jasig-ssp IRC Logs-2013-03-21

Owned by Apereo Infrastructure
Last updated: Mar 21, 2013
5 min read

[16:22:24 CDT(-0500)] <JasonElwood> The external data conversations are getting complicated over email. There are several factors involved now. How about a hangout?

[16:23:10 CDT(-0500)] <js70> sounds good.

[16:24:40 CDT(-0500)] <dmccallum54> and i see unicon mail servers are quarantining some ssp-dev messages. how helpful

[17:23:25 CDT(-0500)] <dmccallum54> opened a unicon support ticket for the mail issue… for all of those who are just dying to get the next batch of ssp-dev mail

[17:25:32 CDT(-0500)] <JasonElwood> JS and Dan- I noticed one other thing with the external_student_transcript table while updating the definitions. the gpa is the only nullable column in the table. that seems completely backwards to me. Reluctantly, I suggest that we also fix it with 1.2 because we are now displaying gpa

[17:26:07 CDT(-0500)] <js70> k

[17:26:20 CDT(-0500)] <JasonElwood> the description (which comes from the legacy spreadsheet) says that SSP will calculate the gpa if provided. that appears to be the original functionality for the system.

[17:27:13 CDT(-0500)] <js70> we are agreed thats not a good idea going forward?

[17:27:34 CDT(-0500)] <JasonElwood> it's the opposite of how we agreed to handle gpa's on the hangout

[17:27:34 CDT(-0500)] <dmccallum54> think of it as an enhancement by removal

[17:27:57 CDT(-0500)] <JasonElwood> hahaha

[17:28:06 CDT(-0500)] <js70> :^)

[17:28:11 CDT(-0500)] <dmccallum54> yes, you can no longer shoot yourself in the foot, but you'll thank us for it

[17:28:25 CDT(-0500)] <JasonElwood> I'll go ahead an create a ticket to remove an enhancement

[17:35:58 CDT(-0500)] <dmccallum54> that's the kind of project management every project would benefit from!

[17:49:37 CDT(-0500)] <JasonElwood> do we want to change the EA roster drop-down functionality in 2.0 then?

[17:49:49 CDT(-0500)] <JasonElwood> making the columns non-nullable, etc

[17:51:29 CDT(-0500)] <dmccallum54> https://issues.jasig.org/browse/SSP-733 was intended as a sort of umbrella for that

[17:52:01 CDT(-0500)] <dmccallum54> think there ended up just being some wholesale implementation of the 1.2 schema changes in one big shot tho

[17:53:41 CDT(-0500)] <JasonElwood> so does the UI enforcement, aka app failing to start, live in a point release of a major/dot release?

[17:55:16 CDT(-0500)] <dmccallum54> that's a tough one. the approach actually proposed in the email thread was a bit more gentle. i.e. once you thought you had it right, you'd throw a switch and the next restart would blow up if you hadn't gotten it right

[17:55:29 CDT(-0500)] <dmccallum54> this is not quite the same as what we discussed in the hangout

[17:55:40 CDT(-0500)] <dmccallum54> which was more "at some point in time you will just have no choice"

[17:56:15 CDT(-0500)] <dmccallum54> but i dont see why we couldn't plan to be more draconian in 2.0

[17:56:36 CDT(-0500)] <dmccallum54> and call it all off if people can't even get the optional stuff right in 1.2

[17:56:38 CDT(-0500)] <JasonElwood> I'd feel better about a major rather than dot

[17:56:42 CDT(-0500)] <dmccallum54> sure sure

[17:57:48 CDT(-0500)] <JasonElwood> JS- I missed highlighting two columns in the spreadsheet I sent you. The Jira issue 857 is accurate.

[17:58:25 CDT(-0500)] <js70> k, I'll make sure to validate between the two.

[17:58:47 CDT(-0500)] <js70> trust AND verify

[17:59:15 CDT(-0500)] <JasonElwood> definitely verify. I've been a little less than 100% accurate lately

[17:59:49 CDT(-0500)] <dmccallum54> 2 points to the person who figures out why unicon mail servers quarantined this thread: http://jasig.275507.n4.nabble.com/Question-About-Case-Management-report-tt4658782.html

[18:00:16 CDT(-0500)] <js70> CUM how fun

[18:00:23 CDT(-0500)] <dmccallum54> we have a winner

[18:00:46 CDT(-0500)] <js70> really changes the meaning of the thread!

[18:00:51 CDT(-0500)] <JasonElwood> if I had the same encouragement that Dan M has throughout the day, I'm sure I would be at least 100%

[18:01:32 CDT(-0500)] <js70> Exactly. A lucky man.

[18:01:54 CDT(-0500)] * dmccallum54 blush

[18:02:26 CDT(-0500)] <js70> So my question is, does this go in my quarterly report?:-$

[18:03:07 CDT(-0500)] <dmccallum54> mine or yours?

[18:03:30 CDT(-0500)] <dmccallum54> cuz it's all over that thread

[18:03:33 CDT(-0500)] <dmccallum54> so to speak

[18:03:38 CDT(-0500)] <dmccallum54> ok. that's probably enough of that

[18:03:52 CDT(-0500)] <dmccallum54> good thing this is being logged

[18:04:10 CDT(-0500)] <JasonElwood> It's time to scrub the logs

[18:04:12 CDT(-0500)] <dmccallum54> hate for the Internet (and HR depts everywhere) to miss out on this

[18:05:36 CDT(-0500)] <dmccallum54> JasonElwood have you seen any trouble with sqlserver since 856 went in

[18:06:44 CDT(-0500)] <JasonElwood> I don't believe so, but let me check the logs

[18:09:01 CDT(-0500)] <dmccallum54> 1.1 blew up on me b/c it thought the UDF function created by 000025.xml already existed. not sure if that was just my own carelessness during development or a genuine problem.

[18:10:34 CDT(-0500)] <JasonElwood> I don't see any errors in the log. I definitely didn't get any deployment errors. from a spot check of 5-6 JEs, they all look good.

[18:11:05 CDT(-0500)] <dmccallum54> thx

[18:12:18 CDT(-0500)] <JasonElwood> Has anybody noticed a slower response within your local SSP this week? It is noticeably slower since Monday.

[18:14:25 CDT(-0500)] <dmccallum54> during app load or everywhere?

[18:14:42 CDT(-0500)] <JasonElwood> pretty much everywhere

[18:15:07 CDT(-0500)] <JasonElwood> except Manage Users

[18:16:22 CDT(-0500)] <dmccallum54> a change went in on tuesday that keeps database connections open slightly longer

[18:16:56 CDT(-0500)] <dmccallum54> did CI slow down too?

[18:17:46 CDT(-0500)] <dmccallum54> not that it has all the 1.2 changes, of course

[18:17:51 CDT(-0500)] <JasonElwood> I really haven't used CI much because it's been flaky

[18:18:06 CDT(-0500)] <dmccallum54> heh

[18:18:27 CDT(-0500)] <JasonElwood> I haven't spent the time to figure it out pushing towards the release. It will take some time to get some useful external data in there

[18:51:20 CDT(-0500)] <js70> The SSP-688 error showcases that reports are not part of the ext.js app. just a jsp page that submits.

[18:52:18 CDT(-0500)] <JasonElwood> so is there any way to catch the timeout like the ext.js does?

[18:52:41 CDT(-0500)] <dmccallum54> yes

[18:53:42 CDT(-0500)] <JasonElwood> and reasonably easy and quick

[18:53:57 CDT(-0500)] <dmccallum54> you get a 403 in this case i think

[18:56:25 CDT(-0500)] <dmccallum54> would have to either… change all the forms to ajax submits or modify the error page that generates the 403 json to "do something else".

[18:56:41 CDT(-0500)] <dmccallum54> the latter is pretty squishy and isn't really going to be pleasant

[18:57:26 CDT(-0500)] <dmccallum54> for the user i mean

[18:59:44 CDT(-0500)] <dmccallum54> there are jquery plugins for the former (http://www.malsup.com/jquery/form/)

[19:00:14 CDT(-0500)] <dmccallum54> not sure how well that's going to work with our PDF responses

[19:01:45 CDT(-0500)] <dmccallum54> might have to monkey with the headers being set server side

[19:02:16 CDT(-0500)] <dmccallum54> so anyway… maybe an hour or so to see if it's even feasible

[19:03:32 CDT(-0500)] <dmccallum54> "it" being submitting forms via ajax, correctly offering to download resulting PDFs, handling error codes gracefully, and special handling for 403 to offer to redirect to login

[19:09:27 CDT(-0500)] <js70> we don' have a 403 jsp page, I'm pretty sure its bringing back the 500. not a big deal but we need to create the 403 as well.

[19:10:56 CDT(-0500)] <dmccallum54> think it might not be using any of the jsp pages

[19:11:06 CDT(-0500)] <dmccallum54> @PreAuthorize(Permission.PERMIT_ALL)

[19:11:06 CDT(-0500)] <dmccallum54> @ExceptionHandler(AccessDeniedException.class)

[19:11:06 CDT(-0500)] <dmccallum54> @ResponseStatus(HttpStatus.FORBIDDEN)

[19:11:06 CDT(-0500)] <dmccallum54> public @ResponseBody

[19:11:06 CDT(-0500)] <dmccallum54> ServiceResponse handleAccessDenied(final AccessDeniedException e) {

[19:11:06 CDT(-0500)] <dmccallum54> getLogger().error(ERROR_PREFIX, e);

[19:11:06 CDT(-0500)] <dmccallum54> return new ServiceResponse(false, e.getMessage());

[19:11:07 CDT(-0500)] <dmccallum54> }

[19:11:34 CDT(-0500)] <js70> k

[19:12:00 CDT(-0500)] <js70> makes sense

[19:12:48 CDT(-0500)] <dmccallum54> trying to use a page seems a little weird in this case, esp if we're going to try to pretty it up

[19:13:23 CDT(-0500)] <dmccallum54> b/c again you're hitting an API and it's barfing. it's not a web page that's barfing where it would make sense to send you to a nice pretty 'sorry about that' page

[19:13:31 CDT(-0500)] <js70> yeah, I like your jquery form suggestion. if it works should be fairly simple to implement.

[19:18:01 CDT(-0500)] <dmccallum54> isn't that always the story (smile)

[19:18:44 CDT(-0500)] <dmccallum54> alright. i gotta run. daycare pickup calls again.