jasig-ssp IRC Logs-2013-10-04

Owned by Apereo Infrastructure
Last updated: Oct 04, 2013
6 min read

[13:01:00 CDT(-0500)] <JasonElwood> My firefox browser keeps locking up when I access Journal Details or Steps on Linux CI. It has done it three times now. Pinwheel on the Mac kinda thing. Anybody else see this?

[13:01:51 CDT(-0500)] <dmccallum54> any user/s in particular?

[13:02:08 CDT(-0500)] <JasonElwood> sorry. this is in admin

[13:02:16 CDT(-0500)] <JasonElwood> I'm trying to test the sorting and active/inactive

[13:02:40 CDT(-0500)] <JasonElwood> I'm not experiencing this in Chrome but no pinwheel. the browser just freezes.

[13:02:49 CDT(-0500)] <JasonElwood> Sadly, hoping it/s not just me

[13:05:03 CDT(-0500)] <dmccallum54> got the beachball on details

[13:05:04 CDT(-0500)] <dmccallum54> FF

[13:06:00 CDT(-0500)] <JasonElwood> good I guess

[13:06:23 CDT(-0500)] <dmccallum54> i'm getting zillions of "Journal Step Detail 1"

[13:06:32 CDT(-0500)] <dmccallum54> maybe a data thing? maybe an infinite loop?

[13:07:28 CDT(-0500)] <JasonElwood> it's definitely new

[13:07:45 CDT(-0500)] <dmccallum54> open a new ticket?

[13:07:47 CDT(-0500)] <JasonElwood> Archna did check in some code for active staus management

[13:07:49 CDT(-0500)] <JasonElwood> I will

[13:08:34 CDT(-0500)] <JasonElwood> And tony's journal associations are there too

[13:08:41 CDT(-0500)] <JasonElwood> both checked in yesterday

[13:12:53 CDT(-0500)] <JasonElwood> Anybody volunteering for it?

[13:13:36 CDT(-0500)] * dmccallum54 ducks

[13:15:47 CDT(-0500)] <JasonElwood> Sorry Dan

[13:17:46 CDT(-0500)] <dmccallum54> dammit

[13:18:33 CDT(-0500)] <dmccallum54> i guess that's the karmic retribution for IGEN magically fixing itself

[13:19:01 CDT(-0500)] <JasonElwood> It's probably more in tony's court, but I know he's focused on CL.

[13:19:09 CDT(-0500)] <dmccallum54> that's fine

[13:28:55 CDT(-0500)] <JasonElwood> unless you guys need something, I'm going to bail

[13:29:03 CDT(-0500)] <dmccallum54> later

[13:29:14 CDT(-0500)] <JasonElwood> DM- if you can let me know when IGEN is stable, I'll complete the configuration

[13:29:23 CDT(-0500)] <dmccallum54> will do

[13:29:28 CDT(-0500)] <JasonElwood> thanks.

[13:29:30 CDT(-0500)] <dmccallum54> vicky's still looking at it

[13:29:36 CDT(-0500)] <dmccallum54> somehow magically fixed itself overnight

[13:29:40 CDT(-0500)] <dmccallum54> which is no good, really

[13:29:55 CDT(-0500)] <JasonElwood> agreed

[13:57:15 CDT(-0500)] <Patty_> does anyone know

[13:57:27 CDT(-0500)] <Patty_> how many colors can be assigned to a specific course in map?

[17:06:08 CDT(-0500)] <dmccallum541> i need to switch to unminified js on Linux CI. this cramp anyone's style?

[17:06:58 CDT(-0500)] <TonyUnicon> no

[17:07:48 CDT(-0500)] <dmccallum541> thx. switched

[17:18:30 CDT(-0500)] <TonyUnicon> is there an issue?

[17:18:55 CDT(-0500)] <dmccallum541> 1795

[17:19:47 CDT(-0500)] <dmccallum541> the underlying API call for the details UI, for example is returning 100K results

[17:19:58 CDT(-0500)] <dmccallum541> which qualifies as "too many"

[17:20:05 CDT(-0500)] <dmccallum541> http://ssp-ci/ssp/api/1/reference/journalStep/aba1440c-ab5b-11e1-ba73-0026b9e7ff4c/journalStepDetail?limit=-1&amp;status=ACTIVE&amp;_dc=1380924579363

[17:20:16 CDT(-0500)] <dmccallum541> that'll crash JSONView, for example

[17:22:18 CDT(-0500)] <dmccallum541> there's a buttload of associations for that step in the db (1557), but not that many

[17:23:49 CDT(-0500)] <dmccallum541> heh.

[17:23:56 CDT(-0500)] <dmccallum541> wonder which step we tend to test with...

[17:24:09 CDT(-0500)] <dmccallum541> ssp=# select journal_step_id, count(star) cnt from journal_step_journal_step_detail group by journal_step_id order by cnt desc;

[17:24:09 CDT(-0500)] <dmccallum541> journal_step_id | cnt

[17:24:09 CDT(-0500)] <dmccallum541> -------------------------------------+-----

[17:24:09 CDT(-0500)] <dmccallum541> aba1440c-ab5b-11e1-ba73-0026b9e7ff4c | 1557

[17:24:09 CDT(-0500)] <dmccallum541> 0a90940a-401b-121b-8140-1bc26d460007 | 8

[17:24:09 CDT(-0500)] <dmccallum541> 0a90940a-401b-121b-8140-1bc248970006 | 6

[17:24:10 CDT(-0500)] <dmccallum541> 0a90940a-4016-1d02-8140-17250f5a0010 | 3

[17:24:10 CDT(-0500)] <dmccallum541> 0a90940a-3ff4-11f8-813f-f417db7b0002 | 3

[17:24:11 CDT(-0500)] <dmccallum541> 0a90940a-3ff0-11f0-813f-f3be88270004 | 3

[17:24:11 CDT(-0500)] <dmccallum541> 0a90940a-4016-1d02-8140-172534b20011 | 3

[17:24:12 CDT(-0500)] <dmccallum541> 0a90940a-412a-163e-8141-2cd5ad610000 | 3

[17:24:12 CDT(-0500)] <dmccallum541> 0a90940a-412a-163e-8141-2cd5e2e60001 | 2

[17:24:13 CDT(-0500)] <dmccallum541> 0a90940a-403e-11bd-8140-402ee0ae0004 | 2

[17:24:13 CDT(-0500)] <dmccallum541> 0a90940a-3efe-1520-813f-00a4d277003b | 2

[17:24:14 CDT(-0500)] <dmccallum541> 0a90940a-3efe-1520-813f-00a4e98f003c | 2

[17:24:14 CDT(-0500)] <dmccallum541> 0a90940a-3efe-1520-813f-00a4ac20003a | 2

[17:24:32 CDT(-0500)] <TonyUnicon> nice

[17:24:38 CDT(-0500)] <dmccallum541> uhoh. freenode just warned me i'm being throttled

[17:25:30 CDT(-0500)] <TonyUnicon> we need to nude that association entity

[17:25:33 CDT(-0500)] <TonyUnicon> nuke*

[17:26:02 CDT(-0500)] <dmccallum541> your first option sounded better

[17:44:35 CDT(-0500)] <TonyUnicon> not sure what the exact cause of it is

[17:44:39 CDT(-0500)] <TonyUnicon> but

[17:44:49 CDT(-0500)] <TonyUnicon> if you're wondering why that screwy logic is in there

[17:45:04 CDT(-0500)] <TonyUnicon> it stems from this mismatch between client and serverside models

[17:45:13 CDT(-0500)] <TonyUnicon> and the 'referencing inactive' issue

[17:52:02 CDT(-0500)] <dmccallum541> well

[17:52:18 CDT(-0500)] <dmccallum541> at the moment i'm just trying to figure out why 1557 db records turns into a 100,000-record JSON payload

[17:52:38 CDT(-0500)] <dmccallum541> i think the 1557 might be demo data insanity

[18:04:13 CDT(-0500)] <dmccallum541> this stuff is insane

[18:04:21 CDT(-0500)] <dmccallum541> (not on you, Tony)

[18:04:51 CDT(-0500)] <TonyUnicon> well partially on me

[18:05:04 CDT(-0500)] <TonyUnicon> but its what happens when you have to build on top of poop

[18:05:14 CDT(-0500)] <dmccallum541> heh

[18:14:39 CDT(-0500)] <TonyUnicon> ok

[18:14:56 CDT(-0500)] <TonyUnicon> I think I know the problem is

[18:15:39 CDT(-0500)] <TonyUnicon> can you give me till monday standup to fix it?

[18:15:48 CDT(-0500)] <dmccallum541> hey. no objection here

[18:16:15 CDT(-0500)] <dmccallum541> not enjoying getting my head back into this little ghetto

[18:16:26 CDT(-0500)] <dmccallum541> and of course my debugger is now refusing to do anything.

[18:16:55 CDT(-0500)] <TonyUnicon> this is my fault in the end, it was hard to notice with my small dataset

[18:17:08 CDT(-0500)] <dmccallum541> i'm opening another ticket to fix the demo data set

[18:17:17 CDT(-0500)] <dmccallum541> it was good in a way

[18:17:27 CDT(-0500)] <dmccallum541> but it's ridiculous for a "demo"

[18:18:44 CDT(-0500)] <TonyUnicon> yeah

[18:18:59 CDT(-0500)] <TonyUnicon> I think if you need 1557 steps to resolve an issue

[18:19:05 CDT(-0500)] <TonyUnicon> it maybe beyond the scope of the app

[18:19:14 CDT(-0500)] <dmccallum541> they're all effectively duplicate associations

[18:19:23 CDT(-0500)] <dmccallum541> they all point to exactly the same detail

[18:20:28 CDT(-0500)] <TonyUnicon> but I think as long as this data model stays as is, there will be a tax on development cost

[18:20:35 CDT(-0500)] <dmccallum541> want to just wipe them out and fix up all the journal_entry_detail records to point at the One True Association

[18:20:36 CDT(-0500)] <TonyUnicon> it made this ordering stuff much more difficult

[18:20:55 CDT(-0500)] <TonyUnicon> probably wouldn't hurt to clean that data up but

[18:20:59 CDT(-0500)] <TonyUnicon> i did notice a mismatch locally

[18:21:03 CDT(-0500)] <TonyUnicon> between the result of that query

[18:21:09 CDT(-0500)] <TonyUnicon> and my json payloads

[18:21:15 CDT(-0500)] <dmccallum541> seems to me it would simpler if this all just routed through the Hib domain model. rather than all the DAOs just for the associations themselves

[18:21:30 CDT(-0500)] <dmccallum541> and of course some client-side model cleanup as well

[18:22:19 CDT(-0500)] <dmccallum541> but almost seems like there needs to be a reboot here, period

[18:23:44 CDT(-0500)] <TonyUnicon> it would be nice to use HQL on that api call

[18:23:52 CDT(-0500)] <TonyUnicon> but then I'd have to write the paging stuff for it

[18:24:18 CDT(-0500)] <TonyUnicon> not that the client cares about paging in these tree models

[18:25:08 CDT(-0500)] <dmccallum541> exactly

[18:26:12 CDT(-0500)] <dmccallum541> for what we actually need it should just be return dao.getStepById(uuid).getDetails();

[18:26:42 CDT(-0500)] <dmccallum541> but i'm sure there's more to it that i'm not thinking of

[18:28:36 CDT(-0500)] <TonyUnicon> its more then that because of object status considerations

[18:28:55 CDT(-0500)] <TonyUnicon> but that old logic did not respect the object status of the association

[18:29:06 CDT(-0500)] <TonyUnicon> only the detail being referenced

[18:29:37 CDT(-0500)] <TonyUnicon> i got a little lazy in that code thinking the data sets were rather small

[18:29:54 CDT(-0500)] <TonyUnicon> but ill replace it with some cleaner code

[18:30:05 CDT(-0500)] <TonyUnicon> how terribly dissapointed would you be if we did not respect the paging?

[18:30:40 CDT(-0500)] <dmccallum541> i can totally live without the paging

[18:30:49 CDT(-0500)] <TonyUnicon> ok

[18:31:08 CDT(-0500)] <dmccallum541> we have no use case where we actually need it. it only exists to prevent user abuse in this case

[18:31:19 CDT(-0500)] <TonyUnicon> i should have it fixed by monday's standup

[18:31:26 CDT(-0500)] <dmccallum541> well, and consistency w/ other APIs. but we've run roughshod over that in other places already

[18:32:57 CDT(-0500)] <TonyUnicon> the beers will go down easy tonight I htink

[18:33:05 CDT(-0500)] <dmccallum541> tell me about it

[18:39:47 CDT(-0500)] <dmccallum541> hmmm… i bet you're right that the SortingAndPaging support was part of the root motivation(s) for a lot of the way this was originally laid out…

[18:40:09 CDT(-0500)] <dmccallum541> and they can't eager load any multi-valued associations with their current implementation...

[18:40:34 CDT(-0500)] <dmccallum541> so navigating the object model for any of this was going to be a different sort of disaster...

[18:40:57 CDT(-0500)] <TonyUnicon> yep