Time
S
Nick
Message
00:11 axfelix joined #dataverse
00:17 cnk left #dataverse
00:36 metamattj joined #dataverse
01:13 cnk joined #dataverse
02:01 axfelix joined #dataverse
03:21 garnett joined #dataverse
03:49 garnett joined #dataverse
03:50 agarnett joined #dataverse
04:37 agarnett joined #dataverse
05:45 metamattj joined #dataverse
07:24 bencomp joined #dataverse
10:06 somebody joined #dataverse
11:04 pdurbin
bencomp: there are some tests in https://github.com/IQSS/dataverse/tree/master/tests that use selenium
11:05 pdurbin
but I think the ones at https://github.com/IQSS/geoconnect-tester/tree/master/weld_test are much newer
12:08 bencomp
pdurbin: are they still run (by Jenkins?)?
12:08 pdurbin
they've never been run by Jenkins
12:08 bencomp
btw I was mostly impressed by the ease of recording a test case.
12:09 pdurbin
Open Sauce would be a better place to run them than Jenkins anyway: Open Sauce: free, unlimited, hosted Selenium testing for Open Source projects … - https://plus.google.com/+PhilipDurbin/posts/CpowbDMDRYH
12:13 bencomp
and are the tests run from Open Sauce? :)
12:13 pdurbin
yeah, it's like how travis does builds in the cloud
12:15 bencomp
can I see the results?
12:15 pdurbin
from open sauce? of course! they even record video of all your tests. it's incredible
12:16 bencomp
can you post a link then? :^)
12:17 pdurbin
I already linked to my writeup of what OpenSauce is.
12:18 bencomp
I don't see a link to the Dataverse Selenium testing results on Open Sauce... Did I misunderstand you?
12:19 pdurbin
bencomp: oh, you think Dataverse is using OpenSauce. No, we aren't. We probably should! We had someone kicking on the tires on it but she left a while ago.
12:20 bencomp
pdurbin: that was what I meant by "and are the tests run from Open Sauce?" ;)
12:20 pdurbin
oh
12:21 bencomp
it's early, I know :P
12:21 pdurbin
sigh
12:21 bencomp
anyway, these results wouldn't add to the code coverage
12:21 pdurbin
I don't even know if the tests at https://github.com/IQSS/dataverse/tree/master/tests even work any more.
12:23 pdurbin
see, they used to be in a folder called "sauce_test": https://github.com/IQSS/dataverse/tree/0b830688b749be17a7a815dbe99372717170065a/sauce_test
12:23 pdurbin
anyway
12:23 pdurbin
I'd rather write unit tests and API tests first. *Then* write tests for the GUI .
12:23 pdurbin
personally
12:23 bencomp
+1
12:24 bencomp
I received a DVN3 question about a strange permission situation
12:25 bencomp
really hope that that's not a bug, because it's pretty much the core of the application to handle permissions correctly. I'd like to see tests for that.
12:25 bencomp
personally :)
12:27 pdurbin
tests in Dataverse 4 or DVN 3? the latter is no longer under active development
12:37 bencomp
tests in 4
12:37 bencomp
I'm investigating in 3
12:39 pdurbin
ok
12:39 pdurbin
bencomp: did you see my comments on Shibboleth: multiple email addresses are stored as a single one · Issue #2512 · IQSS/dataverse - https://github.com/IQSS/dataverse/issues/2512 ?
12:43 bencomp
pdurbin: yes, thanks
12:44 pdurbin
bencomp: hopefully you see that I'm treating it as a bug that bad data (an invalid email address) got into the database
12:44 pdurbin
so I want to fix that first
12:47 bencomp
fine with me :)
12:47 pdurbin
but then
12:48 pdurbin
the easiest fix is to simply sort the multiple email addresses and pick the first one, like we do for names
12:48 pdurbin
bencomp: are you ok with that for now?
12:50 bencomp
I think we may not see this in production in the coming years, but it's not pretty
12:50 bencomp
same as for names
12:50 pdurbin
may not see what?
12:52 bencomp
multiple email addresses
12:52 bencomp
or even multiple email addresses in random order
12:52 bencomp
s/or even/let alone/
12:53 pdurbin
I've never seen an Identity Provider assert multiple email addresses.
12:53 bencomp
but if emails may be provided in random order, users may prefer one over another
12:54 bencomp
I imagine users with multiple names may have stronger feelings about using their preferred name
12:55 pdurbin
well, I already have a solution for multiple names, even if it's not an ideal solution
12:55 pdurbin
I'm talking about multiple email addresses, which again, I've never seen.
12:55 pdurbin
"mail: Email address of record in the Harvard IdDB. Note that this could be an address not ending in harvard.edu." http://iam.harvard.edu/resources/saml-shibboleth-attributes
12:56 pdurbin
it kind of blows my mind that "mail" is Multi=Y at https://www.incommon.org/federation/attributesummary.html
12:59 pdurbin
from https://webauth.service.ohio-state.edu/~shibboleth/attributes.html it sounds like OSU usually sends a single value
13:00 pdurbin
anyway, it seems like Dataverse *does* need to be able to handle multiple email addresses being asserted by the Identity Provider. for now I'll probably just sort and pick the first one
13:01 bencomp
it would be good to have a generic solution for multivalue attributes
13:01 bencomp
gotta go now
13:02 pdurbin
yeah, it would be good
14:48 axfelix joined #dataverse
15:28 metamattj joined #dataverse
16:20 bencomp
pdurbin: who is Naomi? she used System.out.println! :O
16:25 bencomp
turns out GlassFish 4 provides no substitute logger name for println statements
16:26 pdurbin
https://github.com/thenaomiday
16:26 pdurbin
bencomp: "blame" from github should tell you such things
16:27 bencomp
I kinda knew that :$
16:27 bencomp
thanks
16:28 bencomp
clearly it's time to go home and relax :)
16:28 pdurbin
:)
19:52 axfelix joined #dataverse
20:55 axfelix joined #dataverse
21:22 metamattj joined #dataverse
22:47 axfelix joined #dataverse
22:52 metamattj joined #dataverse
22:52 agarnett joined #dataverse
