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

Using timezone: Eastern Standard Time
* kaarefc joins02:23
* cbeer joins02:52
* awoods joins
* mifistochka joins
* pivotal-bot____ joins
* fcrepo-bot joins02:57
* benpennell leaves03:59
* benpennell joins04:02
* kaarefc leaves04:24
* fcrepo-bot leaves04:44
* fcrepo-bot joins04:46
* fcrepo-bot leaves05:53
* mifistochka leaves06:12
* nbanks joins07:40
<pivotal-bot____>Andrew Woods added comment: "Resolved with FedoraAuthorizationDelegate.FEDORA_SERVLET_REQUEST" https://www.pivotaltracker.com/story/show/6637579407:45
Andrew Woods accepted "Request attributes like IP address should be accessible to implementations of FedoraPolicyEnforcementPoint" https://www.pivotaltracker.com/story/show/66375794
* tecoripa joins07:47
* tecoripa leaves07:48
<pivotal-bot____>Andrew Woods added "Remove dead code in AbstractRolesPEP" https://www.pivotaltracker.com/story/show/6710023208:15
* dwilcox joins08:58
<pivotal-bot____>Benjamin Pennell added comment: "@awoods Frank is following up on some comments on the PR" https://www.pivotaltracker.com/story/show/6651423609:00
Benjamin Pennell rejected "Add SPARQL actions to bench-tool" https://www.pivotaltracker.com/story/show/6651423609:02
<benpennell>rejected is such a strong term09:03
<awoods>benpennell: A pivotal "rejection" is like missing the very first bus to school, not like a middle school heartbreak.09:19
<benpennell>i'll just have to hope i didn't break frank's heart09:23
if a PR can't be automatically merged because its not up to date with master should the reviewer do that or the submitter?09:35
<awoods>benpennell: the submitter should rebase...09:36
because they are the ones that know how the code should end up.
<benpennell>that's what i figured, thanks09:37
* nbanks leaves09:48
* nbanks joins09:49
* ajs6f joins10:03
* ajs6f leaves10:11
* ajs6f joins10:15
<awoods>ajs6f: Do you know of any way in github to diff the 13 commits you pushed to "ontology" https://github.com/futures/ontology/commits/master to review https://www.pivotaltracker.com/story/show/65507092 ?10:18
<pivotal-bot____>feature: Develop a formal, documented model for the mapping from JCR through a Fedora 4 ontology to LDP (finished) / owner: A. "Torcrexilly" Soroka
<ajs6f>awoods: I am the _last_ person to ask. My reaction to Git generally is like a five-year old who has discovered an unlocked gun safe. Is there no flag on 'git diff' to compare two commits in the same tree?10:20
<awoods>ajs6f: git diff is no problem, I was looking to see it in github.
<daines>awoods: Start from here? https://github.com/futures/ontology/compare10:21
<ajs6f>Nine out of ten Github questions on Stackoverflow seem to begin with "There's an undocumented API call for that." and end with "They got rid of that call."
<awoods>ajs6f: in the absence of a review, maybe you can answer the question, "are the predicate URLs that we send in F4 responses still all found in the updated ontology"? Meaning, are our links still accurate?10:22
ajs6f: Ideally, we would publish the updated ontology
<ajs6f>Our links are not to Github— until we republish the changes (and we still owe each other a conv about that) nothing should change at all.10:23
<awoods>ajs6f: but of course I am not going to publish it if the links would break.
* tecoripa joins
* tecoripa leaves
<ajs6f>awoods: I don't think you should publish it at all until at least one other person knows what you are doing.
Or rather, how you are doing it.
I broke the data properties out into a separate vocabulary.
awoods: I think we may want to bring them down out of OWL into something simpler, like SKOS.10:24
<awoods>ajs6f: so it sounds like there are two tickets: update responses to new ontology, and have someone besides awoods publish the ontology.
* ajs6f1 joins10:25
* ajs6f leaves
<ajs6f1>awoods: The second ticket: yes. I didn't change the name or meaning of any of the predicates. I introduced new restrictions on some of them, and we may need to check to see that we aren't violating them, but I don't think we are.
<awoods>ajs6f1: that was my main question. So good, I am glad nothing should have broken (if we were to publish the ontology)10:26
<ajs6f1>awoods: But the publishing will have to change— we have at least two files where we had one. We could also think about publishing the LDP mapping.10:27
<awoods>ajs6f1: hmm... if we have two files where we had one, how could the predicate links still be good?10:28
<ajs6f1>The links aren't two either file. The links are to resources on the Web. We can merge the files or do something more complicated.10:29
aren't _to_ either file.
<awoods>ajs6f1: actually the links are served by specific files currently.10:34
which are directly mapped to files in "ontology" git repository. We can do pre-processing hocus pocus to merge files, but it seems like we just added complexity where it did not exist before.
<ajs6f1>I don't know why we are publishing anything out of our source code repo. That seems wrong. And we are adding complexity to the publishing workflow because maintaining simple data properties in OWL is unnecessary complexity in the ontology itself. If, on the other hand, we do end up mapping some of those properties into the PREMIS ontology, we could stay with OWL.10:37
<awoods>ajs6f1: it seems like maintaining and publishing our ontologies from git is reasonable. Do you have something else in mind?10:39
<ajs6f1>I have no problem with maintaining them in Git. I think we should be publishing them as separate artifacts, not links into our source. That could be as simple as making (modified) copies of the data onto a webserver, or as complicated as using a hosted service to the purpose, like Stanford's WebProtege.10:40
<pivotal-bot____>Mike Daines finished "Refactor authz to provide more information to PEPs and make API clearer for implementers" https://www.pivotaltracker.com/story/show/6694612210:43
<ajs6f1>Actually, as escowles and I discussed in Texas, the ideal might be to enable the repository itself to publish the ontologies it is using.
So that links out of an out-of-the-box ontology don't rely on external resources (like Github).10:44
<awoods>ajs6f1: fyi, the three ontologies are hosted on a duraspace webserver (http://fedora.info)10:46
<pivotal-bot____>Mike Daines started "Refactor authz to provide more information to PEPs and make API clearer for implementers" https://www.pivotaltracker.com/story/show/6694612210:48
Andrew Woods added comment: "This ticket includes the 13 commits starting from: ""10:58
https://github.com/futures/ontology/commit/45107b09c32b6..." https://www.pivotaltracker.com/story/show/65507092
Mike Daines finished "Refactor authz to provide more information to PEPs and make API clearer for implementers" https://www.pivotaltracker.com/story/show/6694612211:06
<ajs6f1>awoods: Okay, so why wouldn't we just publish four? We can just alter the namespace of the data properties to indicate the difference (which would probably be a good thing anyway).11:09
* ajs6f1 leaves
<awoods>ajs6f1: That sounds like a great idea... but it will be accompanied by updates to the predicate names in our responses, no?11:10
* ajs6f joins11:11
* ksclarke joins11:27
<daines>Fixes for my authorization refactoring PR are ready to be reviewed when anyone has a chance. https://github.com/futures/fcrepo4/pull/263 https://www.pivotaltracker.com/story/show/6694612211:43
<pivotal-bot____>feature: Refactor authz to provide more information to PEPs and make API clearer for implementers (finished) / owner: Mike Daines
Gregory Jansen edited "Demonstrate that we can reliably deploy, start and stop a cluster." https://www.pivotaltracker.com/story/show/6550515411:46
Gregory Jansen edited "Demonstrate that we can reliably deploy, start and stop a cluster at UNC." https://www.pivotaltracker.com/story/show/65505154
<awoods>daines: I will give it a look in an hour or so.
<daines>awoods: great, thanks!
<pivotal-bot____>Andrew Woods added comment: "@gregoryjansen, It is safe to say that transactions are not currently supported within a clustered context...." https://www.pivotaltracker.com/story/show/6550515411:48
* ajs6f leaves11:50
<pivotal-bot____>Gregory Jansen added "Demonstrate linear scaling of a cluster by adding nodes at UNC." https://www.pivotaltracker.com/story/show/67115528
Gregory Jansen added comment: "@awoods I was able to run benchtool in TX mode COMMIT against the first cluster node, but not against a l..." https://www.pivotaltracker.com/story/show/6550515411:53
Gregory Jansen added comment: "@awoods I would say that load balancers are not yet fully supported, since not all of our endpoints work ..." https://www.pivotaltracker.com/story/show/6550515411:54
* ajs6f joins11:55
<pivotal-bot____>Benjamin Pennell added "Benchtool option to change logging level" https://www.pivotaltracker.com/story/show/6711615411:59
* gregjansen joins12:00
finally here, freenode login not working normally for me12:01
awoods: I would like to move the current clustering page under our Administrator section and replace it with a higher-level overview page. Does that work for you?12:02
<awoods>gregjansen: hit it
<gregjansen>awoods: thx12:03
<benpennell>i'm thinking of adding a pivotal ticket to add an reporting format option for benchtool, mainly so that you could tell it to just give you the condensed, one-liner report to make it easier to use the data from multiple runs in spreadsheets or other reports.12:08
do others that have used benchtool think that'd be helpful? i noticed you were listing that output in your cluster benchmarking ticket, gregjansen12:09
<awoods>benpennell: what is currently required to get that same info from a benchtool run?
<benpennell>comment out a bunch of lines in FCRepoBenchRunner.java and rebuild12:10
<gregjansen>benpennell: I think the condensed results is helpful. It would be great to have some field separators or output a little json or something
<benpennell>it already outputs the condensed version, it is just surrounded by a lot of other text which makes it a bit difficult to extract if you run 10 runs in a row
<gregjansen>without field separators it is hard to plug into a spreadsheet consistently12:11
i.e. there may be nulls in some fields depending upon the test
<benpennell>i tried tab delimiting it but it came out as spaces in my terminal. json would probably be good as another format though
<awoods>benpennell: If you add a ticket, please be very descriptive. As the UNC crew is not going to be around for some time.
<benpennell>right now it is space delimited
<gregjansen>whatever works, commas work12:12
<awoods>benpennell: or... you may be interested in a weekend project?
<gregjansen>spaces are really hard to see
commas would be easier to import into a spreadsheet than json I suppose12:13
<benpennell>i'll make a ticket for it, it should be pretty simple to implement12:14
awoods: you never know!
<awoods>benpennell: I am open to surprises
<pivotal-bot____>Benjamin Pennell added "Benchtool option to change the reporting format" https://www.pivotaltracker.com/story/show/6711878212:29
<ksclarke>there is a discrepancy in the checkstyle and fcrepo eclipse line split settings (and documentation): checkstyle says 120 and eclipse/new dev docs says 80... it should be 120 now?12:35
https://github.com/futures/fcrepo-build-tools/blob/master/src/main/resources/fcrepo-checkstyle/checkstyle.xml vs. https://github.com/futures/fcrepo4/blob/master/src/site/eclipse/fedora-formatter.xml and https://github.com/futures/fcrepo4/tree/master/src/site/eclipse
<awoods>ksclarke: yes, the project standard is now 120.13:16
* github-ff joins13:51
[fcrepo4] ajs6f pushed 1 new commit to StrategiesForEventMapping: http://git.io/NCG_Xg
fcrepo4/StrategiesForEventMapping ec0f0c0 ajs6f: All tests passing with new strategies
* github-ff leaves
* github-ff joins13:52
[fcrepo4] ajs6f opened pull request #264: Strategies for event mapping (master...StrategiesForEventMapping) http://git.io/sI8NbA
* github-ff leaves
<pivotal-bot____>A. "Torcrexilly" Soroka added comment: "https://github.com/futures/fcrepo4/pull/264" https://www.pivotaltracker.com/story/show/66603072
A. "Torcrexilly" Soroka finished "Improve event reporting facilities for node property changes" https://www.pivotaltracker.com/story/show/66603072
A. "Torcrexilly" Soroka edited "Improve event reporting facilities for node property changes" https://www.pivotaltracker.com/story/show/6660307213:53
* ksclarke leaves14:02
* visteria joins14:04
* travis-ci joins14:07
[travis-ci] futures/fcrepo4#1616 (StrategiesForEventMapping - ec0f0c0 : ajs6f): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/1af3129f7fa2...ec0f0c0e2496
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/20305043
* travis-ci leaves
* ksclarke joins14:28
* ksclarke1 joins14:42
* edInCo joins14:46
<pivotal-bot____>Gregory Jansen added comment: "Still trying to replicate this failure. The test that fails is in fcrepo-kernel. Since the failure is in ..." https://www.pivotaltracker.com/story/show/6645409614:48
* ksclarke1 leaves14:53
* ajs6f1 joins14:56
* ajs6f leaves
* ajs6f1 leaves15:01
* ajs6f joins
<pivotal-bot____>Andrew Woods added comment: "Attached is the fcrepo-kernel maven build output." https://www.pivotaltracker.com/story/show/6645409615:33
Gregory Jansen added comment: "https://wiki.duraspace.org/display/FF/University+of+North+Carolina+at+Chapel+Hill+-+High+Volume+of+Concur..." https://www.pivotaltracker.com/story/show/6550482415:40
Gregory Jansen finished "Establish performance goals for UNC cluster" https://www.pivotaltracker.com/story/show/65504824
<gregjansen>awoods: thank you15:43
* ajs6f leaves
<awoods>gregjansen: hope it helps.
<pivotal-bot____>Gregory Jansen added comment: "@awoods Think I found it. Not only were we not shutting modeshape down, we also were not ensuring that it..." https://www.pivotaltracker.com/story/show/6645409616:04
* visteria leaves16:08
<gregjansen>awoods: I am going to keep cranking on the startup/shutdown issue tomorrow. I think I have fixed it, but it is definitely a timing issue.16:20
<awoods>gregjansen: on a call16:21
<pivotal-bot____>Gregory Jansen added comment: "I push a change to the PR that should address the timing issue by ensuring start in our factory's post co..." https://www.pivotaltracker.com/story/show/6645409616:23
* gregjansen leaves16:24
* tecoripa joins16:49
* tecoripa leaves
* dwilcox leaves
* dwilcox joins16:50
* dwilcox leaves16:54
* nbanks leaves17:07
* nbanks joins17:34
* nbanks leaves17:39
* edInCo leaves17:46
* github-ff joins18:17
[fcrepo4] awoods pushed 1 new commit to master: http://git.io/WFCDzA
fcrepo4/master 3bf9967 Mike Daines: Authorization refactoring...
* github-ff leaves
<pivotal-bot____>Andrew Woods added comment: "Resolved with: https://github.com/futures/fcrepo4/commit/3bf996787709247b32e5942fd0befef61f72a91c" https://www.pivotaltracker.com/story/show/66946122
* ksclarke leaves
<pivotal-bot____>Andrew Woods delivered "Refactor authz to provide more information to PEPs and make API clearer for implementers" https://www.pivotaltracker.com/story/show/66946122
* travis-ci joins18:33
[travis-ci] futures/fcrepo4#1620 (master - 3bf9967 : Mike Daines): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/63fe103a9cdd...3bf996787709
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/20321999
* travis-ci leaves
* dwilcox joins18:51
* dwilcox leaves18:55
<pivotal-bot____>Andrew Woods accepted "Refactor authz to provide more information to PEPs and make API clearer for implementers" https://www.pivotaltracker.com/story/show/6694612219:29
* github-ff joins19:31
[fcrepo4] awoods closed pull request #263: Authorization refactoring (master...authz-pep-refactor-remove-filter) http://git.io/C4WhCw
* github-ff leaves
* sfurmanova joins19:34
<pivotal-bot____>Andrew Woods accepted "Establish performance goals for UNC cluster" https://www.pivotaltracker.com/story/show/6550482419:40
* scossu joins20:00
* scossu leaves
* nbanks joins
* nbanks leaves20:01
* github-ff joins22:27
[fcrepo4] awoods pushed 1 new commit to master: http://git.io/X5Al1w
fcrepo4/master 649a3dd ajs6f: Improve event reporting facilities for node property changes...
* github-ff leaves
<pivotal-bot____>Andrew Woods added comment: "Resolved with: https://github.com/futures/fcrepo4/commit/649a3dd5970522d4a5d9d78254670aca97fff262" https://www.pivotaltracker.com/story/show/66603072
Andrew Woods delivered "Improve event reporting facilities for node property changes" https://www.pivotaltracker.com/story/show/66603072
* github-ff joins22:28
[fcrepo4] awoods deleted StrategiesForEventMapping at ec0f0c0: http://git.io/-m8VyA
* github-ff leaves
<pivotal-bot____>Andrew Woods delivered "Develop a formal, documented model for the mapping from JCR through a Fedora 4 ontology to LDP" https://www.pivotaltracker.com/story/show/6550709222:40
* travis-ci joins
[travis-ci] futures/fcrepo4#1621 (master - 649a3dd : ajs6f): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/3bf996787709...649a3dd59705
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/20330863
* travis-ci leaves
<pivotal-bot____>Andrew Woods added "Ensure response predicates conform to ontology" https://www.pivotaltracker.com/story/show/6715099622:42
Andrew Woods accepted "Develop a formal, documented model for the mapping from JCR through a Fedora 4 ontology to LDP" https://www.pivotaltracker.com/story/show/6550709222:43
Andrew Woods accepted "Improve event reporting facilities for node property changes" https://www.pivotaltracker.com/story/show/66603072
* scossu joins23:06
* scossu leaves
* fcrepo-bot joins23:33

Generated by Sualtam