master and re-deploy
17:19 towodo did branch tweak-branches get merged?
17:19 josephwb For this study: http://devtree.opentreeoflife.org/curator/study/view/pg_2817
17:20 josephwb "Regulus calendula" is in OTT, but it is not being found.
17:20 towodo isn't found where?
17:21 josephwb When I "Start mapping"
17:21 josephwb Result: Failed mapping (edit or add hints)
17:21 towodo it unclassified.
17:22 josephwb so is "Regulus regulus", but it is found and mapped
17:23 towodo NCBI had it in an 'unclassified X' container in Jan. Maybe the TNRS suppresses those. It's no longer unclassified now.
17:23 josephwb that is why i am confused
17:23 josephwb good
17:23 towodo look at OTTFlags.java in taxomachine
17:23 towodo (sorry, I'm in the middle of something else)
17:24 josephwb i understand about the flags. but both taxa have the same flag, and only one gets mapped. :-S
17:27 towodo got me.  if you want to look at the January version of ncbi taxonomy it's in files.opentreeoflife.org/ott/2.6-inputs/feed/in/ncbi/taxdump.tar.gz ...
17:31 josephwb jimallman: we definitely need to do something about batch mapping. Even just increasing the number of taxa on a page from 20 to 100 (possibly a setting) would be helpful.
17:33 jimallman yes, pagination is definitely the path of least resistance.. up to 100, or 100, or Show All
17:33 josephwb word
17:35 jimallman I’ve updated the issue about bulk mapping: https://github.com/OpenTreeOfLife/opentree/issues/255
17:36 blackrim joined #opentreeoflife
17:36 jimallman aside from the quality-control issues, i was worried about performance with thousands of OTUs live in the UI. but we’ve switched to a more lightweight approach to Knockout, so i’m guessing it won’t be a problem.
17:49 josephwb jimallman: can a study not be saved if quality < 100%?
17:49 jimallman oh no, that shouldn’t stop you.
17:49 jimallman are you having trouble saving?
17:50 josephwb yes. do i have to do mapping to save?
17:51 mtholder joined #opentreeoflife
17:51 jimallman nope, we should happily save stuff in a very incomplete state. do you get an error when trying to save? what’s the behavior?
17:53 josephwb Sorry, there was an error in the study data.
17:53 josephwb {"description": "ot_validation failed: {\"annotationEvent\": {\"@preserve\": false, \"@description\": \"Open Tree NexSON validation\", \"@passedChecks\": false, \"@wasAssociatedWithAgentId\": \"peyotl-validator\", \"message\": [{\"@severity\": \"ERROR\", \"@code\": \"INCORRECT_VALUE_TYPE\", \"data\": [{\"expected\": \"integer\", \"type\": \"unicode\", \"key\": \"^ot:focalClade\"}], \"refersTo\": {\"@idref\": \"study\", \"@top\": \
17:54 jimallman ah, thanks…
17:54 jimallman mtholder: please take a look at validation error msg above
17:55 josephwb study is here: http://devtree.opentreeoflife.org/curator/study/edit/ot_8
17:55 jimallman if i’m reading this correctly, focalClade should be a true integer (or null), versus a string? wouldn’t this be coerced to integer? checking actual value now...
17:56 josephwb huh, that's not right
17:57 jimallman josephwb: thanks! i assume i’m seeing the previous value (null)
17:58 jimallman i just set this to ‘Mammalia’, which stored its ottid as “244265” (a string)
17:59 jimallman josephwb: did you set the focalClade while you were editing? if you’re comfortable with your browser’s Javascript console, you can report its current value like so:
17:59 jimallman > viewModel.nexml['^ot:focalClade']
18:00 * jimallman is not saving any of his changes, FYI
18:02 mtholder I did recently turn on the "reject if validation fails"
18:02 mtholder I can disable that or we can get the curator app coercing those #'s to strings.
18:03 mtholder I'd sort of like to avoid  much coercion server side.
18:03 mtholder that seems risky.
18:04 mtholder but I can do the server side coercion if it is tough to do in the app.
18:06 mtholder josephwb, do you have a lot of work in your browser's memory? I could disable validation and redeploy to let that study in...
18:06 josephwb not really. just uploaded the tree.
18:08 josephwb So, how to do the "Focal clade"?
18:08 josephwb i just loss my changes, so no need to turn things on/off
18:10 josephwb starting over.
18:10 josephwb got study saved by not setting Focal clade
18:11 jimallman sorry about that, josephwb.
18:11 josephwb no problem
18:12 jimallman mtholder, re: “we can get the curator app coercing those #'s to strings”, the curator app DOES store the focalClade value as a string, or null.
18:12 jimallman so not sure why it wasn’t handled/coerced to integer by the validator. maybe there was some other (unexpected) value there?
18:13 jimallman if you meant “get the curator app coercing those strings to #’s”, then i can definitely do that.
18:13 mtholder yes that is what I meant.
18:13 jimallman ah, OK. will do.
18:19 josephwb mtholder and jimallman: so, no way to set the Focal clade at the moment?
18:20 mtholder no
18:20 jimallman josephwb: i’m working on the fix right now, it’s trivial… should be ready to deploy in a few minutes.
18:20 jimallman can you save your work and back out soon?
18:20 josephwb great
18:20 josephwb yup.
18:21 josephwb i'm out
18:25 blackrim towodo: taxonomy question. in order to do more tests with new trees, we need to settle on the next taxonomy (so we don't have to retest everything). is 2.8 draft 4 the one that needs to be checked and if good will be the next one?
18:26 josephwb towodo: you asked me to remind you about refreshing the deployed treemachine code.
18:26 * jimallman is pushing changes now, will test next…
18:26 towodo I would think 2.8 draft 5.  is it ok if it's rough around the edges?
18:27 towodo give me 5 minutes to go over the logs, then I can push a new tarball if I don't see any egregious errors
18:27 blackrim ok
18:27 towodo it has a brand new refresh of ncbi taxonomy, and has species from IRMNG
18:28 towodo ok, hang on.
18:30 jimallman josephwb: you should be good to go (i’m coercing focalClade to integer before save). be sure to clear your browser cache to get the latest!
18:37 towodo # OTUs used but hidden has dropped from 5477 in 2.6 to 3023 in 2.8… good
18:38 blackrim towodo: that is good
18:39 towodo 1800 taxa have moved (changed parent) since 2.7 ...
18:39 mtholder thanks, jimallman!
18:40 towodo 19 used OTUs/taxa deprecated since 2.7 … e.g. Puccinia has been split… that could be a mistake but not worrying about it. 19 lost OTUs isn't so bad
18:40 jimallman mtholder: no sweat. thanks for clear validation messages!
18:42 towodo oh my… IRMNG has a gajillion Puccinias
18:42 josephwb jimallman: works. thanks.
18:42 blackrim towodo: perfect! I am sure there are a gajillion
18:43 towodo no, I mean homonyms at the genus level.
18:43 blackrim I see. i think it is just a fungal genus, but apparently not
18:44 towodo fungus, chordate, mollusc, plant
18:45 blackrim well that is definitely not right. hopefully the ones in those other groups are deprecated or synonyms or something
18:47 towodo no plant.
18:47 * blackrim is confused by towodo's fragment
18:48 towodo i was wrong about there being a plant Puccinia. was fooled by the -eae family name it's in
18:48 blackrim gotcha
18:48 towodo forgot fungi use that suffix too
18:49 towodo I only see one in the current IRMNG online. maybe I can get Tony to make me another dump.
18:49 towodo oh well, this is minutia.  2.8 looks fine, I will make a tarball now
18:50 mtholder so... putting functions that are not part of the @request.restful() bit of web2py's hackery does not work well.
18:50 mtholder I don't know how to get things like phylesystem_config to map to v1/phylesystem_config
18:51 mtholder (without our apache redirection which causes problems for some clients).
18:51 towodo I kludged one of those using apache redirection.
18:51 mtholder that is what make some ajax calls unhappy.
18:52 mtholder smart clients can deal with it, but the raw call fails using jQuery
18:52 mtholder makes the api a PITA until we figure out how to do it the web2py way.
18:52 mtholder I've spent about an hour on that, but the docs are horrible.
18:52 mtholder and (as always) the web2py way is only used by web2py.
18:53 mtholder "never reuse any prior art" that is the web2py motto.
18:53 jimallman :)
18:53 mtholder I think you mean :-(
18:53 towodo why not revise the front end and back end in concert?  forget about maintaining backward compatibility?
18:53 jimallman i make my emoticons the web2py way
18:53 mtholder lolo
18:54 mtholder minus the last o
18:54 mtholder what do you mean JAR?
18:54 jimallman let me dig into this a bit.. example API call that’s failing from jQuery?
18:54 mtholder one sec on that.
18:57 towodo maybe I don't know what the problem is. I thought you were moving URL x to v1/x . if you don't care that x works afterward, isn't it straightforward?
18:59 mtholder http://api.opentreeoflife.org/phylesystem/v1/study/pg_719
19:00 mtholder just a get on that use redirection.
19:00 mtholder maybe any js coder knows how to deal with it, but I'm a real novice
19:00 mtholder towodo. we get the v1 in the route by putting the study stuff in a v1 func with a restful decorator
19:01 mtholder we can't throw any old function in there.
19:01 mtholder if I understood web2py routes well, I'm confident that we can tell web2py to route v1/blah to the function blah
19:01 mtholder but I don't
19:02 mtholder and the fact that we use parameter routing in some apps constrains us to use that through out
19:03 towodo i should shut up, I don't know this stuff.  you are aware of what's in the apache vhost file, yes?
19:03 mtholder yes.
19:04 mtholder but if I rely on that, then all of my testing machines have to have the same apache modifications
19:04 towodo that was all a very hasty hack to support the urls given in the api documentation
19:04 mtholder or my test code has to swap routes
19:04 mtholder I know
19:04 towodo good point.  yes, not depending on redirects would be highly desirable.
19:04 mtholder it works fine, but I'm trying to get my tests working locally and remotely
19:05 towodo yep.  I made a start on that too
19:05 towodo this relates to the makefile business.  i want 'make' to work either locally or remotely
19:06 towodo blackrim, josephwb, OTT2.8draft5 is in the usual place, and linked from http://files.opentreeoflife.org/ott/
19:06 towodo IRMNG may need some attention but I think that's the one to use for now
19:10 mtholder jimallman. I have to run soon (and don't sweat the web2py routing stuff. I'll figure it out.
19:10 jimallman ok, but fwiw, i’m seeing a completely different problem:
19:10 mtholder the goal is to have all the buttons on this work: http://phylo.bio.ku.edu/status/status.html
19:10 jimallman missing ‘Access-Control-Allow-Origin’ header.
19:11 jimallman OK, grab me later, might have some solutions alrready.
19:11 mtholder great thanks
19:11 blackrim towodo: ok, will grab it and try it
19:11 blackrim towodo: not found
19:13 towodo try again now.  that vhost is on a different server & I hadn't deployed there before
19:15 blackrim got it
19:22 towodo jimallman, put some names in the credits file
19:22 towodo I did that is.
19:22 jimallman thanks!
19:23 towodo looked at it in browser, seems fine.  pull request ready for your action
19:23 josephwb blackrim: finishing purging files
19:23 josephwb finshed
19:28 jimallman towodo: Credits page looks good, merged (will re-deploy soon)
19:29 towodo I know the curators list is incomplete but I want to get the page up soon, since we're getting visits
19:30 towodo (need to figure out how to run webalizer or whatever on a machine that's different from the one where the logs are found)
19:30 jimallman hrm..
19:30 towodo that is, i think we should put the page up even if the curators list is incomplete, and then finish it in a second pass
19:30 jimallman on a related note, it might be good to capture snacktavish’s draft About pages..
19:30 towodo yes
19:31 jimallman maybe we can ask for a quick round of comments, then put up the latest.
19:31 snacktavish joined #opentreeoflife
19:33 josephwb i can grep the nexson files to get a list of curators
19:33 mtholder left #opentreeoflife
19:35 towodo snacktavish, did you curate any studies?  I didn't put you in the curators list, since I wasn't aware you had, but it can be changed
19:35 towodo oh wait, I *did* put you in
19:36 towodo jimallman, yes, we should give the people listed a chance to change their names, remove themselves, etc.  so put on dev site before pushing to prod site
19:39 josephwb towodo: "bryan drew",
19:39 josephwb "Jiabin Deng",
19:39 josephwb "William Wysocki",
19:39 josephwb "Romina Gazis",
19:39 josephwb "Chris Owen",
19:39 josephwb "Rick Ree",
19:39 josephwb "Evgeny Mavrodiev",
19:39 josephwb "Angela Oliverio",
19:39 josephwb "Laura Katz",
19:39 josephwb "Joseph Brown",
19:39 josephwb "Kristen Swithers",
19:39 josephwb "Dail Laughinghouse",
19:39 josephwb "Stephen Smith",
19:39 josephwb "Karen Cranston",
19:39 snacktavish @jimallman I never did actually! Just messed with testing ones.
19:41 josephwb those are all the curators involved with studies for the paper (synthesis and not)
19:42 josephwb for all studies in phlyografter, it is:
19:42 josephwb "Rick Ree",
19:42 josephwb "William Wysocki",
19:42 josephwb "bryan drew",
19:42 josephwb "please delete",
19:42 josephwb "Jiabin Deng",
19:42 josephwb joined #opentreeoflife
19:43 josephwb got disconnected somehow.
19:43 josephwb towodo: i hope that is useful.
19:43 josephwb if it got through.
19:44 towodo yes, thanks
20:12 kcranstn joined #opentreeoflife
20:14 blackrim joined #opentreeoflife
20:14 blackrim josephwb: wireless at ashleys is pretty good
20:16 josephwb sweet
20:17 josephwb i got the gcmdr repo down to less than half the size reported by towodo
20:18 blackrim that is good. so do we need to move it?
20:18 towodo why are you storing derived files in a repo at all?
20:19 blackrim where else? we need to refer to them for publication
20:19 towodo and for the nexsons, is it that you have qc'd particular nexson versions, and don't have a way to access particular nexson versions from phylesystem?
20:21 towodo i.e. you're afraid of losing them?
20:22 josephwb this might not be the best place for things, but this all predated phylsystem, and this was the workhorse for synthesis/testing
20:22 blackrim possibly . more that we want to be able to provide actual files for folks. i image we will want this for each synth. the actual artefact stored.
20:23 josephwb right: one can recreate all of the paper anALyses from these files and nothing more
20:24 josephwb WE can recreate all of the analyses
20:24 kcranstn agree that we want to store the artifacts for each synth; also agree with @towodo that github not the right place (not critiquing the current setup, i.e. making it work for version 1.0)
20:24 josephwb gcmdr is not currently set up to work with phylesystem
20:24 blackrim yeah ,happy for other solutions
20:24 jimallman towodo: OK for me to push HTTPS config to ot10? we need it for all services, and it looks like our apache config will handle the job.
20:24 josephwb +!
20:24 josephwb +1
20:24 towodo I understand the context. you can ignore my issue if you want (WONTFIX), I just find the space and speed an impediment when I'm dealing the asterales system. I want to script system setup end-to-end
20:25 josephwb +1!
20:25 towodo yes
20:25 towodo but it has to be easy to set up a server that doesn't have credentials.
20:25 towodo it's really the same thing as the gcmdr situation.
20:26 towodo I'm working toward a script that anyone can run that creates and runs the entire pipeline on the asterales system: taxonomy system, tree synthesis, browser, curator (without write).
20:27 towodo having a 300M gcmdr repo is not going to look friendly to people trying to do this.
20:27 josephwb agreed
20:27 josephwb takes me 9 seconds to clone, FWIW
20:28 josephwb this was not supposed to be permanent
20:28 josephwb just a place to collect files before sending to e.g. DRYAD
20:28 towodo no rush, I just filed the issue because it was an issue
20:28 towodo jimallman, the non-SSL version can wait until after SSL is deployed
20:28 towodo but I really would like it done soon, so I can my asterales thing.
20:34 jimallman understood. at this point, i’m setting up the whole system as “passive” (driven by the current request in the browser). if a user asks for HTTPS, we’ll try to do it. if they stick to HTTP, no harm done.
20:35 towodo well, if we use HTTP and credentials are passed, that might lead to harm, is the app competent to protect credentials by specifying HTTPS?
20:37 jimallman maybe i misunderstood you. do you want SSL cert + key to be optional as well? or just the GitHub OAuth stuff (registered apps)?
20:37 towodo all of it
20:38 jimallman got it. should i assume the two work in tandem (no OAuth without SSL)? or leave it up to the discretion of the admin to allow OAuth and logged-in sessions over HTTP (as we’ve been doing)?
20:38 towodo my requirement is: a random user on the internet should be able to set up an asterales system without having to get into relationships with a third party such as a cert authorizer, github, etc.
20:39 towodo a *working* system (modulo study modification)
20:39 jimallman right. so it’s all or nothing (SSL+Oauth, or neither)?
20:39 towodo I would say if there's no ssl then credentials shouldn't be passed (in the clear)
20:40 towodo so yes, no oauth without ssl
20:40 towodo we could allow it to be configured, but the default should be fail-safe
20:41 jimallman i’d agree. i believe i can force the issue with web2py’s auth config, forcing to HTTPS for logged-in users. for now, i’m just untangling the mixed-content issues.
20:41 towodo but if it's easier to just prevent oauth without ssl, do it that way.  *we* got benefit from that configuration, so it must have some value, but it's not terribly important
20:42 towodo and it is living dangerously.
20:42 jimallman indeed
20:44 towodo my intent is that the asterales system should be a way to draw people in, let them learn about & gain confidence in the system
20:44 towodo it's about the longer-term survival ofthe project.
20:44 jimallman right, so our site is a special case
20:48 towodo yes, at least we should think of it that way.
20:48 towodo I'm mainly thinking local instances for debugging and pull requests by people not on the project
20:48 towodo but parallel worlds with different taxonomies and studies ought to be doable
20:49 * jimallman nods
20:50 towodo something like that could conceivably happen at the end of the project, if it continues under different management.  or if it falls into neglect, some entrepreneurial soul ought to be able to revive it, without asking permission.
21:29 josephwb joined #opentreeoflife
21:31 josephwb joined #opentreeoflife
21:37 jimallman towodo: i believe i’ve got HTTPS working on ot10, and the study list working in FF and other “careful” browsers. (In other words, no mixed-content warning.)
21:39 towodo good.  I was never bothered by mixed content, never understood the attacks
21:39 jimallman the main threat is JS and other interpreted stuff that’s pulled in via HTTP, apparently a successful vector for session hijacking and the like.
21:40 jimallman dangit, looks like this broke the arguson fetch!?
21:40 jimallman https://devtree.opentreeoflife.org/
21:41 jimallman i’m using scheme-relative URLs to slip from HTTP to HTTPS and back. looks like MY code is tripping over the lack of ‘http(s)’… hm, fixing now.
22:44 kcranstn joined #opentreeoflife
22:45 towodo well it's ok to give up and always use either one or the other.
23:05 mtholder joined #opentreeoflife
23:25 jimallman it’s good, but phylopic is failing in HTTPS (i’ve submitted a ticket).. no biggie, we can do without critter pics i guess :-(
23:46 kcranstn joined #opentreeoflife
23:54 towodo joined #opentreeoflife

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