Time
S
Nick
Message
08:27 jri joined #dataverse
08:28 jri joined #dataverse
09:09 jri joined #dataverse
13:15 jri joined #dataverse
13:21 donsizemore joined #dataverse
14:16 jri joined #dataverse
14:59 donsizemore joined #dataverse
15:09 pdurbin
mornin donsizemore
15:17 jri joined #dataverse
15:18 jri joined #dataverse
15:52 donsizemore
@pdurbin how are you sir?
15:53 pdurbin
Warming up. The school bus was late and my 11 year old was texting me but eventually it came. It was 11F with windchill this morning.
16:00 donsizemore
we determined something weird this morning. after going back and forth with our campus mail folks (hybrid exchange/office365 service) because e-mail verification tokens were getting stripped from e-mails sent to UNC accounts... the problem appears to somehow occur within dataverse
16:00 donsizemore
akio is looking into it more closely
16:02 pdurbin
donsizemore: ok. Is this new behavior? It used to work? I'm wondering if you're suffering from this issue: https://github.com/IQSS/dataverse/issues/3407
16:02 donsizemore
it works for everybody except UNC folks (on this end). gmail, you at harvard, all fine
16:03 donsizemore
and yes, we were all converted
16:03 pdurbin
ok
16:04 pdurbin
donsizemore: speaking of Akio do you know if he happended to see my comment at https://github.com/IQSS/dataverse/issues/2331#issuecomment-351180763 ?
16:09 donsizemore
@pdurbin akio says yes, he's carving out the ingest module to make it a stand-alone application
16:10 donsizemore
@pdurbin the issues so far surround preserving extracted metadata (i'm paraphrasing, he's in my office now telling me this)
16:11 pdurbin
donsizemore: ok. When he's done, if a bug is found in that code, does he plan to fix the bug in two places? It sounds like the code will exist in two places.
16:18 jri joined #dataverse
17:19 jri joined #dataverse
17:44 jri joined #dataverse
18:42 donsizemore joined #dataverse
18:56 pdurbin
donsizemore: you sounds busy with other things but at some point I'd like to check in with you about https://github.com/IQSS/dataverse/issues/4384 . I brought it up in standup this morning, which is why it got moved from "Code Review" to "Development".
18:56 pdurbin
sound*
18:57 donsizemore
@pdurbin i'm actually in a holding pattern until my 2:30 - what'd i screw up?
18:58 pdurbin
donsizemore: heh. Nothing. We just need to agree on "definition of done" before we can move it to QA.
18:59 donsizemore
@pdurbin oh, I was sure I've missed something. but most of the EL package names carry across majors
19:00 pdurbin
Sure. Makes sense. Did you want to hack on the Docker images at all?
19:01 donsizemore
absolutely, but i need to map out Pete's workflow across those scripts (and set up my own little Docker registry for testing)
19:02 pdurbin
Awesome. So we'll leave that issue and pull request in the Development column at https://waffle.io/IQSS/dataverse rather than "Code Review". There's still work do to. Docker, Installation Guide. Maybe other stuff.
19:03 pdurbin
Also, I'm not sure if Leonid is aware of this effort yet. I see you noticed his comment that we plan to use el7 on AWS.
19:04 pdurbin
donsizemore: do you mind if I add it to the "dev efforts by the community" spreadsheet at https://docs.google.com/spreadsheets/d/1pl9U0_CtWQ3oz6ZllvSHeyB0EG1M_vZEC_aZ7hREnhE/edit?usp=sharing ?
19:09 donsizemore
not at all! i just hope i can do the community effort justice
19:12 pdurbin
donsizemore: cool, will do. Any should I remove the "rewrite installer in Python" one? Are we still doing that?
19:28 donsizemore
@pdurbin it's a distant dream of mine but if someone would like to pick it up instead don't let me stop 'em
19:49 pdurbin
donsizemore: no worries. I change the status to "distant dream"
19:50 pdurbin
bjonnh: oh and I switched the issue for Docker support. It was 3938 (closed) but now it's 4040 (open).
19:51 pdurbin
donsizemore: and I was just thinking more about Docker stuff for your el7 branch. I'm concerned that you (or anyone) is going to struggle with how we currently advise to hack on Docker stuff since it's so tied to Minishift: http://guides.dataverse.org/en/4.8.4/developers/dev-environment.html#docker
19:58 bjonnh
pdurbin: great
20:01 pdurbin
bjonnh: 4040 is sort of about using Docker in production. You might want to look at this issue which is more about using Docker to run integration tests: https://github.com/IQSS/dataverse/issues/4369
20:04 bjonnh
yep
20:31 donsizemore joined #dataverse
20:31 donsizemore
@pdurbin quick question?
20:32 pdurbin
hit me
20:33 donsizemore
@pdurbin on the converted users not able to receive verification link issue. akio and i are looking for the source logic which checks for user type on login - we were browsing but he thought it might be faster just to ask you =)
20:33 pdurbin
ok
20:34 donsizemore
though it looks to me like it's all in LoginPage.java - and the pertinent columns between builtinuser and authenticateduser look identical. trying to suss out where that's breaking down, but akio keeps asking about precedence in user type
20:35 pdurbin
I just opened up LoginPage.java
20:37 pdurbin
donsizemore: you do know that the "Verify Email" stuff doesn't really do anything yet, right? All it does is record in the database if you clicked the link or not.
20:37 pdurbin
Clicking the link puts a timestamp in the "emailconfirmed" column of the "authenticateduser" table: http://phoenix.dataverse.org/schemaspy/latest/tables/authenticateduser.html
20:37 donsizemore
@pdurbin correct, but Thu-Mai is keen to use it once it becomes effective
20:38 donsizemore
@pdurbin also correct, but we have a lot of users =) Akio keeps suspecting some code remnant... he went back to his office
20:39 pdurbin
I see. You like it when these timestamps are recorded.
20:39 pdurbin
The link is missing when someone creates a new Shib account?
20:40 donsizemore
no, the issue you reminded me of earlier holds true for us - users who are "born shibbolized" are fine
20:40 donsizemore
i'm going to let akio pore over it for a while
20:41 pdurbin
Ok, I'm confused about what the user just did. Did they sign up? Did they click "verify email"? Something else?
20:42 donsizemore
@pdurbin we were looking into #3407 and he was asking me about the relationship between builtinuser and authenticateduser, and whether there might be any preference for a user type. consider any question retracted, as he's back in his office going through the source code
20:44 pdurbin
Every user appears in the authenticateduser table. Only builtin users appear in the builtinuser table. You'll also want to look at the authenticateduserlookup table which is where we store the "authenticationproviderid" which can be "shib" or "github" or "builtin" or whatever.
21:09 jri joined #dataverse
21:38 pdurbin
Have a good weekend, everyone.
21:38 pdurbin left #dataverse
22:39 jri joined #dataverse
