Time
S
Nick
Message
00:48 axfelix joined #dataverse
07:06 bencomp joined #dataverse
07:55 jri joined #dataverse
12:09 donsizemore joined #dataverse
12:56 jri joined #dataverse
13:28 djbrooke joined #dataverse
13:30 jri joined #dataverse
14:12 jri_ joined #dataverse
14:14 donsizemore joined #dataverse
14:18 pdurbin
donsizemore: shibboleth accounts are not different than local accounts with regard to permissions. And good morning. :)
14:22 jri joined #dataverse
14:22 donsizemore
@pdurbin morning =) i'm shaking my fists at Condor this morning, then updating the Shib metadata for dataverse-test.irss.unc.edu (and getting that typo corrected)
14:37 pdurbin
donsizemore: ah, we use Condor too: http://projects.iq.harvard.edu/rce/book/working-your-jobs
15:10 donsizemore joined #dataverse
15:10 axfelix joined #dataverse
15:11 donsizemore
@pdurbin knock knock?
15:16 pdurbin
who's there?
15:23 donsizemore
@pdurbin the interrupting panda!
15:24 donsizemore
@pdurbin quick Q: you sent me a specific war file to use for Shibboleth testing, 4.3.1 build 28
15:24 donsizemore
@pdurbin but Akio encountered some trouble with upload requirements, so he had me deploy 4.3.1 build 94
15:25 donsizemore
@pdurbin which broke the Shibboleth login UI (okay, fine) and my plan is to redeploy the war you sent me, and wipe out another DVN4 test machine to deploy 4.3.1 proper
15:25 donsizemore
@pdurbin unless you tell me to sit tight pending any pending release or other test war of which i'm not aware?
15:26 donsizemore
@pdurbin this is all coming about because he's working on a dockerized Dataverse publishing component https://github.com/DICE-UNC/publisher
15:28 donsizemore
@pdurbin (the problem he ran into with build 28 being that the data upload Description field was being detected as null so he couldn't upload any test data)
15:28 jri joined #dataverse
15:33 jri_ joined #dataverse
15:33 pdurbin
build 94... what's the git commit sha?
15:35 donsizemore
https://github.com/IQSS/dataverse/commit/be5b26ec4b873ae07f76584a9f3a156e91a6f5c1
15:35 donsizemore
(it's the latest release proper)
15:38 pdurbin
right, right, we're running v. 4.3.1 build 94-be5b26e in production at https://dataverse.harvard.edu ... you're saying there's a bug in this release? I'm confused.
15:39 donsizemore
@pdurbin eek! must run! but the bug was in the earlier "shib" version you sent. was going to downgrade that box since the shib stuff was working and give akio 4.3.1 proper on another machine
15:39 donsizemore
@pdurbin just checking to see if that sounds solid with you or if there's another, newer war you'd like us to test
15:43 pdurbin
https://build.hmdc.harvard.edu:8443/job/phoenix.dataverse.org-build-develop/ is the job I'm using to build from the "develop" branch. It produces the war file I deploy to http://phoenix.dataverse.org for integration testing (using API calls) of the latest code in "develop".
16:11 axfelix joined #dataverse
17:20 metamattj joined #dataverse
17:50 bencomp joined #dataverse
17:52 donsizemore joined #dataverse
18:34 balo joined #dataverse
18:47 djbrooke joined #dataverse
19:00 donsizemore joined #dataverse
19:03 pameyer joined #dataverse
19:04 axfelix joined #dataverse
19:50 pdurbin
pameyer: hi! I should send that JSON via POST, right? (As opposed to PUT.) To what path? http://dcm.example.com/path/to/resource ?
19:50 pameyer
pdurbin: you noticed that the repo was empty ;)
19:50 pdurbin
I didn't find my answer there. :)
19:51 pameyer
it was a philosophical statement about the contemplation of emptyness :)
19:52 pdurbin
lol
19:52 pameyer
just kidding - I can check now if you're taking a look
19:52 pameyer
it'll become non-empty reasonably soon
19:52 pdurbin
nice!
19:52 pdurbin
pameyer: does that mean you'd like me to run it myself?
19:53 pameyer
I was assuming you might want to, at least for testing
19:54 pdurbin
It would be nice to have something to test against to see what the expected responses are.
19:55 pdurbin
pameyer: if you want to host a test instance for me it would be less work for me :)
19:55 pameyer
especially if we're thinking on having it talk to something inside our vlan
19:55 jri joined #dataverse
19:55 pameyer
I'd thought you were going to be busy with other stuff this week
19:56 pdurbin
I can go work on something else if that's better. :)
19:57 pameyer
but back to your original question, I was surprised to find that the request method isn't currently set :<
19:57 pameyer
so go with POST
19:57 pdurbin
POST it is
19:57 pdurbin
and the path?
19:57 pameyer
"/ur.py"
19:58 pdurbin
huh. you're the boss
19:59 pameyer
in the fortran naming scheme, that translates to "/upload_request_handler.py"
19:59 pdurbin
heh. ok
20:00 pdurbin
pameyer: should the email address sent to the DCM be the email address of the user who just created the dataset? I assume so.
20:01 pameyer
it won't be sending emails anymore - so the dv username might be better
20:01 pdurbin
could we use the id from the authenticated users table instead? 42 or whatever?
20:02 pameyer
we need an identifier for the user and dataset - but I think we've got flexibility as to what those identifiers are
20:03 pdurbin
I'm going to favor primary keys such as the id from the authenticated user table or the dataset table.
20:06 pameyer
I'm going to favor your opinion of what makes sense for dataverse identifiers
20:10 pdurbin
hmm, come to think of it, you'll probably need the "identifier" of the user. I guess you can think of this as the dv username
20:16 pameyer
which identifier would we use if we wanted to use the dv api to send a notification to that user?
20:17 pdurbin
well, I need to code up that endpoint first :)
20:17 pdurbin
so we can do whatever
20:17 pameyer
so we've got flexibility ;)
20:18 pdurbin
yep
20:19 pdurbin
pameyer: the notification you're sending includes an rsync script, from what I understand.
20:20 pameyer
right
20:21 pameyer
we can wrap that in json if helpful (which it probably will be)
20:22 pdurbin
right now the rsync script is just emailed, right?
20:22 pameyer
right
20:22 pdurbin
but you want Dataverse to email it
20:23 pameyer
I'm reasonably sure we're switching to dataverse sending the user a notification, and the userdownloading it
20:23 pdurbin
yeah, Dataverse needs the script
20:23 pameyer
yup
20:23 pdurbin
so it makes sense for Dataverse to email the user once the script is in Dataverse's hot little hands
20:24 pameyer
I'd been thinking that Dataverse would host the script for download through the browser
20:24 pdurbin
yep. that too
20:26 pameyer
partly because a script could be something else from the dcm for a different protocol, and partly because some email clients have creative interpretations of what to do with character encodings
20:26 pdurbin
:)
20:35 pameyer
pdurbin: how much ugliness are you willing to tollerate?
20:36 pdurbin
for the summer of code, plenty :)
20:37 pameyer
that's what I was asking - I'm doing some de-coupling of existing code for this module
20:37 pameyer
we can worry about the scar tissue later
20:37 pdurbin
heh
20:53 sivoais joined #dataverse
20:53 sivoais joined #dataverse
