Log of the #fcrepo channel on chat.freenode.net

Using timezone: Eastern Standard Time
* benpennell joins07:58
* dwilcox joins08:09
* dwilcox leaves08:24
* peichman joins08:53
* peichman1 joins08:58
* peichman leaves08:59
* dwilcox joins
* dwilcox leaves09:01
* yamil_ joins09:23
* whikloj joins10:44
* bseeger joins10:54
https://wiki.duraspace.org/display/FF/2017-11-16+-+Fedora+Tech+Meeting in 6 minutes...
<benpennell>i'm on the call11:03
<bseeger>google thread: https://groups.google.com/forum/#!topic/fedora-community/Ok9L9oosPsE11:33
whikloj: was that Carrick who was doing the testing?11:43
<whikloj>peichman1: Colin Brittle
<benpennell>thanks bethany!11:56
this isn't related to any ongoing tickets, but are etags supposed to change when a generated membership property is added to a container?11:57
<whikloj>benpennell: I remember a discussion about this...there might be a JIRA ticket11:58
<bseeger>there is this one: https://jira.duraspace.org/browse/FCREPO-174211:59
<benpennell>and re: whether to do a beta for fcrepo5, not totally relevant, but hyrax has recently been working on their release workflow for that as well and have a document that might be worth looking at for how they handle moving from beta to rc https://wiki.duraspace.org/display/samvera/Hyrax+Testing+and+Release+Process12:00
bseeger whikloj: thanks for finding that, i'm not clear on if inbound reference means "membership references child" or "container adds member"12:02
<bseeger>benpennell — thanks for sharing that document - very interesting to look at12:04
<whikloj>benpennell: basically it means the reference is stored on another resource, and points to the resource you are looking at
<benpennell>so inbound reference to child would be: <container> <pcdm:hasMember> <child>?12:06
the situation i'm trying to work out how to deal with is the container adds a pcdm:hasMember relation from an IndirectContainer, but the container's etag isn't updated so caching mechanisms don't know the properties have changed12:08
<whikloj>Oh indirect containers, they seem to be the most problematic
<benpennell>yeah, pretty much, although i've found it kind of hard not to use them12:17
<whikloj>benpennell: Earlier in CLAW land I ran into this https://jira.duraspace.org/browse/FCREPO-209212:19
<benpennell>whikloj: hadn't seen how to do an inverted indirect container, that's good to know. is CLAW still using indirect containers?12:25
<whikloj>benpennell: no
<benpennell>whikloj: what approach instead?
<whikloj>benpennell: We are going to maintain the references manually (or rather via a PUT)12:26
<benpennell>whikloj: okay, so they will be actual properties on the objects rather than container magic
<whikloj>benpennell: exactly12:27
<benpennell>whikloj: bit more management, but more reliable i guess
<whikloj>benpennell: the key was making changes idempotent, so dhlamb has just done a tonne of work to make all our changes to Fedora as only PUTs and DELETEs12:28
<benpennell>whikloj: i'll need to consider switching to something like that, i keep running into reasons to not rely on indirectContainers12:31
<whikloj>benpennell: yeah, I really liked the idea of auto-creating those links. But it becomes hidden and harder to manage12:32
* whikloj leaves12:54
* whikloj joins
* whikloj leaves12:58
* whikloj joins13:46
* whikloj leaves13:58
* bseeger1 joins14:01
* bryjbrown joins14:02
* bseeger1 leaves
* bseeger leaves14:04
* whikloj joins14:06
* benpennell leaves14:30
* bseeger joins15:01
* bryjbrown_ joins15:09
* bryjbrown leaves15:11
* bryjbrown_ leaves16:00
* bseeger leaves16:05
* dwilcox joins16:07
* benpennell joins16:23
* dwilcox leaves16:52
* peichman1 leaves17:07
* whikloj leaves17:25
* yamil_ leaves17:48
* benpennell leaves20:20
* benpennell joins20:39
* benpennell leaves21:27
* benpennell joins21:35
* benpennell leaves21:56
* benpennell joins22:21
* benpennell leaves22:52