Log of the #duraspace-ff channel on chat.freenode.net

Using timezone: Eastern Standard Time
* fcrepo-bot joins02:47
* fcrepo-bot leaves03:50
* jcoyne joins07:35
* jcoyne leaves08:34
* jcoyne joins08:56
* jbaiter joins08:58
* jbaiter leaves
* jbaiter joins
* ermadmix joins09:17
* fasseg joins
* dwilcox joins09:18
* dwilcox leaves09:24
* dwilcox joins
* dwilcox_ joins09:27
* dwilcox leaves09:29
* fcrepo-bot joins09:32
* aawoods_ joins10:03
* aawoods_ leaves
* aawoods_ joins10:04
* awoods leaves
* scossu joins10:09
* gregjansen joins10:14
<scossu>awoods: What time is the content model validation call?10:20
<awoods>scossu: 4pm ET
<scossu>awoods: thanks.
scossu: and you are on the agenda for the committers call: https://wiki.duraspace.org/display/FF/2014-03-20+-+Fedora+Committer+Meeting10:21
* dwilcox_ leaves10:22
* dwilcox joins10:23
* dwilcox leaves10:27
* ksclarke joins10:33
* dwilcox joins10:35
* fcrepo-bot leaves10:40
<pivotal-bot_____>Andrew Woods accepted "Uploading to federated filesystem via REST API breaks browse (but not content)" https://www.pivotaltracker.com/story/show/6766365810:52
Andrew Woods accepted "Copy between Federation and Repo" https://www.pivotaltracker.com/story/show/65600152
* edInCo joins10:56
* ksclarke leaves10:57
* mikeAtUVa joins11:00
* barmintor joins11:01
ok, ok, I'm on my way
awoods: are we on readytlk?11:02
<awoods>barmintor: yes
* ksclarke joins
* ajs6f joins11:05
<barmintor>good, good, I was worried about that uncomitted hour of the afternoon :)11:06
<gregjansen>cannot hear ya11:13
* fasseg leaves
<gregjansen>good idea
<ajs6f>Let's don't reverse engineer our own platform.
<barmintor>ajs6f: I think I just terrorized B Daigle for an hour. It's not just you!11:14
<gregjansen>seems like nt:file or nt:folder are the only node types possible in a filesystem connector...11:15
<ajs6f>barmintor: Careful. Bradley is a black belt in karate. He's not someone you want to push too far. Don't let his pleasant and urbane demeanor fool you!
* barmintor was fooled11:16
* martind_ joins11:17
<gregjansen>I think it is more about the node types than their names??
<ajs6f>gregjansen: org.modeshape.jcr.federation.FederatedDocumentWriter.setPrimaryType(Name) implies that nodes from a connector can have arbitrary primary type...
gregjansen: perhaps this is occurring somewhere in the FIlesystemConnector or subclasses?11:18
<gregjansen>ajs6f: sounds right11:21
<ajs6f>gregjansen: Then we _should_ be able to do something about that to buy ourselves some flexibility...11:22
I don't like that either
we could solve it by storing all the ACLs in a separate tree, then linking to them from the fcr:resource11:26
i.e. only add the property to the federated content (which would be in the sidecar or in mode)
note that persisting ACLs as part of the content was a design goal w/respect to preservation at one point..11:39
so if we put them in "system space" we may want to consider the impact
<escowles>gregjansen: maybe a node can link to a ACL either as a child node, or somewhere else? that would would allow nodes in the repo to have their ACLs bundled with them, and federated filesystem nodes to link to ACLs in the repo11:41
<ajs6f>escowles/gregjansen: What about "sharable nodes" (which I think MDOE supports)?11:43
<gregjansen>escowles: I like that pattern
escowles: we were just talking locally about some kind of linking that could be a special way to connect ACLs to "projected nodes"11:44
<ajs6f>http://www.day.com/specs/jcr/2.0/3_Repository_Model.html#3.9 Shareable Nodes Model
<gregjansen>"shareable" sounds like the reuse use case, whereas some kind of hard ref (cascade delete) would work for the non-reuse case11:45
I think there is a cascade delete type of reference in JCR
<escowles>ajs6f: shareable nodes could be a good way for nodes in the repo to share the same ACL and also have them bundled under each node too
<ajs6f>escowles: Yep.
<escowles>gregjansen: i assume the cascade delete would delete the link to the shared node, not the original node
<ajs6f>As long as it all scales in some way.11:46
<escowles>sounds like a good IT to write...
<gregjansen>escowles: I think you can do either in JCR
<ajs6f>Cal you guys hear me?11:48
<escowles>ajs6f: nope
<ajs6f>&^(^&(6 ReadyTalk.11:49
<escowles>ajs6f: you're just Mr. Technical Difficulties this week
<ajs6f>A sequencer can execute arbitrary code, but it is instantiated by MODE, not us.
So it can't get access to the internal mesage bus or anything that we inject with our DI framework.
Or at least not easily at all.
People expecting to emit messages from a sequence should be aware of that.11:50
all: Got to run. Talk to y'all later. If you can hear me then.11:54
* ajs6f leaves
<escowles>my presentation: https://www.icloud.com/iw/#keynote/BALjEpLgqhdUVnjSeVWBO2OFFbx1AUb3IQ2F/ff11:56
<scossu>Minutes are online: https://wiki.duraspace.org/display/FF/2014-03-20+-+Fedora+Committer+Meeting - let me know if I missed / misinterpreted something.12:33
<awoods>thanks, scossu.
<scossu>escowles, I will add the presentation link to the minutes.
<awoods>ermadmix: It looks like the error you were seeing with the filesystem connector is just a log warning, not an actual error.13:00
re: namespaces13:02
* ermadmix leaves13:08
* ermadmix joins13:20
awoods: re 13:00, yes I realized that too, looking into how the mode file system connector and AccessRoles handles the ID using just the local part that the filesystem connector appears to persist in this case...13:59
<ksclarke>awoods: looks like I'm not going to be able to make the 4pm meeting after all so that drops the number back down to 10 if you wanted to use the hangout14:06
<awoods>on a call
<pivotal-bot_____>Esme Cowles added comment: "I've created a MODE issue for this: https://issues.jboss.org/browse/MODE-2176" https://www.pivotaltracker.com/story/show/6730811014:37
* scossu leaves14:43
* ajs6f joins14:45
<pivotal-bot_____>Esme Cowles added comment: "Also, one partial workaround for this issue: if you change the repo path the filesystem is mapped to, then t..." https://www.pivotaltracker.com/story/show/6730811014:49
<ajs6f>awoods: I'll be three-four minutes late to sprint wrapup. But I will be there. Whether or not you will be able to hear or whether or not you will be interested in anything I have to say remains to be seen.14:52
* scossu joins14:57
<ajs6f>Can you hear me?15:06
awoods: You already know I'm out.15:15
<pivotal-bot_____>Eric James added comment: "Documentation of this issue: ""15:39
https://wiki.duraspace.org/display/FF/Federation+and+Authentication+ongoing+issues" https://www.pivotaltracker.com/story/show/67557532
Andrew Woods edited "federated configuration changes in repository.json aren't reflected in the repository" https://www.pivotaltracker.com/story/show/6730811015:45
<ajs6f>Sprint wind-down folks: I'm off for a bio-break, back for the next meeting.15:55
<pivotal-bot_____>Andrew Woods edited "changing an existing externalSource name fails Context initialization" https://www.pivotaltracker.com/story/show/6748690815:56
Esme Cowles started "changing an existing externalSource name fails Context initialization" https://www.pivotaltracker.com/story/show/67486908
<awoods>ajs6f, barmintor: ?? coming16:05
<ajs6f>awoods: I can't believe you said "Jeepers!" during the last meeting. That's awesome. I didn't think anyone really said that outside of the Andy Griffith Show.16:06
I can't believe Ive memorized our ReadyTalk code.
I guess they probably said that in the Dobey Gillis Show too.
<barmintor>whoa, Martin!16:07
<ajs6f>martind_ has been all over F4.
Like brown on rice.16:08
* nbanks joins16:13
<awoods>Content modeling use cases: https://wiki.duraspace.org/label/FF/uc-content-modeling16:14
<awoods>ajs6f: also mikeAtUVa and edInCo16:39
<ajs6f>awoods: Thanks!
<barmintor>ajs6f: I do like the notion of having an object representing a certain validation in your repo, and asking the repo for an assertion of current applicability of same to another object/node16:42
<ajs6f>barmintor: Didn't we talk about that yesterday!?!
<barmintor>ajs6f: yes, but not here.
<ajs6f>barmintor: True.16:43
barmintor: It's all about the _hook_.16:44
<barmintor>HOISTS AND CRANES http://www.allaboutthehook.com/16:46
<ajs6f>I can't believe that's a thing.
<martind_>"Entity layer"?16:48
<barmintor>nbanks: how does your interest in restrictive types gibe with your interest in mixin types?16:51
<edInCo>ajs6f: you do remember there is a website for everything http://www.websiteforeverything.com/16:52
<nbanks>barmintor: I'd like to be abled to start with a locked down object then add mixin's to add new functionality / properties etc.16:53
<ajs6f>edInCo: We better get that archived. What if it goes down? We could lose EVERYTHING!
<nbanks>barmintor: but at the same point we will have use cases for completely open objects, but that could be implemented as a mixin type as well?16:54
<barmintor>nbanks: that's… hard.
<jcoyne>nbanks: boo!
<nbanks>jcoyne: Why boo?
<cbeer>awoods: do you have a location in mind for the next hackfest?16:56
<awoods>on a call
<jcoyne>nbanks: because I prefer the exact opposite. If someone thinks it's important, then let them store it. Don't make them configure the system first.
<ajs6f>jcoyne: But nbanks deals with a lot of people who _turn out to be expensively wrong_.16:57
Speaking as someone who is usually wrong.
cbeer: Wish you were here.
<nbanks>jcoyne: We could potentially have both could we not?
<jcoyne>"expensively wrong" <- I like that.
<ajs6f>_I'm_ expensively wrong.
Expensively and expansively16:58
<barmintor>In FCR3, I've gravitated increasingly toward very loose models, and REL-INT/EXT describing structure.
<ajs6f>barmintor: Why wold nbanks' desire be hard? Why couldn't there be a "any property or sub-resource" you want mixin?
<jcoyne>RELS-EXT is too damn strict!
<nbanks>barmintor: There wasn't much of a choice in FCR3 otherwise?16:59
<barmintor>ajs6f: b/c you have one mixin that says "not X", and another one that says "yes X"
<ajs6f>barmintor: I thought mixin's couldn't say "not X"?
<barmintor>ajs6f: they can't; this is why I'm saying the concept is at odds with proscriptive types17:00
* dwilcox leaves17:01
<jcoyne>This is going to come down between the Java programmers and the Python/Ruby people?
* dwilcox joins
<barmintor>jcoyne: that'd be the first time the Fedora 3 devs fell on the dynamic side of the equation17:02
<ajs6f>barmintor: Seriously? did you never notice the painfully dynamic nature of dissemination in F3?17:03
<barmintor>awoods: I need to drop17:04
* mikeAtUVa leaves
<barmintor>ajs6f: dude. SOAP? Dude.
* dwilcox leaves17:05
* barmintor leaves17:06
<ajs6f>awoods: You took notes. Right? Right?!? RIGHT!??!17:08
* nbanks leaves
<ajs6f>Cause I didn't.
<awoods>ajs6f: yes, I took notes.17:11
cbeer: I was thinking east coast, but no specific city yet. Do you have a suggestion?
* ermadmix leaves
* scossu leaves
<ajs6f>awoods: Thanks. I'm out for the day. See y'all.17:12
<awoods>ajs6f: be safe
<ajs6f>awoods: Be brave.
* ajs6f leaves17:13
<cbeer>awoods: just selfish ones. i already have some travel lined up the following week that i'd like to tack onto either this or OR, just for my sanity.17:14
awoods: so, selfishly, a delta hub or somewhere in the midwest would be better for me.17:15
* scossu joins17:19
<jcoyne>cbeer: MSP is both
<cbeer>jcoyne: yep. MSP would be my #1 pick, but i'm not sure how committed awoods is to the east coast17:20
<jcoyne>cbeer: lots of things to see here: http://mnbeer.com/breweries/
* gregjansen leaves17:21
<awoods>cbeer: fyi, I am looking at the week of 5/1217:23
* scossu leaves
<cbeer>awoods: oh. tcramer asked me about the week after that.
awoods, oh, no. that's right.17:24
both this and OR are inconveniently scheduled, but i don't want to do 3 cross-country (and then some) in a month.17:25
* ksclarke leaves17:38
* nbanks joins17:55
* nbanks leaves18:39
* edInCo leaves19:04
* fcrepo-bot joins19:15
* fcrepo-bot leaves20:24
* jcoyne leaves21:29
* jcoyne joins21:44
* jbaiter leaves22:10
* fcrepo-bot joins23:21
* jcoyne leaves00:01
* fcrepo-bot leaves00:25

Generated by Sualtam