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

Using timezone: Eastern Standard Time
* ksclarke leaves00:39
* dwilcox joins07:53
* awead leaves08:13
* osmandin joins08:57
* ksclarke joins09:02
* acoburn joins09:17
<mikeAtUVa>ajs6f: that PR you're looking at is within another one that you can comment inline on... https://github.com/fcrepo4-labs/migration-utils/pull/109:20
<ajs6f>mikeAtUVa: thanks!
<mikeAtUVa>ajs6f: also, thanks for the comments, we appreciate the pointers!
<ajs6f>mikeAtUVa: This is Java. We don't call them pointers, we call them references.
* ksclarke leaves09:31
* awead joins
* ksclarke joins09:32
* Nianli joins09:50
ajs6f: thanks for the comments! I will change it based on your feedbacks.09:59
<ajs6f>Nianli: No prob. Hope they help.10:00
<Nianli>ajs6f: Oh, yes! I just need time to learn it.10:02
* github-ff joins10:21
[fcrepo4] ajs6f force-pushed BNodesAreBNodes from bb4129c to 86dca3a: http://git.io/p5iJ
fcrepo4/BNodesAreBNodes 64cd450 ajs6f: First draft, blocked on unintelligible auth ITs
fcrepo4/BNodesAreBNodes 4583b24 ajs6f: Down to failing tests for hash URIs
fcrepo4/BNodesAreBNodes d42b5b5 ajs6f: Working true blank-node implementation
* github-ff leaves
* travis-ci joins10:41
fcrepo4/fcrepo4#3495 (BNodesAreBNodes - 86dca3a : ajs6f): The build passed.
Change view : https://github.com/fcrepo4/fcrepo4/compare/bb4129cb8e93...86dca3a4b2a7
Build details : http://travis-ci.org/fcrepo4/fcrepo4/builds/54875607
* travis-ci leaves
<f4jenkins>Yippee, build fixed!10:52
Project fcrepo4-T2 build #178: FIXED in 4 min 52 sec: http://jenkins.fcrepo.org/job/fcrepo4-T2/178/
awoods: Remove equals() for a private utility class with no fields
* whikloj joins11:04
* dwilcox leaves11:14
* dwilcox joins11:29
* Nianli leaves12:06
* osmandin leaves12:13
* dwilcox leaves12:24
* dwilcox joins12:28
* awead leaves
* awead joins13:17
* osmandin joins13:43
* Nianli joins13:57
* awead leaves15:07
<whikloj>mikeAtUVa: If I made a new class like your ConsoleLogger (for migration-utils), can I get the whole foxml document? I just want to make a test class to push the foxml at a camel-service.15:11
<mikeAtUVa>whikloj: well, here's the thing, sometimes the underlying foxml is a complete representation of the object, other times it isn't.15:13
<whikloj>duh duh duNNN
sooo you are pulling it together in the app?15:14
<mikeAtUVa>whikloj: if your model is to throw a bunch of foxml onto camel, you don't need the framework build in the migration-utils app, you can just iterate over a directory of foxml files.
whikloj: I am pulling it together...
* github-ff joins
[fcrepo4] ajs6f force-pushed Java8 from 7aae527 to 2002521: http://git.io/F98X
fcrepo4/Java8 013012e ajs6f: Step to Java 8
fcrepo4/Java8 1686301 Aaron Coburn: updated travis-ci configuration
fcrepo4/Java8 308f6db ajs6f: Removing obsolete PermGem flag
* github-ff leaves
<mikeAtUVa>whickloj, there's just no serialization format... just an in-memory object.
whickloj: specifically an ObjectReference implementation is in memory, that exposes access to the whole object... https://github.com/fcrepo4-labs/migration-utils/blob/master/src/main/java/org/fcrepo/migration/ObjectReference.java15:15
<whikloj>mikeAtUVa: ok, but that is pulling in an external object correct? Like I am migrating A, and A B C, this grabs me C.15:16
<mikeAtUVa>whikloj: I'm not sure I understand what you mean by "external object".15:17
<whikloj>mikeAtUVa: sorry, when your class refers to "references". Are these metadata references or references to the datastream content?15:18
* awead joins15:21
<mikeAtUVa>whikloj: sorry for the confused terminology, there's a lot of reuse in words like "object" "reference" etc... basically from within the application, when you have a java object that is an implementation of ObjectReference class, that class exposes access to every bit of information about its subject fedora 3 object. (you can get datastream content, you can get properties, you can get versions, etc.)15:22
<whikloj>mikeAtUVa: ok, so this is supposed to use archive format foxml (if you pass it a directory of files) correct?15:23
<mikeAtUVa>whikloj: there are two types of foxml it's been tested to handle right now...15:24
<whikloj>archive and migrate
<mikeAtUVa>whikloj: 1, foxml that was exported with the "archive" context.
whikloj: and foxml that exists on disk in the native store.
whikloj: the later is appealing because it's already there for fedora 3... you don't have to export, you don't have to make additional copies... heck, if no one's updating your fedora 3 instance you don't even have to stop the service.15:26
<whikloj>mikeAtUVa: ok I think I'm starting to get it. The archives foxml has everything in it, but this can be problematic. The native store is already there (no need to export) but you need to "pull it together"
<mikeAtUVa>whikloj: indeed... and the act of "pulling it together" involves having walked through the datastream store directory to create a mapping from ids to file names.15:27
whikloj: the current version of the code you're looking at really only creates a general abstraction for a fedora3 object that works in either of those contexts to hide the complexity of the serialization.
whickloj: I'd be interested to know how you envision camel fitting into the process... ie, what goes over the wire and where the transformation occurs.15:28
<whikloj>mikeAtUVa: ok ok, so the camel service danny started on looks to use the whole of a foxml document. However, ('cause I'm new to camel) I'm imaging we could use your abstraction to send parts of the object to different camel endpoints.
<mikeAtUVa>I'd worry about sending the whole foxml, for a couple reasons... one, when many versions are involved or large binaries exist, it can be huge...15:29
two: the existing serializations are poor... generating the archive foxml involves base 64 encoding into an XML doc...15:31
<whikloj>mikeAtUVa: true, but in theory we could leave the datastream content where it is if we could get a URI, then pass that to the camel component which would request it.
mikeAtUVa: ok that I can't speak to
miekAtUVa: ok, so I am going to try and write a ConsoleLogger type class that pushes an object to a camel endpoint. Just to see what happens.15:32
mikeAtUVa: thanks for the info, I'm sure I'll be back for more later. Meeting time. cheers.
<mikeAtUVa>whikloj: sounds good... here's one last tip...15:33
<ruebot>whikloj: fwiw, the upgration-text-fixtures repo just has migrate context foxml.15:34
<mikeAtUVa>whikloj: the consoleLogger class uses the streaming API, which doesn't expose the whole object... look at implementing FedoraObjectHandler and using the ObjectAbstractionStreamingFedoraObjectHandler to have whole-object access.15:35
<ruebot>whikloj: i'm waiting for a whole bunch of stuff to get merged so we can have some nice archive context foxml
* dwilcox leaves15:42
* osmandin leaves15:50
* ajs6f leaves15:56
* ajs6f joins15:57
<ruebot>awoods: dhlamb won't be there tomorrow. he is on vacation. shall I put the star on my name?15:59
* ajs6f leaves
* ajs6f joins16:00
<awoods>ruebot: that would be very nice of you. thanks.
<ajs6f> awoods: quick Confluence qestion: do you if there's a way to make a link to a Jira issue, like a plain link with your choice of text, as opposed to the little icon the macro makes?
do you kno16:01
<awoods>ajs6f: you can always just use a "web link"16:02
<ajs6f>awoods: Fair enough. That would certainly work.
<ruebot>awoods: and whikloj is going to join us tomorrow for our pre-call... or whatever we call that :-)
<awoods>ruebot: glad to hear it.16:03
* github-ff joins
[fcrepo4] ajs6f pushed 2 new commits to BNodesAreBNodes: http://git.io/h85a
fcrepo4/BNodesAreBNodes 2abc9d1 ajs6f: Better JavaDocs
fcrepo4/BNodesAreBNodes 279c421 ajs6f: Better use of constants
* github-ff leaves
<ajs6f>awoods: That ^^^ should do it for the bnodes thing, I think.
<awoods>ajs6f: I am still poking around...16:04
<ajs6f>awoods: Did you have any qualms about the functionality?
<awoods>ajs6f: I am still in the "trying to understand what you have put in place" phase.
<ajs6f>awoods: I'd ideally like to see this merged today so I can do it all over again for Java8 tomorrow.
awood: Fair enough.16:05
<barmintor>awoods, ajs6f https://docs.google.com/document/d/1n7XpX3cN01t2sr3jXnqAioB5x-_SdbSwf71TiXEs6VY/edit#16:20
<awoods>barmintor: thanks. How would you like to move the related conversation forward?16:21
<barmintor>awoods: I don’t know.16:22
<ajs6f>barmintor: Interesting that what you're calling the LDP approach would work with OPTION, too.
barmintor: I'm not sure why you're distinguishing between the two approaches as "LDP" and "Fedora". What's specific to Fedora about the second approach?16:23
barmintor: Woudn't the full example include GETing the fcr:metadata and following the fixity link there?16:24
<barmintor>ajs6f: LDP/Fedora describes an aesthetic to API development: “there’s a REST endpoint” vs “there’s a Link, and LDP interaction models"16:25
<ajs6f>barmintor: Wouldn't there be a middle ground of "There's a REST endpoint and it's advertised via a Link"?16:26
barmintor: I'm not any more in favor of "magically well-known URIs" than are you. That's not really REST.
<barmintor>ajs6f: You could introduce a hop but reduce some customization by linking through the metadata
<ajs6f>barmintor: I thought that's what we are currently doing...?
* travis-ci joins16:27
fcrepo4/fcrepo4#3497 (BNodesAreBNodes - 279c421 : ajs6f): The build passed.
Change view : https://github.com/fcrepo4/fcrepo4/compare/86dca3a4b2a7...279c421faae6
Build details : http://travis-ci.org/fcrepo4/fcrepo4/builds/54926546
* travis-ci leaves
<barmintor>ajs6f: the other part of it the Link is “that leads to a resource that has LDP interaction models”
<ajs6f>barmintor: Oh, I'm in no way defending the current response you get from a fcr:fxitiy (or any other similar endpoint). I'm sure they do need to be worked over to be more LDPy.16:28
barmintor: If that's what you're saying, then +1.
Got to run for the day. Say hi to Fedo for me.16:30
* ajs6f leaves
<barmintor>ajs6f: I’m *also* saying that we should be able to describe all these REST endpoints in terms of LDP interactions. “fcr:fixity is a RDFResource describing a NonRDFResource, like the LDP-NR’s description” “fcr:versions is a Container of dated versions of a Resource"
durn it
awoods: I guess I’d want to propose like documents for the REST api bits that aren’t clearly understood as LDP interactions, and get feedback from RSanderson about how reasonably they might be understood to be extensions16:36
awoods: and move to defining the API in terms of the Links and references rather than the endpoints
awoods: in a way this is all related to the versionable API bit16:37
<awoods>barmintor: that sounds like a great approach for moving more towards an LDP-based model, and having it documented...
barmintor: too bad you will not be at ldcx
barmintor: this would make a good session16:38
<barmintor>I’m not the road warrior you are (thank goodness)
tech leads have to go to lots of places16:39
* mikeAtUVa leaves16:43
* awead leaves17:07
* whikloj leaves17:57
* acoburn leaves18:02
* ksclarke leaves18:41
* pivotal-bot_ leaves19:38
* pivotal-bot_ joins
* ksclarke joins20:54
* ksclarke leaves20:58
* ksclarke joins21:14
* awead joins22:16
* awead leaves22:30
* awead joins22:44
* awead leaves22:47
* awead joins22:59
* awead leaves23:32

Generated by Sualtam