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

Using timezone: Eastern Standard Time
<ff-pivotal>Chris Beer added comment: "+1; what should happens when one checksum is swapped for another?" https://www.pivotaltracker.com/story/show/4570968300:27
Chris Beer added comment: "ok, I have something working. I'll push it up as a pull request , but I'd appreciate a review and learn why i..." https://www.pivotaltracker.com/story/show/4562856100:39
* jcoyne leaves
* github-ff joins00:40
[fcrepo4] cbeer created path-to-modeshape-config (+2 new commits): http://git.io/gzIWYA
fcrepo4/path-to-modeshape-config 46fa184 Chris Beer: allow the repository to be passed in as a property
fcrepo4/path-to-modeshape-config 9525e5d Chris Beer: use spring resources to pass in the classpath/filepath to the modeshape config
* github-ff leaves
<ff-pivotal>Chris Beer added comment: "https://github.com/futures/fcrepo4/pull/30" https://www.pivotaltracker.com/story/show/45628561
* github-ff joins
[fcrepo4] cbeer opened pull request #30: Path to modeshape config (master...path-to-modeshape-config) http://git.io/ziMLbA
* github-ff leaves
<ff-pivotal>Chris Beer edited "Update the kitchen-sink and/or webapp repositories to nicely expose the configs for implementors to modify" https://www.pivotaltracker.com/story/show/4562856100:41
Chris Beer edited "Update the kitchen-sink and/or webapp repositories to nicely expose the configs for implementors to modify" https://www.pivotaltracker.com/story/show/45628561
Chris Beer finished "Update the kitchen-sink and/or webapp repositories to nicely expose the configs for implementors to modify" https://www.pivotaltracker.com/story/show/45628561
Chris Beer added comment: "I created the digest and digestAlgorithm properties for this data.. but now I'm wondering if I should have cr..." https://www.pivotaltracker.com/story/show/4562814300:43
* kaarefc joins01:40
* kaarefc leaves02:10
* kaarefc joins02:38
<ff-pivotal>Edwin Shin edited "Store fixity data when a datastream is created or modified" https://www.pivotaltracker.com/story/show/4562814303:52
* github-ff joins04:00
[fcrepo4] eddies created 45628143 (+1 new commit): http://git.io/q25iUQ
fcrepo4/45628143 b9862ac Edwin Shin: 45628143 : Store fixity data when a datastream is created or modified...
* github-ff leaves
* github-ff joins04:01
[fcrepo4] eddies opened pull request #31: 45628143: Store fixity data when a datastream is created or modified (master...45628143) http://git.io/CJtq0Q
* github-ff leaves
<ff-pivotal>Edwin Shin added comment: "I can see arguments for either way. Also see: https://github.com/futures/fcrepo4/pull/31 which would really m..." https://www.pivotaltracker.com/story/show/4562814304:05
Edwin Shin added "Add ops to eddies on #duraspace-ff" https://www.pivotaltracker.com/story/show/4575150904:13
Edwin Shin started "CMIS endpoint NPE on kitchen sink" https://www.pivotaltracker.com/story/show/45666379
Edwin Shin added comment: "In an attempt to avoid NIH syndrome and practically, not to re-invent the wheel, some other references: ""04:22
- ht..." https://www.pivotaltracker.com/story/show/45457823
Edwin Shin edited "Expose object fixity state in Hydra" https://www.pivotaltracker.com/story/show/4569138504:34
* eddies leaves05:21
* eddies joins05:31
* eddies leaves
* eddies joins
<ff-pivotal>Edwin Shin added comment: "See this use case articulation from Richard Green: http://permalink.gmane.org/gmane.comp.cms.fedora-commons.u..." https://www.pivotaltracker.com/story/show/4568841105:40
Edwin Shin added "Switch from #duraspace-ff to #fcrepo for IRC" https://www.pivotaltracker.com/story/show/4575692706:11
Edwin Shin edited "Switch from #duraspace-ff to #fcrepo for IRC" https://www.pivotaltracker.com/story/show/45756927
Edwin Shin edited "Switch from #duraspace-ff to #fcrepo for IRC" https://www.pivotaltracker.com/story/show/4575692706:12
* eddies leaves06:27
* eddies joins06:43
* eddies leaves
* eddies joins
* eddies leaves07:01
* eddies joins07:07
* eddies leaves
* eddies joins
<ff-pivotal>Frank Asseg added "Versioning features form eSciDoc " https://www.pivotaltracker.com/story/show/4575952107:11
Frank Asseg edited "Versioning features for eSciDoc " https://www.pivotaltracker.com/story/show/45759521
Frank Asseg edited "Versioning features for eSciDoc " https://www.pivotaltracker.com/story/show/4575952107:18
Frank Asseg added comment: "Inquired about versioning requirements at FIZ Karlsruhe" https://www.pivotaltracker.com/story/show/45759521
Frank Asseg started "Add eSciDoc stories for Fedora versioning needs" https://www.pivotaltracker.com/story/show/4569270507:19
Frank Asseg edited "Versioning features for eSciDoc " https://www.pivotaltracker.com/story/show/45759521
Frank Asseg added comment: "Adde ticket: https://www.pivotaltracker.com/story/show/45759521" https://www.pivotaltracker.com/story/show/4569270507:20
* VincentNG joins07:54
* ppound joins07:57
* jonathangee joins08:14
* jonathangee leaves08:23
* jonathangee joins08:24
* jcoyne joins08:29
* jonathangee leaves08:40
* jonathangee joins
* jcoyne leaves08:47
* jcoyne joins08:50
* jcoyne leaves09:03
* kaarefc leaves09:16
<ff-pivotal>A. "Vombulo" Soroka added comment: "Immediately, nothing. IOW, adding a new checksum impl shouldn't change the state of the content at a..." https://www.pivotaltracker.com/story/show/4570968309:18
* ajs6f joins09:22
<ff-pivotal>Chris Beer added comment: "barmintor suggested URIs were one way to go, in order to include both the digest value and the algorithm. In ..." https://www.pivotaltracker.com/story/show/4562814309:48
<ajs6f>+1 to URIs. I'm writing a test for it now.09:49
<cbeer>eddies: you're suggesting today's call starts in 40 minutes?09:50
* kaarefc joins09:51
<eddies>actually, we said it was unkind to expect you to make that call
i'll take proposals for alternative slots09:52
* github-ff joins
[fcrepo4] ajs6f pushed 1 new commit to master: http://git.io/B2l6CQ
fcrepo4/master 0b7fafd ajs6f: Added simple test for ContentDigest helper class
* github-ff leaves
<ff-pivotal>Chris Beer added comment: "Close, but not really the right command. and +O is silly. Make CS op you as needed: /cs op #duraspace-ff" https://www.pivotaltracker.com/story/show/4575150909:58
Chris Beer started "Add ops to eddies on #duraspace-ff" https://www.pivotaltracker.com/story/show/45751509
Chris Beer finished "Add ops to eddies on #duraspace-ff" https://www.pivotaltracker.com/story/show/45751509
<ajs6f>cbeer: just think of this schedule as a way to have at least one meeting in the week where you don't have to get out of bed. :)09:59
<ff-pivotal>Chris Beer added comment: "But the wheel is obviously so easy to make!" https://www.pivotaltracker.com/story/show/45457823
<eddies>cbeer: i didn't know that
<cbeer>eddies: yup. it's a freenode-ism to not auto-op, afaict.10:00
anyway, i also gave you all the chanserv flags beyond +o
<ff-pivotal>A. "Gorbulak" Soroka added comment: "Doesn't Duracloud do something like this that we could use?" https://www.pivotaltracker.com/story/show/45457823
<eddies>cbeer: how to inquire about the ownership/status of #fcrepo?10:01
<ff-pivotal>Chris Beer added comment: "Can you split these into separate tickets?" https://www.pivotaltracker.com/story/show/45759521
<cbeer>j #fcrepo
<ff-pivotal>Edwin Shin added comment: "There's: https://svn.duraspace.org/view/duracloud/trunk/services/hadoop/fixity-processor/" https://www.pivotaltracker.com/story/show/4545782310:02
<cbeer>eddies: "freenode-staff"
<ff-pivotal>Frank Asseg edited "Version top level objects feature for eSciDoc " https://www.pivotaltracker.com/story/show/4575952110:03
<eddies>i saw that earlier and didn't grok if that was some default freenode channel business or someone on freenode staff really took #fcrepo
<ff-pivotal>Frank Asseg added "Retrieve a top level object version based on a time stamp" https://www.pivotaltracker.com/story/show/4576882310:04
A. "Gorbulak" Soroka added comment: "Mm. That looks bound to Hadoop. Although INSPN _does_ offer MapReduce jobs of some kind. Maybe the ..." https://www.pivotaltracker.com/story/show/45457823
<cbeer>i don't know, maybe it's some masking. or maybe the sign of a registered channel?
s/registered/registered with freenode/
<ff-pivotal>Frank Asseg added comment: "Sure...split to: https://www.pivotaltracker.com/story/show/45768823" https://www.pivotaltracker.com/story/show/45759521
<cbeer>s/freenode/PDPC/
<ff-pivotal>Frank Asseg added comment: "split from: https://www.pivotaltracker.com/story/show/45759521" https://www.pivotaltracker.com/story/show/4576882310:05
<cbeer>ajs6f: objections to merging in https://github.com/futures/fcrepo4/pull/30#issuecomment-14562545 for now, and filing a ticket to de-spring it later? (e.g. when we do CDI, if we do CDI)?
or, it sounds like you're suggesting some work in that ticket I don't know how to do it.10:06
s/it.$/yet/
anyone know if there's a ticket for poking the checksum info through the REST API?10:08
<ajs6f>No, no objection— it's a step forward. I'd rather see it unbound from Spring _and_ CDI, but let's move forward and keep improving it.
The improvement I would want is to handle the MODE JSON config purely as an URL, but using Spring Resource to pick up the classpath protocol is where we are now, so all this change does is improve by adding parameterization: +1.10:09
* github-ff joins
[fcrepo4] cbeer pushed 1 new commit to master: http://git.io/SM4ZgA
fcrepo4/master 3f523f6 Chris Beer: Merge pull request #30 from futures/path-to-modeshape-config...
* github-ff leaves
<ff-pivotal>Chris Beer added "De-springify ModeshapeFactoryBean" https://www.pivotaltracker.com/story/show/4576926110:10
<ajs6f>I don't see a ticket like that. You mean something like : /objects/{pid}/datastreams/{dsid}/fixity/{algorithm} ?
<cbeer>ajs6f: i was just thinking in the datastream profile10:11
i don't know if VincentNG's ticket overs that (although it doesn't say that, really)
<ajs6f>Oh, cool. That will be a minor break to back-compatibility (the profile XSDs from fcrepo3 are surprisingly strict) so maybe it should appear in fcrepo-http-api?
<cbeer>i thought there was digest info there already?10:12
<ajs6f>Hmm. Might be. Let me go check.
* barmintor joins
* kaarefc leaves10:14
<ajs6f>cbeer: Yep. It's in there.10:15
We can just use it.
<cbeer>whew.
<ajs6f><xs:element name="dsChecksumType" >
<xs:simpleType>
<xs:restriction base="xs:string">
<xs:enumeration value="DEFAULT"/>
<xs:enumeration value="DISABLED"/>
<xs:enumeration value="MD5"/>
<xs:enumeration value="SHA-1"/>
<xs:enumeration value="SHA-256"/>
<xs:enumeration value="SHA-385"/>
<xs:enumeration value="SHA-512"/>
<xs:enumeration value="HAVAL"/>
<xs:enumeration value="TIGER"/>
<xs:enumeration value="WHIRLPOOL"/>
</xs:restriction>
<ff-pivotal>Chris Beer added "Add digest and algorithm info to the datastream profile output" https://www.pivotaltracker.com/story/show/45769627
Chris Beer edited "Add digest and algorithm info to the datastream profile output" https://www.pivotaltracker.com/story/show/45769627
<ajs6f>WHIRLPOOL? Did we really support that?
<ff-pivotal>Chris Beer edited "Add digest and algorithm info to the datastream profile output" https://www.pivotaltracker.com/story/show/4576962710:16
Chris Beer estimated "Add digest and algorithm info to the datastream profile output" as 1 point https://www.pivotaltracker.com/story/show/45769627
A. "Gorbulak" Soroka edited "Add digest and algorithm info to the datastream profile output" https://www.pivotaltracker.com/story/show/4576962710:17
<cbeer>oh, that's where that ticket went
too fast.
<ajs6f>What ticket?
<cbeer>the one you just started.10:18
<ajs6f>It looks simple enough. We already have slots in the JAXB.
<cbeer>yep10:19
eddies: for this joint call, are we timeboxing it to 30m? or are we going to let it slop into the standup?10:20
<eddies>30mins10:21
<cbeer>ok. sad trombone is coming out after 30m :P
<barmintor>cbeer: do you have a second before the hangout to talk about sha1 and ds keys?
<cbeer>sure
<barmintor><-- just went to the hangout10:22
<cbeer>ah, TALK talk.10:23
<eddies>no. Talk Talk: Talk Talk. http://www.youtube.com/watch?v=Eqjttpl3peI10:24
<ff-pivotal>Frank Asseg edited "Version top level objects feature for eSciDoc " https://www.pivotaltracker.com/story/show/4575952110:30
* travis-ci joins
[travis-ci] futures/fcrepo4#172 (master - 3f523f6 : Chris Beer): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/0b7fafdd8868...3f523f6dadd4
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/5315421
* travis-ci leaves
* github-ff joins10:31
[fcrepo4] ajs6f pushed 2 new commits to master: http://git.io/_WR4LQ
fcrepo4/master 2ad89e2 ajs6f: Produced checksum info in datastream profile for: https://www.pivotaltracker.com/story/show/45769627
fcrepo4/master acdcf66 ajs6f: Merge branch 'master' of github.com:futures/fcrepo4
* github-ff leaves
* jcoyne joins
<ff-pivotal>A. "Gorbulak" Soroka added comment: "Finished with: https://github.com/futures/fcrepo4/commit/2ad89e2eaee4a10ee3c9ed9e04c2dc13d50b20e8" https://www.pivotaltracker.com/story/show/45769627
A. "Gorbulak" Soroka finished "Add digest and algorithm info to the datastream profile output" https://www.pivotaltracker.com/story/show/45769627
<cbeer>ajs6f++10:32
* fasseg joins10:33
<ajs6f>No sweat. I'm going to add some more XML tests to fcrepo-legacy-api to check on the back compatibility of our responses. Those schemata define ordering (why? dunno).
<eddies>https://github.com/futures/fcrepo4/pull/31#issuecomment-1456659910:35
<ff-pivotal>Frank Asseg added "Update a whole set of datastreams while incrementing the top level object version by one only" https://www.pivotaltracker.com/story/show/4577145910:36
<cbeer>still reading through it: https://github.com/ModeShape/modeshape/blob/master/modeshape-jcr/src/main/java/org/modeshape/jcr/value/BinaryKey.java
ajs6f: have you seen jcd.value.UrlBinaryValue? is that federation?10:42
<ajs6f>No idea— but if it's in JCR API, then no. Federation is a MODE-specific functionality.10:44
<cbeer>eddies--10:45
maybe i should make you rebase this.
<jcoyne>I just saw a job posting for a Software Configuration Engineer.10:46
$10 says it's a java shop
* travis-ci joins10:47
[travis-ci] futures/fcrepo4#173 (master - acdcf66 : ajs6f): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/3f523f6dadd4...acdcf664ea5c
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/5315901
* travis-ci leaves
<cbeer>should we be nervous that the BinaryKey isn't making any claim about the kind of hash they are creating?10:48
* github-ff joins10:50
[fcrepo4] cbeer force-pushed 45628143 from b9862ac to ea0f800: http://git.io/sAmcPg
fcrepo4/45628143 ea0f800 Edwin Shin: 45628143 : Store fixity data when a datastream is created or modified...
* github-ff leaves
* github-ff joins
[fcrepo4] cbeer pushed 1 new commit to master: http://git.io/AYe_5w
fcrepo4/master e8911df Chris Beer: Merge pull request #31 from futures/45628143...
* github-ff leaves
* github-ff joins10:51
[fcrepo4] cbeer deleted path-to-modeshape-config at 9525e5d: http://git.io/QEnA-g
* github-ff leaves
* github-ff joins
[fcrepo4] cbeer deleted multipart-form-data-datastream-upload at 4d3d1bd: http://git.io/VGBWBA
* github-ff leaves
* github-ff joins
[fcrepo4] cbeer deleted batch-update-single-object at c70656f: http://git.io/MviL7A
* github-ff leaves
<cbeer>(10 minute warning)10:52
3 minutes!10:57
eddies: time to move on? table this to the list?11:01
<jonathangee>whats the url for the hangout again?11:02
<VincentNG>https://plus.google.com/hangouts/_/46ed3fe3f2d8606cbc4c4056cc8bb4cce9176ebc?authuser=0&hl=en
<jonathangee>thansk
<barmintor>if you go to google+, and click on the meeting in the future, you can join the hangout there
* travis-ci joins11:07
[travis-ci] futures/fcrepo4#175 (master - e8911df : Chris Beer): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/acdcf664ea5c...e8911df8bc7a
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/5316639
* travis-ci leaves
<cbeer>VincentNG??11:12
<ff-pivotal>Chris Beer added comment: "does this need to be "backported" to the legacy API?" https://www.pivotaltracker.com/story/show/4576962711:14
<cbeer>ajs6f: can you deliver https://www.pivotaltracker.com/story/show/45628561?11:15
barmintor: https://www.pivotaltracker.com/story/show/45628143
<ff-pivotal>A. "Gorbulak" Soroka edited "Update the kitchen-sink and/or webapp repositories to nicely expose the configs for implementors to modify" https://www.pivotaltracker.com/story/show/45628561
Chris Beer edited "Multipart POST vs POST body content stream negotiation for add/modify datastream" https://www.pivotaltracker.com/story/show/45062931
Chris Beer edited "Add digest and algorithm info to the datastream profile output" https://www.pivotaltracker.com/story/show/45769627
A. "Gorbulak" Soroka added comment: "It's in there." https://www.pivotaltracker.com/story/show/45769627
<cbeer>ok, didn't see it in the commit, but didn't look at the actual output11:16
oh, i disabled auto-deploys for the iteration end
let me re-enable that
we're talking about https://www.pivotaltracker.com/story/show/45689187?11:17
<VincentNG>cbeer: so if the user provides the checksum on addDatastream, then we validate the checksum before actually adding the datastream right?11:20
* github-ff joins
[fcrepo4] ajs6f pushed 2 new commits to master: http://git.io/4XIdrQ
fcrepo4/master 12d8608 ajs6f: Added XML validity test for legacy datastream profile response
fcrepo4/master aefb665 ajs6f: Merged upstream changes
* github-ff leaves
<cbeer>VincentNG: yes; since we have transactions, though, you can do all the work and then check before the session.save() call
<VincentNG>and if you're storing the fixity on addDatastream, then we compute checksum twice?11:21
<cbeer>i don't think we have to. i think if you...
hm, so you could either extend https://github.com/futures/fcrepo4/blob/master/fcrepo-kernel/src/main/java/org/fcrepo/Datastream.java#L75 to support passing in a checksum11:22
and raise an exception or something around here:
https://github.com/futures/fcrepo4/blob/master/fcrepo-kernel/src/main/java/org/fcrepo/Datastream.java#L111
or..
so something at the REST API level, I guess11:23
i think putting it in setContent makes sense, though.. right?
i wouldn't want to read through the input stream multiple times if we can avoid it
<VincentNG>yea, that's why i was asking11:24
<cbeer>eek:
Researcher A creates a data management plan and stores in Fedora via Fedora User Interface or API.
Note: Improve SOAP/REST API to handle deposit, edit and query final research data outputs.
https://wiki.duraspace.org/display/FF/Sponsor+Use+Cases
VincentNG: yeah, i think you'd be fine adding an optional checksum argument to setContent11:25
and raise an exception (that would trigger aborting the current session, i guess)
hm. looks like my config-location change doesn't play nice with the jetty run-war stuff?11:31
i'll look at it more in the office.11:32
<barmintor>OMG ADAM DISAGREES WITH ME?!?!?!?
<cbeer>(but only a little bit!)11:34
<ajs6f>I disagree.
I say, "ZOMG BEN DISAGREES WITH ME?!?!?!" You see the subtle semantic difference?
<barmintor>:D
* travis-ci joins11:36
[travis-ci] futures/fcrepo4#176 (master - aefb665 : ajs6f): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/e8911df8bc7a...aefb665db862
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/5317585
* travis-ci leaves
<cbeer>man, if only I wrote this in the ticket:11:38
"
If a checksum fails to match, it should record this failure in the audit logs for the object and repository."
<barmintor>hush you
<ajs6f>What audit log? :)
<cbeer>ajs6f: eddies made it at code4lib.11:39
i thought.
<ajs6f>Ohhhh, right. :)
<barmintor>GO BACK INSIDE! WE'RE FIGHTING!
<VincentNG>The checksum for the file is different everytime i do an addDatastream. any idea?11:40
same file, dif checksum everytime
<ajs6f>Is the checksum being calculated against the bitstream or some larger thing, like the node?11:41
<VincentNG>bit bitstream
*the bitstream
<ajs6f>Mm.
<cbeer>VincentNG: what content do you get when you pull it back out?11:42
i wonder if you have the same weird multi-part post thing islandora was doing
<VincentNG>same content. put in a pdf and the same pdf. same file size
<cbeer>the new endpoint only takes the POST body and pushes it into the datastream
hm11:43
how are you posting content?
* kaarefc joins11:45
<VincentNG>curl -H "Content-Type:application/pdf" -X POST -F "file=@file.pdf" "http://localhost:8080/rest/objects/{PID}/datastreams/testdatastream"11:46
<cbeer>what's -F?
<ajs6f>Doesn't that make it a form post?
-F "form paramter"11:47
<VincentNG>how should i be POSTing?
<cbeer>yeah, you don't want that, i think..
<ajs6f>—data-ascii @myfile
or
—data-binary @myfile
<VincentNG>let me try that
<cbeer>so, maybe this is an argument for pulling that legacy change forward..11:48
if VincentNG and the Islandora people want to do it
<VincentNG>ya, we used it at the CDC for our ingests of external content
<ajs6f>But maybe specifying the behavior pretty carefully.11:49
<cbeer>right. i think I did a "if you have a form part called "file", use it"11:50
in fcrepo 4
at the very least, maybe we should check for that
and complain loudly
HTTP 500 or something.
i wonder what documentation or whatever suggested using form parts was a good idea.11:51
<ajs6f>Didn't this have something to do with HTML forms acting as simple clients?
<cbeer>that was my thought too. but i don't think we're interested in that use case any more11:52
(unles you're new south wales, apparently.)
<ajs6f>Sounds like something you could do against the "normal" endpoint by smearing a little Javascript on your form.11:54
<VincentNG>ok, i think --data-binary param worked. good catch on the form param, didn't think it mattered.
thanks11:55
takin a quick lunch break...be right back.
<jcoyne>barmintor: You can store an arbitrary checksum with FC3?
<barmintor>jcoyne: as ong as its one of a supported set of algorithms11:56
<jcoyne>barmintor: would that be a suitable place to put the checksum for an external datastream?
<barmintor>I can't remember if external DS's store it or not. It's unclear what it means if an external DS's checksum doesn't validate.11:57
<jcoyne>Right, I'm just looking for a good place to store it along with its dsLocation
<eddies>jcoyne: the DC datastream, of course11:58
<jcoyne>Currently I'm putting that in a second datastream.
Okay, so I'm doing it correctly now.
<eddies>everything goes there
<jcoyne><your mom joke>
<ajs6f>barmintor: for reference, here's what working with JCR versioning looks like:11:59
https://github.com/ModeShape/modeshape/blob/master/modeshape-jcr/src/test/java/org/modeshape/jcr/JcrVersioningTest.java
<barmintor>eddies: you laugh, but I have seriously considered storing checksum infor as a dc:identifier
<eddies>i pity you
<barmintor>it's a way you could answer the question "is this file already in Fedora"?
eddies: I'm evangelizing hydra and blacklight at Yale tomorrow morning, so I won't be on the standup12:04
<eddies>well, that's an interesting hack. i'd probably try to use the resource index to query for the existence of a resource w/ the checksum12:05
i didn't know yale needed more evangelizing
* barmintor shrugs
<eddies>we already rana hydracamp there a couple months ago
<barmintor>I do what I'm told. They tols me to be at Yale tomorrow at 9am to talk about hydra and blacklight
<cbeer>can you run hydra on top of dspace?12:06
cbeer--
<eddies>cbeer--
<ajs6f>Can you run Hydra on top of Sharepoint?
<barmintor>Sup dawg! We heard you like repositories, so we put a hydra on your dspace...12:07
<eddies>see, sound effects for irc would be so useful right now
<ajs6f>{raspberry}
* kaarefc leaves12:31
* VincentNG leaves12:32
<ff-pivotal>A. "Gorbulak" Soroka added comment: "One such example exists: /rest/objects/{pid} from https://github.com/futures/fcrepo4/commit/fa8191f..." https://www.pivotaltracker.com/story/show/4517870512:35
* VincentNG joins12:47
<ff-pivotal>Edwin Shin added comment: "That was my intent in having Vincent look at https://www.pivotaltracker.com/story/show/45423729" https://www.pivotaltracker.com/story/show/4517870512:49
Edwin Shin added comment: "I wanted to avoid taking on the one extreme adopting something like spring-mvc but something closer to straig..." https://www.pivotaltracker.com/story/show/4517870512:51
A. "Gorbulak" Soroka added comment: "Yes. That's (IMO) exactly where we want to go-- JAX-RS + views." https://www.pivotaltracker.com/story/show/4517870512:53
A. "Gorbulak" Soroka finished "Provide examples of HTTP conneg for fcrepo-http-api" https://www.pivotaltracker.com/story/show/4517870513:01
A. "Gorbulak" Soroka edited "Provide examples of HTTP conneg for fcrepo-http-api" https://www.pivotaltracker.com/story/show/4517870513:02
Edwin Shin added comment: "I'm very open to opinions about the various view frameworks listed under http://code.google.com/p/htmleasy/wi..." https://www.pivotaltracker.com/story/show/4517870513:05
Edwin Shin accepted "Update the kitchen-sink and/or webapp repositories to nicely expose the configs for implementors to modify" https://www.pivotaltracker.com/story/show/4562856113:06
Edwin Shin accepted "Add digest and algorithm info to the datastream profile output" https://www.pivotaltracker.com/story/show/4576962713:07
Edwin Shin accepted "Provide examples of HTTP conneg for fcrepo-http-api" https://www.pivotaltracker.com/story/show/45178705
* github-ff joins13:10
[fcrepo4] ajs6f pushed 1 new commit to master: http://git.io/cg4aEg
fcrepo4/master e9feb54 ajs6f: Removed unused qualifer annotation
* github-ff leaves
* github-ff joins13:12
[fcrepo4] ajs6f pushed 1 new commit to master: http://git.io/RuhDKA
fcrepo4/master 7bc9e57 ajs6f: Removed Java file for unused qualifer annotation
* github-ff leaves
* kaarefc joins13:19
<ff-pivotal>A. "Gorbulak" Soroka added comment: "I'm not sure this issue makes sense in isolation from the question of DI framework. This is because..." https://www.pivotaltracker.com/story/show/4576926113:27
A. "Gorbulak" Soroka edited "De-springify ModeShapeRepositoryFactoryBean" https://www.pivotaltracker.com/story/show/45769261
A. "Gorbulak" Soroka added comment: "Object owner is recorded in JCR props now as "fedora:ownerId". Object state is not, because we have..." https://www.pivotaltracker.com/story/show/4559570913:30
A. "Gorbulak" Soroka added comment: "I don't see anything for object label right now, but I think JasonDGI is (or was) working on that." https://www.pivotaltracker.com/story/show/4559570913:32
<cbeer> gah. ajs6f, i think i need your help with that configuration resource thing..
* travis-ci joins
[travis-ci] futures/fcrepo4#177 (master - e9feb54 : ajs6f): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/aefb665db862...e9feb54aa23d
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/5320475
* travis-ci leaves
<cbeer>i'm getting:
java.io.FileNotFoundException: ServletContext resource [/config/repository.json] cannot be resolved to URL because it does not exist
<ff-pivotal>A. "Gorbulak" Soroka added comment: "Do we have some consensus on whether object label is single-valued or not?" https://www.pivotaltracker.com/story/show/4559417913:33
<cbeer>so, i guess something is different when run from tests vs jetty?
(i'm going to try to recreate it locally now..)
<ajs6f>Shouldn't be. That URL arises from Resource.getURL(). Should matter what the container outside Spring is or is doing.
_Shouldn't_ matter.
<cbeer>keep telling yourself that. i'm still not sure how the kitchen-sink app even works13:34
so i'll check the fcrepo-webapp first
<ajs6f>Mm.
<cbeer>if it's not a problem there, time to learn something about kitchen-sink..
<ajs6f>When you changed the Spring config to get rid of the @Qualifier annotation, did you change the SPring config in kitchen sink?13:35
<cbeer>there isn't a spring config in kitchen sink, as far as i can tell
<ajs6f>I believe that kitchen-sink overlays some of that.
<cbeer>it magically gets it from the webapp
<ajs6f>Oh, okay.
<cbeer>i think
<ajs6f>No, I don't see any overlay there.13:36
<fasseg>eddies: I have finished a first iteration of the fixity service which does logging only and added this repository: https://github.com/futures/ff-fixity-service
<cbeer>yay, it's a problem in the webapp too
<fasseg>you can check this test to see how it works: https://github.com/futures/ff-fixity-service/blob/master/src/test/java/org/fcrepo/services/TestFixityService.java
<ajs6f>cbeer: are you doing mvn jetty:run for kitchen-sink, or mvn jetty:run-war.
<cbeer>run-war13:37
(checking run on the webapp now too..)
<fasseg>ben: you're very much invited to take a look as well
<ajs6f>Weird. It doesn't do that for me.
I run mvn jetty:run in webapp fine.
<cbeer>with the latest code?
<ajs6f>Yup.
<cbeer>really really?
<ajs6f>Well, I'll try it righ tnow.
<cbeer>hm. i wonder if it's interpreting that as a file-path and not a classpath.13:39
ooh.
yep.
<ff-pivotal>Frank Asseg added comment: "started a ff-tech thread at https://groups.google.com/forum/?hl=en&fromgroups=#!topic/ff-tech/LkKrIdGLxrQ" https://www.pivotaltracker.com/story/show/4575952113:40
<ajs6f>Mm… but it's getting it from the Spring resource...
<ff-pivotal>Frank Asseg added comment: "Added a first implementation which does logging only at: https://github.com/futures/ff-fixity-service" https://www.pivotaltracker.com/story/show/45689187
<ajs6f>Wow! I get it now too!
Weird.
You kinxed me.
s/kinxed/jinxed
But it says ServletContext resource
* github-ff joins13:41
[fcrepo4] cbeer pushed 1 new commit to master: http://git.io/B-TmAA
fcrepo4/master a438fd5 Chris Beer: explicitly use the modeshape config off of the classpath
* github-ff leaves
<ajs6f>as tho' it were trying to get it as a relative URL from the URL tree.
<cbeer>hm. well, whatever. my commit seemed to make the webapp happy
i'm stop questioning it
especially if you'll fx it all up after we have a DI framework
<ff-pivotal>Chris Beer edited "Switch from #duraspace-ff to #fcrepo for IRC" https://www.pivotaltracker.com/story/show/4575692713:42
<cbeer>not sure what that ticket's about, what needs to happen, or why you assigned it to me, eddies.
<ajs6f>Oh, cbeer. I bet I know what it was.13:43
We were explicitly using the constructor for the Spring Resource.
Then you switched it to creating a Spring Resource implicitly, as a property. Must be different semantics.
No biggie.13:44
I'll make a ticket for Eddie to explain what he means by the ticket he just gave Chris.13:45
<jcoyne>tickets++13:47
<ajs6f>But first someone make me a ticket to do that.
* jcoyne is building a ticket based finite state machine to do that13:48
<ajs6f>There better be a ticket for your work on that, jcoyne, or we'll have no way to track progress.13:50
<cbeer>hm.13:52
Caused by: java.lang.Exception: Invalid cache loader configuration!! Only ONE cache loader may have fetchPersistentState set to true. Cache will not start!
that seems.. unfortunate.
<ajs6f>Well, you can see where they're coming from. What if two stores disagree about the state of a persisted resource? First one in wins? Voting? It's not an impossible thing to figure out, but it could be hard to offer enough configurability.13:54
And do the "winners" make the "losers" change their state? Or do you go ahead with inconsistent persistence?13:55
And what about Scarecrow's brains!
* travis-ci joins14:06
[travis-ci] futures/fcrepo4#179 (master - a438fd5 : Chris Beer): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/7bc9e57c4ee5...a438fd57745c
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/5321123
* travis-ci leaves
<cbeer>oh, totally.
<ajs6f>On my way to the office for a bit. I'll check in later.14:08
<cbeer>anyway, the inability to do that makes the whole "some stuff might be off in glacier" harder.14:09
<barmintor>I need to think about it more, but I don't understand why that difference is understood as a problem, and not just an assertion by any given cache of whther it persisted those bytes.14:12
<cbeer>barmintor: so, when infinispan first starts, it needs to have a single, authoritative cache of everything it stores (apparently.. as I understand it)14:13
<barmintor>that's… unexpected. If so, Glacier is always mirrored to local disk somewhere.14:14
* barmintor shrugs
<cbeer>yeah, maybe i'm wrong. happened before.
of course, last time i tried to start it, i hung my machine14:15
so maybe i really don't know what i'm talking about.
* ajs6f leaves
<cbeer>cool. chained loaders work as advertised, at least.14:18
ok, or i don't understand what persistent state means14:21
which seems more and more likely
after reading a little more, i think it means:
when you join a cluster, go ask the cluster for data and persist it in this store.14:22
that seems more reasonable.
* barmintor groans14:42
my train from grand central to new haven leaves at 6:23AM tomorrow.14:43
* ajs6f joins14:52
Hey, no whining. Grand Central is beautiful at any time of day or night.14:54
<barmintor>it's not so beautiful when you've been on a subway since 5:15am or so.14:56
<ajs6f>Okay, I would probably be pretty unhappy at 6:23, too.
Still, it could be worse. It could be pretty much any of the airports in any city in which any of us live.14:57
<ff-pivotal>A. "Schwarmerei" Soroka added comment: "The start of an example: ""15:02
http://modeshape.wordpress.com/category/federation/" https://www.pivotaltracker.com/story/show/45464181
A. "Schwarmerei" Soroka added comment: "I believe that sitemaps are limited to 50,000 entries. Anyway, I would think we_could_ do this o..." https://www.pivotaltracker.com/story/show/4545888315:04
A. "Schwarmerei" Soroka added comment: "My real claim is that _any_ authN is a container-level concern. JAX-RS maps Java objects into HT..." https://www.pivotaltracker.com/story/show/4545824915:06
A. "Schwarmerei" Soroka added comment: "Hurting us like slowing us down? Because we moved it onto disk?" https://www.pivotaltracker.com/story/show/44012979
Chris Beer added comment: "gobbling up RAM." https://www.pivotaltracker.com/story/show/4401297915:08
Chris Beer added comment: "maybe i just need an example. I want to see fcrepo4 and OAuth work together, maybe I don't care about how." https://www.pivotaltracker.com/story/show/4545824915:09
A. "Schwarmerei" Soroka added comment: "Mm. Weird. I wonder how accessible the Lucene config is from MODE. Time for a little trawling." https://www.pivotaltracker.com/story/show/44012979
A. "Schwarmerei" Soroka added comment: "Right. I'm arguing that we don't care about OAuth. We FFfolk only care about auth. How the conta..." https://www.pivotaltracker.com/story/show/4545824915:13
A. "Schwarmerei" Soroka added comment: "IOW, if you want an example, that would be composed in (say) Tomcat code/config. Not Fedora." https://www.pivotaltracker.com/story/show/4545824915:15
Chris Beer added comment: "yes. so I want an example. I think having an example of how to do X within our context is sufficient to satis..." https://www.pivotaltracker.com/story/show/4545824915:16
A. "Schwarmerei" Soroka added comment: "Fair enough. The first thing to do would be to make fcrepo4 in any way aware of auth, which it i..." https://www.pivotaltracker.com/story/show/4545824915:17
* fasseg leaves15:20
<ff-pivotal>Chris Beer added comment: "And there's a separate ticket for that." https://www.pivotaltracker.com/story/show/45458249
<cbeer>so much more efficient to talk there than here.15:21
* fasseg joins
<cbeer>barmintor: ok, so i was totally wrong about fetchPersistentState, i think.15:25
<ff-pivotal>A. "Schwarmerei" Soroka added comment: "Groovy. Maybe this is an epic? "Fedora vs. OAuth, the Final Battle"? There's using authNZ inside..." https://www.pivotaltracker.com/story/show/45458249
<cbeer>but i'm still waiting on #infinispan telling me what fetchPersistentState's behavior actually is
<ajs6f>My persistent state is tired and awfully cranky.15:26
<cbeer>ok, now i have absolutely no idea what fetchPersistentState controls15:28
* ppound leaves
<cbeer>i started reading the docs thinking one thing
after reading them, i thought something different
in #infinispan, the first answer was the thing i originally thought15:29
but that was wrong, it's a third thing.
<ajs6f>Sounds like #infinispan needs a voting protocol to decide what Infinispan thinks about itself.
Anyway, does it block us on using INPSN to replicate?15:30
<cbeer>nope, not at all. the thing i thought it was at first would have been a problem
<ajs6f>We win by the power of confusion!15:31
<cbeer>so, with fetchPersistentState on
if a node joins the cluster
the cache store configured with fetchPersistentState will send the data in that cache store15:32
to the node that just joined.
apparently.
i've now asked for clarification in a replication/distribution world.
<ajs6f>So only one node gets to decide on authoritative reality?
Good.
<cbeer>only one cache store on a node gets to decide reality.
different nodes may elect different cache stores, i think
<ajs6f>I'm so tired of consensusless reality.
<cbeer>i think i need to catch the #infinispan people earlier15:33
i think they're mostly GMT.
<ajs6f>I;m starting to wonder if we shouldn't be trying to assemble a more formal relationship with them, and with rhauch & co. at MODE.
I don't know what that would look like.15:34
But we have a lot of good questions for them.
afk15:35
<VincentNG>Do i check code into the master or create a branch and pull request?15:56
* github-ff joins16:14
[fcrepo4] vnguyen745 pushed 1 new commit to master: http://git.io/AisOQg
fcrepo4/master 53858f8 Vincent Nguyen: Updated REST API for user provided checksums during add DS
* github-ff leaves
<ff-pivotal>Vincent Nguyen added comment: "https://github.com/futures/fcrepo4/commit/53858f813285477f4da0832147d57d80becbf175" https://www.pivotaltracker.com/story/show/4569075516:15
Vincent Nguyen edited "Update REST API for posting datastreams to allow user-provided checksums" https://www.pivotaltracker.com/story/show/45690755
* fasseg leaves
<ff-pivotal>Vincent Nguyen added comment: "I created a WIKI to compare implementations and view frameworks: https://github.com/futures/fcrepo4/wiki/..." https://www.pivotaltracker.com/story/show/4517870516:25
* travis-ci joins16:30
[travis-ci] futures/fcrepo4#180 (master - 53858f8 : Vincent Nguyen): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/a438fd57745c...53858f813285
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/5325179
* travis-ci leaves
<ff-pivotal>Chris Beer started "Configure Infinispan to store multiple copies of the data in different directories" https://www.pivotaltracker.com/story/show/4569056916:31
* VincentNG leaves16:32
<ff-pivotal>Chris Beer added comment: "https://gist.github.com/cbeer/5112004" https://www.pivotaltracker.com/story/show/4569056916:33
* kaarefc leaves16:37
<cbeer>hm. should i commit that config somewhere? i guess probably so16:41
<ff-pivotal>Chris Beer added comment: "https://github.com/futures/fcrepo4/commit/46dd92a6e54e3e21c846e9ef9495adb35f7c8aae" https://www.pivotaltracker.com/story/show/4569056916:43
* github-ff joins
[fcrepo4] cbeer pushed 1 new commit to master: http://git.io/URGQ1Q
fcrepo4/master 46dd92a Chris Beer: add an example of a chained infinispan configuration
* github-ff leaves
<ff-pivotal>Chris Beer added comment: "Run in the webapp as e.g., MAVEN_OPTS="-Dfcrepo.infinispan.cache_configuration:config/infinispan_configuratio..." https://www.pivotaltracker.com/story/show/4569056916:46
Chris Beer edited "Configure Infinispan to store multiple copies of the data in different directories" https://www.pivotaltracker.com/story/show/4569056916:47
Chris Beer finished "Configure Infinispan to store multiple copies of the data in different directories" https://www.pivotaltracker.com/story/show/45690569
<cbeer>ok, now to figure out why the fcrepo4 is so unhappy.16:50
<ff-pivotal>Chris Beer added "Fix fcrepo4.fcrepo.org deploy to persist data across deploys." https://www.pivotaltracker.com/story/show/4580583516:51
Chris Beer started "Fix fcrepo4.fcrepo.org deploy to persist data across deploys." https://www.pivotaltracker.com/story/show/45805835
Vincent Nguyen added comment: "Used Query Params: http://localhost:8080/rest/objects/{pid}/datastreams/{dsid}?checksumType=SHA-1&checksu..." https://www.pivotaltracker.com/story/show/4569075516:57
* travis-ci joins16:58
[travis-ci] futures/fcrepo4#181 (master - 46dd92a : Chris Beer): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/53858f813285...46dd92a6e54e
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/5325844
* travis-ci leaves
* ajs6f leaves17:03
<ff-pivotal>Chris Beer added comment: "Should this be back-ported to the legacy api? (Eddie: should we even be backporting stuff at this point?)" https://www.pivotaltracker.com/story/show/4569075517:18
<cbeer>barmintor? eddies? i've got a maven question.17:53
<barmintor>I'm here
<cbeer>ok, so, the fcrepo-kitchen-sink doesn't seem to be pulling in the LATEST dependencies
i've seen something about the updatePolicy tag
should we be using that?17:54
<barmintor>the latest dependencies? from snapshot you mean?
<cbeer>yeah17:55
<barmintor>b/c as far as I know you have to actually specify the versions you want of rleeased deps
(unlike eg bundler)
yeah, we need a snashot policy. otherwise it only checks once a day17:57
<cbeer>the default updatePolicy is "daily"
yeah
* jcoyne leaves
<cbeer>or, we only re-deploy once a day
that's fine too, i guess
<barmintor>I think that erodes the utility of snapshots, myself
<cbeer>good
<barmintor>maybe the once-per-session option?17:58
* barmintor pounds his head
this library sucks.
is this channel logged?17:59
<ff-pivotal>Chris Beer added "Update fcrepo-kitchen-sink updatePolicy for snapshots to be more frequent than 1-a-day." https://www.pivotaltracker.com/story/show/45810447
Chris Beer edited "Update fcrepo-kitchen-sink updatePolicy for snapshots to be more frequent than 1-a-day." https://www.pivotaltracker.com/story/show/45810447
<cbeer>Hi, ff_logger!
<barmintor>noted
I don't like untested java libraries that break
<cbeer>eddies: it'd be super if you picked up that ticket and made it work before I wake up ^^^18:00
barmintor: tested libraries don't generally break.
might be a correlation there.
<barmintor>Mmm
* ajs6f joins18:14
* barmintor leaves18:57
<ff-pivotal>Chris Beer edited "Update fcrepo-kitchen-sink updatePolicy for snapshots to be more frequent than 1-a-day." https://www.pivotaltracker.com/story/show/4581044719:07
* ajs6f leaves19:23
<cbeer>ah, hm: mvn clean install -U19:50
* cbeer leaves20:04
* ff-pivotal leaves
* cbeer joins20:10
* jcoyne joins20:50
* jcoyne leaves20:56
<cbeer>https://community.jboss.org/thread/22220121:15
taking it to the source.
* futures-git joins21:46
[icemelt] cbeer pushed 1 new commit to master: http://git.io/I3TgEw
icemelt/master b580fc4 Chris Beer: Update README.md
* futures-git leaves