uPortal IRC Logs-2013-04-04

[11:17:04 CDT(-0500)] <EricDalquist> drewwills: some time when I get my head above water we should talk about the attachements portlet (smile)

[11:17:12 CDT(-0500)] <EricDalquist> like I'm not even sure where it lives code wise (tongue)

[11:22:58 CDT(-0500)] <drewwills> https://github.com/Jasig/SimpleContentPortlet/tree/master/src/main/java/org/jasig/portlet/attachment

[11:23:36 CDT(-0500)] <drewwills> if it's not deployed & available, announcements doesn't show that area of the UI

[11:23:42 CDT(-0500)] <EricDalquist> ah ok

[11:24:05 CDT(-0500)] <EricDalquist> is there a file upload size limit?

[11:24:14 CDT(-0500)] <drewwills> yes, definitely

[11:24:51 CDT(-0500)] <EricDalquist> and have you tested this on postgres?

[11:24:56 CDT(-0500)] <EricDalquist> it has very different LOB handling

[11:25:06 CDT(-0500)] <EricDalquist> we got bit hard by that in uPortal :/

[11:25:15 CDT(-0500)] <drewwills> i'm not sure... do you think that platform will have issues with large text?

[11:26:10 CDT(-0500)] <drewwills> i imagine the necessary treatment will be similar to portlet preferences

[11:26:24 CDT(-0500)] <EricDalquist> it isn't large text

[11:26:32 CDT(-0500)] <EricDalquist> it is the transactional requirements of LOBs

[11:26:35 CDT(-0500)] <EricDalquist> with postgres

[11:26:46 CDT(-0500)] <EricDalquist> postgres requires a TX when reading lobs

[11:27:04 CDT(-0500)] <EricDalquist> all other DBs see a big performance hit if you are always doing TXs on reads

[11:27:29 CDT(-0500)] <drewwills> sounds like the discussions that went down w/ portlet preferences

[11:27:36 CDT(-0500)] <EricDalquist> yup

[11:27:46 CDT(-0500)] <EricDalquist> we solved it with a bunch of custom transaction handling code in uportal

[11:27:50 CDT(-0500)] <EricDalquist> but it would be worth testing

[11:27:54 CDT(-0500)] <EricDalquist> before having to port all that over

[11:27:59 CDT(-0500)] <drewwills> yeah

[11:29:51 CDT(-0500)] <drewwills> i expect we'll have to refine this one over time... i'm just glad to begin the plugging of this big hole

[11:29:57 CDT(-0500)] <EricDalquist> yeah

[11:30:13 CDT(-0500)] <EricDalquist> just want to make sure it gets checked before we get stuck and have to figure out a schema change

[11:39:55 CDT(-0500)] <drewwills> a schema change would stink... i hope it doesn't come to that... perhaps we'll have a near-term chance to set it all up for a postgress-using client... in that case, i'm certain it would get hit pretty hard