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

Using timezone: Eastern Standard Time
* kaarefc joins04:04
* nbanks_ joins05:27
* eddies1 joins05:28
* eddies1 leaves
* eddies1 joins
* eddies leaves05:36
* nbanks leaves
* eddies1 leaves06:19
* eddies joins06:58
* eddies leaves
* eddies joins
* eddies1 joins07:33
* eddies1 leaves07:34
* eddies1 joins
* eddies leaves07:35
* eddies1 leaves08:16
* eddies joins08:31
* eddies leaves
* eddies joins
* VincentNG joins08:43
<pivotal-bot>Vincent Nguyen edited "Add HTTP 1.1-friendly cache headers to our REST API responses" https://www.pivotaltracker.com/story/show/4545854508:55
Vincent Nguyen edited "Add datastream delete to the batch-modify-object endpoint" https://www.pivotaltracker.com/story/show/4550656308:57
Vincent Nguyen started "Object search-like API built on top of JCR query API" https://www.pivotaltracker.com/story/show/4546471108:59
Edwin Shin added "Backup a running repository" https://www.pivotaltracker.com/story/show/4610822909:27
Edwin Shin edited "Backup a running repository" https://www.pivotaltracker.com/story/show/46108229
Chris Beer added comment: "http://planet.jboss.org/post/new_repository_backup_and_restore_in_modeshape_3" https://www.pivotaltracker.com/story/show/4610822909:33
Chris Beer estimated "Backup a running repository" as 1 point https://www.pivotaltracker.com/story/show/46108229
Chris Beer added "Install Fedora via an RPM" https://www.pivotaltracker.com/story/show/46108605
Chris Beer edited "Install Fedora via an RPM" https://www.pivotaltracker.com/story/show/46108605
* nbanks_ leaves09:46
* nbanks joins10:10
* kaarefc leaves10:18
<pivotal-bot>Chris Beer added comment: "I failed at doing this in this branch: https://github.com/futures/fcrepo4/tree/broken-authn" https://www.pivotaltracker.com/story/show/4545805510:26
* barmintor joins10:28
* kaarefc joins10:35
<pivotal-bot>Chris Beer added "HTTP/1.1 Range requests" https://www.pivotaltracker.com/story/show/4611349110:39
Chris Beer added comment: "Extra points if the filesystem federation maintains directory hierarchy present on the filesystem." https://www.pivotaltracker.com/story/show/4603960110:41
<cbeer>pretty sure that's what federation does, but might be a killer feature for us.
* kaarefc leaves10:50
* kaarefc joins
<barmintor>Jiminy crickets, OR is an expensive registration.10:55
<cbeer>always is.10:58
that's how they pay for the fancy dinners.
<eddies>on a call…will be right there11:01
<cbeer>barmintor: joining us?11:03
<barmintor>I always forget it's a two-click process11:04
<pivotal-bot>Chris Beer edited "Fixity-service should support delegating check to fcrepo4" https://www.pivotaltracker.com/story/show/4595137911:06
Benjamin Armintor edited "Fixity-service should support delegating check to fcrepo4" https://www.pivotaltracker.com/story/show/45951379
* fasseg joins11:07
<cbeer>eddies: http://fcrepo4.fedora4lib.org/rest/objects/sufia:td96k2517/datastreams/content/fixity11:10
<pivotal-bot>Chris Beer added comment: "http://fcrepo4.fedora4lib.org/rest/objects/sufia:td96k2517/datastreams/content/fixity" https://www.pivotaltracker.com/story/show/4569065911:12
Edwin Shin accepted "Add REST API service to request fixity check of an object/datastream" https://www.pivotaltracker.com/story/show/4569065911:13
Edwin Shin accepted "Fixity checks against every instance of a data blob in a distributed-cache mode" https://www.pivotaltracker.com/story/show/45628025
Edwin Shin edited "Extend "RepositoryService" to handle performing fixity checks and return a status for each cache store" https://www.pivotaltracker.com/story/show/4600325711:14
<cbeer>barmintor: i wonder what's in getProperties here: http://docs.jboss.org/infinispan/5.1/apidocs/org/infinispan/loaders/AbstractCacheLoaderConfig.html#properties11:15
or getCacheLoaderClassName
<barmintor>class name won't be unique11:16
necessarily anyway
but the properties thing is promising, if we can find some info about what they are.11:17
<cbeer>yeah, i'd check properties first. i think there's an id attribute or something we could use
<barmintor>I feel kind of weird about not seeing that when I was working on the issue. I guess I was sleepy.11:18
guess should I make a seperate ticket for that, since eddies accepted the old one?11:19
* kaarefc leaves
<cbeer>yeah, probably.
<pivotal-bot>Benjamin Armintor added "Pull in a cache id from the cache config properties to include in the fixity status" https://www.pivotaltracker.com/story/show/4611773511:20
Benjamin Armintor edited "Pull in a cache id from the cache config properties to include in the fixity status" https://www.pivotaltracker.com/story/show/4611773511:21
* ajs6f joins11:22
<cbeer>oh, i'm probably not poking that through the BinaryCacheStore class though, am I11:23
<barmintor>I think you have access to the lowlevel store
i remember looking at the cacheconfig class
* barmintor goes to look11:24
<cbeer>yeah, that might have been old though.
at some point i added an abstraction over the ISPN cachestore because modeshape might store binaries outside ISPN in some configurations
ok, let me figure out a method to add to the LowLevelCacheStore that returns some reasonable information11:25
<barmintor>we could push the fixity check for the blob down into the LowLevelCacheStore11:26
LowLevelCacheStore#fixity(BinaryKey key)
have that return a status object, have LowlevelService aggregate them11:27
seem like appropriate encapsulation?
<cbeer>i considered that at some point and didn't do it.. let me try to remember if i had a good reason
<cbeer>i also wouldn't do it now either, because i think clustering is going to cause trouble.11:28
<barmintor>I don't have your perspective on that, but it seems like LowLevelService is always going to be aggregating reports from some other things, and we could rely on those things to give an individual status and identify themsewlves11:30
OR, I don't know how the clustering works.
which seems pretty likely
<cbeer>not sure anyone does.11:31
hm. that's annoying11:35
the CacheStoreConfig interface doesn't define properties11:36
it's on the AbstractCacheStoreConfig
<barmintor>well, isn't that a poke in the eye11:43
<cbeer>and looks like we should file a featurerequest with MODE to expose something useful from their BinaryStore11:47
ok, just need to figure out how to write a test for this11:51
i really don't want to learn how to construct modeshape repositories programmatically11:53
* avessey leaves11:58
* VincentNG leaves
<cbeer>i hate writing unit tests in java12:01
more trouble than it's worth.12:02
* VincentNG joins12:03
<cbeer>ok, running tests.12:05
<pivotal-bot>Chris Beer started "Pull in a cache id from the cache config properties to include in the fixity status" https://www.pivotaltracker.com/story/show/4611773512:07
<cbeer>i guess i should just steal that whole ticket
<pivotal-bot>Hello, cbeer
<pivotal-bot>Benjamin Armintor edited "Pull in a cache id from the cache config properties to include in the fixity status" https://www.pivotaltracker.com/story/show/4611773512:08
<barmintor>^^ making myself the requestor then
<pivotal-bot>Edwin Shin added comment: "fyi: https://github.com/futures/fcrepo4/commit/18014998a5c780b721b29ac005f27332c60dfaf0?w=1
makes it a littl..." https://www.pivotaltracker.com/story/show/44962643
Edwin Shin edited "Extract (and/or update-in-place) the LowLevelStorageService. applyDigestToBlobs method to return FixityStatus objects " https://www.pivotaltracker.com/story/show/4600543112:09
* github-ff joins12:13
[fcrepo4] cbeer pushed 1 new commit to master: http://git.io/Ap8yvw
fcrepo4/master 30d3b24 Chris Beer: update fixitystatus to contain an identifier for the cache, expose a 'friendly' from the lowlevelcachestore. fixes #46117735
* github-ff leaves
<cbeer>hm. guess that wasn't the right syntax12:16
oh. stupid. i'd need to start the commit with [Fixes ...]
<pivotal-bot>Chris Beer edited "Pull in a cache id from the cache config properties to include in the fixity status" https://www.pivotaltracker.com/story/show/4611773512:17
<cbeer>also, that jax-b thing in the test is pretty clever. i don't think that occurred to me before.12:20
<pivotal-bot>Chris Beer added "(Placeholder ticket to do something useful with webhooks)" https://www.pivotaltracker.com/story/show/46122883
<cbeer>i'm kinda at a loss what to do with them that we wouldn't just do with the jms or eventbus right now.12:22
<barmintor>cbeer: It'd be easier to outsource indexing to hydra with a webhook then by implementing an eventlistener on the hydra side12:23
off the top of my head- might not be the best use case12:24
<cbeer>although you lose all sorts of flow control and things you actually care about in a non-artificial scenario
<barmintor>Hmm. Does hydra care about flow control in that scenario? Or are you saying it's artificial?12:25
<cbeer>well, if you really did this, you'd want to dedupe messages, filter them, make sure hydra actually got your message and successfully processed it, etc.. things that a message queue is actually good at12:26
<pivotal-bot>Vincent Nguyen added comment: "A check to see if the node was created was happening before the createDatastreamNode call. updateReposit..." https://www.pivotaltracker.com/story/show/44962643
* barmintor shrugs12:27
I can only think of on-the-cheap integrations as use cases for it
oh well. i made the placeholder ticket. that's all i was obligated to do today :)12:28
now to figure out why the fcrepo build was unstable.
<barmintor>there's also the whole "does fcrepo4 authenticate for webhook calls" can of worms
<cbeer>i think we just suck at asynchronous messages in integration tests12:29
i'll blissfully assume that's either a solved problem, or someone elses problem.
* travis-ci joins12:32
[travis-ci] futures/fcrepo4#222 (master - 30d3b24 : Chris Beer): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/0f8a3ff83bbe...30d3b24f7c5e
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/5475616
* travis-ci leaves
<cbeer>hm. barmintor: on http://fcrepo4.fedora4lib.org/rest/objects/sufia:td96k2517/datastreams/content/fixity12:44
where are the two statuses coming from?
is getfixity returning them twice?
<barmintor>looking now12:45
<cbeer>hm. now i'm also confused why computedChecksum doesn't have the namespace part too12:46
i guess that's what the debugger is for12:49
* jcoyne joins12:50
<cbeer>oh, i see what's happened
my tests were only with a modeshape transientbinarystore12:51
when you're using ISPN, you actually have two different references to the same store (one for the "metadata" store and one for the "binary" store)
* barmintor grumbles about some bad code fomr a JISC project
<cbeer>at least their projects release code.12:54
<barmintor>new StreamSource(new FileInputStream(new File(url.getFile())))
^^ this is a thing a crazy person does
but yes, at least they get some things out there
afk a sec
* cbeer checks that he's never written code for a JISC project12:56
whew. got worried for a sec.
<barmintor>I'm being unfair. I'm using the code because there's not many options in the space, and I don't want to write my own z3950 client13:08
but that's a really frustrating bug
* github-ff joins
[fcrepo4] cbeer pushed 2 new commits to master: http://git.io/S79UIA
fcrepo4/master 4cc69e4 Chris Beer: be nice and expose an #equals method for LowLevelCacheStores
fcrepo4/master 9521529 Chris Beer: use an ImmutableSet for the list of caches, mostly to de-dupe metadata + binary caches.
* github-ff leaves
* nbanks leaves13:22
* travis-ci joins13:28
[travis-ci] futures/fcrepo4#223 (master - 9521529 : Chris Beer): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/30d3b24f7c5e...9521529dd2bb
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/5477105
* travis-ci leaves
<cbeer>ok, looks like the duplicate fixity check is fixed.13:32
<fasseg>cbeer: the fixity-webapp woul dbe ready for deployment on jenkins i guess. Just let me know if you got time..13:35
argl im off for ten minutes, sry13:36
ill come back to you chris...13:37
<VincentNG>are we currently setting the label/dc:title at all?13:45
* github-ff joins13:47
[fcrepo4] cbeer pushed 1 new commit to master: http://git.io/XvDjbg
fcrepo4/master b22a74c Chris Beer: make the FixityResult computedChecksum a URI, and pass it consistently through the stack.
* github-ff leaves
* nbanks joins13:48
* avessey joins13:54
* nbanks leaves14:02
<cbeer>fasseg: is it possible to wire the fixity app into the fcrepo-kitchen-sink?14:03
(and, i guess, have jenkins build the snapshot artifact?)
i'd definitely prefer that over deploying two different webapps, wiring them in through the proxy, etc, etc.14:04
<VincentNG>cbeer: are we currently setting the label/dc:title at all?14:05
* travis-ci joins14:06
[travis-ci] futures/fcrepo4#224 (master - b22a74c : Chris Beer): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/9521529dd2bb...b22a74cb3681
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/5478013
* travis-ci leaves
<fasseg>cbeer: re14:08
cbeer: I'll take a look at kitchen sink from an integration point of view..
hmm shoul dbe possible, although the fixity-webapp uses a different jax-rs impl: resteasy...14:09
dont know if this can be run side by side with cxf14:10
but i guess one could wire the resources together using cxf and drop resteasy for the kitchen sink integration14:11
but this would be something cooked up, not really updating itself when an update to the fixity-webapp is done...
so this would involve some manual work each time the fixity webapp gets updated :/
Do you want me to try to wire them together?14:15
<pivotal-bot>Frank Asseg edited "Create HTML representation for fixity-service results" https://www.pivotaltracker.com/story/show/4595022114:19
* nbanks joins14:25
<cbeer>darn. you can't inject arbitrary properties to the cache store configuration14:39
fasseg: don't we want to switch the other fcrepo stuff to resteasy?14:40
and, aren't they both jax-rs implementations? are we using anything unique to one of them?
(all I know about is the difference in the jax-rs:server context)
<eddies>cbeer, fasseg: if the automated deploy to tomcat is working…better off deploying the two separate webapps14:47
<cbeer>we're not deploying to tomcat.14:48
<eddies>well then, there's that
well, i think it's still going to be easiest (for now) to run two separate jetties on two ports14:49
when we refactor fcrepo4 to use resteasy, then i think the overlay on kitchen-sink might go more smoothly14:50
<eddies>(or jersey for that matter)
yeah, i know
<cbeer>what are we doing that isn't portable anyway?14:51
or is that just how these jax-rs things work?
<eddies>i'm not sure what would happen with two jax-rs providers in the same classloader
that doesn't sound like a good thing though
<cbeer>no, i mean, why can't we just use CXF with the fixity app?
<eddies>cause we already know that we don't want cxf14:52
it's a bit of a toss-up to me if we should use resteasy or jersey, but we definitely don't want cxf
no component scanning, so we have to keep editing config rather than just dropping in a jar with new functionality14:53
<cbeer>yeah, but that's all external to the code itself
<eddies>the config?
<cbeer>yeah, isn't that just that one spring file?
(or, i really don't want to set up an entirely parallel infrastructure to do this one thing if it's easy enough to shoe-horn it into the existing app.)14:54
or we should make the other app right
both of which i (in my naivity) think should be faster.14:55
<eddies>well, we should deploy to a running servlet container (e.g. tomcat) and not use the maven jetty plugin
and then we could deploy other webapps there too, e.g. solr
<cbeer>ok, let me remind myself what the problem with dropping the kitchen sink war into tomcat was.14:56
<fasseg>cbeer: it's all portable, but it's configured using spring-resteasy beans, and this would have to be changed to a cxf config...14:57
and im not sure if you can have both impls run side by side for different resources
<fasseg>so there would be some work involved
<cbeer>isn't that just in https://github.com/futures/fcrepo-fixity/blob/master/fcrepo-fixity-webapp/src/main/resources/context-fixity-web.xml though?14:58
<fasseg>and web.xml...
<cbeer>both of which the kitchen sink is providing, right?14:59
sorry, i'm still not understanding why this is a big deal.
of course, i don't understand servlet containers at all
or jax-rs implementaitons
<fasseg>no big deal: but "there would be some work involved"15:00
certainly not a day or sth
<eddies>cbeer: in your gist, there aren't any actual errors though
<cbeer>picky picky.15:01
<eddies>we get similar SEVERE warnings from fcrepo3
well, i mean, the webapp still runs, right?
<cbeer>gist updated.
<fasseg>but what's the problem with running two differnt instances, that would mean no work, and we could build directly from github
<eddies>now who's being picky?
oh fine. *now* there's a stacktrace
<cbeer>no, the webapp doesn't run15:02
fasseg: it doesn't mean no work. it means:
- setting up a new jenkins build
- setting up a new post-commit hook with a new ssh key
- setting up the same janky listener thing i'm doing on futures6
- setting up a new init.d service for it
- setting up an apache proxy to the new port
and probably other things i've forgotten15:03
<fasseg>i see so it's a matter of effort on your side of things...
<cbeer>and if it's just a matter of putting 1 line in the kitchen sink app, i'd rather see us do that. but i guess that isn't the case here?
(and, in which case, i'd really like to understand why it isn't... or are we in this situation for each and every external service we're making until we do something to the kitchen sink?)15:04
<fasseg>like i said im usre its possible to wire those two together, but i do not neccessarily think they belpng together...
<eddies>fasseg: well technically, it's your ticket so it's more work for you either way ;-)
the fixity webservice was meant to be standalone as i understand it and, that would mean wiring it to the kitchen sink app15:05
<eddies>yeah, the fixity service is technically scoped as a standalone service
<fasseg>well id be happy to setup jenkins
or the jenkins project
<cbeer>but isn't the kitchen sink app designed to contain all these random things we want but treat as external or optional services?
<eddies>i think that getting it to run on tomcat is the "right" thing to do
kitchen sink is for wiring in optional dependencies15:06
not combining different webapps
like, i wouldn't put solr in the kitchen sink
cbeer: is that stacktrace from fcrepo4 webapp or kitchen-sink?15:07
<cbeer>kitchen sink.15:08
<eddies>and it doesn't happen w/ fcrepo4 or just haven't tried?
<cbeer>haven't tried. we're not deploying fcrepo4 on futures6 at all any more, so i don't have those artifacts handy.
<eddies>because it's a pretty big bug anyway if we can't deploy to tomcat
<cbeer>hm. could tomcat and jetty be sharing the same jmx whatever it is?15:10
<eddies>no idea. i don't know squat about jmx15:11
here's a link to the latest snapshot build of fcrepo-webapp: https://oss.sonatype.org/content/repositories/snapshots/org/fcrepo/fcrepo-webapp/4.0-SNAPSHOT/fcrepo-webapp-4.0-20130313.175043-83.war
<fasseg>so should i go ahead do the jenkins setup for the fixity webapp?15:14
<cbeer>same error on fcrepo-webapp15:15
i guess that's good.
<eddies>and you still have jetty running?15:17
worth a shot temporarily killing jetty?
i have to present in a bit, so afk15:18
<cbeer>i've killed jetty
hm. it seems like it's a different error on a fresh restart of tomcat15:19
maybe i should remove fedora 3 from the webapps.. 40s to restart is ridiculous
after removing solr and fcrepo 3.x and gsearch it starts fine
or not.
hm. it's bad when i search the new error i'm seeing and see my own gist about it15:24
and here's what i said last month:15:25
ok, lost interest in deploying to tomcat, ran into an NPE: https://gist.github.com/4693633 14:51
<ajs6f> That's werid. It can't build the repo because of something going wrong with the transaction manager.
ajs6f: want to help now?
* fasseg leaves
<cbeer>i wonder if it has anything to do with those ObjectStore and PutObjectStoreHere directories we see in the webapp under jetty?
<ajs6f>I just got back here from a smoke. What's wrong?15:27
when i try to deploy fcrepo-webapp under tomcat
<ajs6f>Hm. Is there a transaction manager already configured for the container (Tomcat)?15:30
<cbeer>from what i've read, tomcat doesn't provide one, no.15:33
<ajs6f>Hm. We're using JBoss JTA. Let me see if JBoss has any info in their docs.15:34
<cbeer>(use JBoss AS of course!)
<eddies>starts up fine for me locally15:35
(fcrepo-webapp deployed to a local copy of tomcat 7.0.37)
<cbeer>hm. this is tomcat 6.
<cbeer>i can try to update
hey, there is a package
is fcrepo 3.x tomcat7 friendly?15:37
<eddies>oh. yeah, technically we're requiring servlet 3.0
and that's tomcat 7
yes to 3.x
<ajs6f>Maybe we can try updating to Tomcat7 and see if the JTA problem goes away. {crosses fingers}15:38
<eddies>i see in the logs that we must be transitively including servlet-2.5-api as a dependency, so should exclude that, but tomcat knows to ignore it, it seems anyway
<cbeer>well, this is promising. i ran out of heap before something else broke
<eddies>success is just a different kind of failure.15:39
sounds like a de-motivation poster
<ajs6f>I don't mind working on problems. I just want the problems to change occasionally.15:41
<cbeer>ran out of heap again.15:42
* nbanks leaves
<cbeer>eh, ram is cheap.15:43
Caused by: org.infinispan.config.ConfigurationException: Directory /var/lib/tomcat7/target/FedoraRepository/storage/FedoraRepository does not exist and cannot be created!
at org.infinispan.loaders.file.FileCacheStore.start(FileCacheStore.java:369) ~[infinispan-core-5.2.1.Final.jar:5.2.1.Final]
at org.infinispan.loaders.CacheLoaderManagerImpl.start(CacheLoaderManagerImpl.java:152) ~[infinispan-core-5.2.1.Final.jar:5.2.1.Final]
... 58 common frames omitted
<ajs6f>A permissions problem?
O, wait, no.15:44
<cbeer>or a, it shouldn't try to put files there problem
<ajs6f>It's the hardcoded repo location.
<cbeer>there's a property i can set
if i can figure out where in tomcat7/ubuntu to do that
<ajs6f>Yay for parameterization!
<cbeer>and i guess i should do the java home explicitly15:45
Caused by: com.arjuna.ats.arjuna.exceptions.ObjectStoreException: ARJUNA012225: FileSystemStore::setupStore - cannot access root of object store: PutObjectStoreDirHere/ShadowNoFileLockStore/defaultStore/
ajs6f: can you figure out how/where to set that?
<ajs6f>Oh, that PutObjectStoreDirHere/
No, I don't. I don't even know what it is or why it exists.15:47
We maybe could ask in #modeshape?
Or #infinispan?
Actually, it looks like that coming from JBossJTA…
which might explain why MODE doesn't mention it at all.15:48
cbeer— check this:15:49
Looks like there's a system property to set?
That's crummy, but it's a start.
<cbeer>is that something i set with -Dblahblahblah?15:50
(looks like activemq has a similar problem)15:51
i'm just creating the directories for now.
ok, tomcat7 seems like it worked15:52
i guess we declare fcrepo4 tomcat7+ huh?
we're servlet 3.0, so we're tomcat7
ok. really afk for a bit15:53
<ajs6f>I think so (just a system property). For ActiveMQ we can use Spring to config properly, but a system prop seems fine for now.
It seems like that's some kind of datastore for JBossJTA.
<cbeer>ok, now to restore solr. i'm going to skip fcrepo 3.x and the gsearch/solr 3.6 nonsense.15:54
<ajs6f>Be sure to add the fcrepo3 demo apps.15:55
* kaarefc joins15:56
<pivotal-bot>Chris Beer added "Update islandora to point at new, tomcat location for fcrepo4" https://www.pivotaltracker.com/story/show/46142275
Chris Beer added "Remove crappy temp files that leak into the container directory" https://www.pivotaltracker.com/story/show/4614232915:57
<cbeer>ok, got that sorted.16:07
taking a break.16:10
* nbanks joins16:11
* ajs6f leaves16:17
* nbanks leaves16:22
* VincentNG leaves16:40
* nbanks joins16:51
<eddies>i think this was easily the latest i have ever had to present17:03
cbeer: is kitchen-sink deploying on tomcat now?
<eddies>you got that all sorted?
<cbeer>eddies++ for figuring out that tomcat7 worked
<eddies>was the other stuff just directory permissions?
well, if there were any "special" things had to be done we should note it, perhaps just in the readme for fcrepo-webapp17:06
and we can sort out later if there's anything we can do to mitigate
like that whole putobjecthere directory creation i thought was something addressed in MODE 3.2.017:07
but i can't find the ticket
* nbanks leaves
<eddies>cbeer: i take it http://fcrepo4.fcrepo.org/fcrepo/ is our new Fedora url17:09
there's tickets in the icebox for fix the stupid crap
<eddies>oh, and to repoint islandora and hydra too
or did you already do hydra?
<cbeer>hydra is good
<eddies>this is blowing up on me: http://hydra.fcrepo.org/files/td96k251717:11
(first link on http://hydra.fcrepo.org/)
<cbeer>RuntimeError (You're connecting to a Fedora 0.0 repository. Rubydora requires Fedora >= 3.4):17:12
how do i get our xml response to describe?
<eddies>you mean from http://fcrepo4.fcrepo.org/fcrepo/rest/describe ?17:14
curl -H "Accepts: application/xml" http://fcrepo4.fedora4lib.org/fcrepo/rest/describe
is giving me html
<eddies>that's better than the 500 error i just got...sec17:15
<cbeer>i'll look again tonight
i think
at least before standup
<eddies>oh that's just cause i typed "Accept" (no 's'")
yeah. only html17:16
<cbeer>i wonder if that's a problem under jetty too
and it's been broken but no one has noticed
<eddies>huh. the resource is annotated with text/xml and application/json17:18
but they all just return html
<cbeer>look up one
there's an html resource there
anyway, bbl.
* kaarefc leaves17:22
<eddies>cbeer: curl -H "Accept: text/xml" http://fcrepo4.fcrepo.org/fcrepo/rest/describe
It *is* "Accept", not "Accepts"
* avessey leaves17:24
* jcoyne leaves17:33
* nbanks joins17:35
* nbanks leaves17:47
* nbanks joins18:17
* nbanks leaves18:28
* nbanks joins18:56
* nbanks leaves19:09
* nbanks joins19:37
* nbanks leaves19:49
<cbeer>eddies: hm. i bet we don't send any particular accept header in rubydora20:06
i wonder how JAX-RS figures out which response to send lacking a header.. order maybe?
* nbanks joins20:17
* nbanks leaves20:29
* nbanks joins20:59
<cbeer>yup, rubydora is getting the html21:10
* nbanks leaves
* jcoyne joins21:26
* nbanks joins21:39
* nbanks leaves21:50
* nbanks joins22:19
* github-ff joins22:24
[fcrepo4] cbeer pushed 1 new commit to master: http://git.io/RYgUAQ
fcrepo4/master ff0e6d6 Chris Beer: does order matter for jax-rs resources?
* github-ff leaves
* nbanks leaves22:30
* travis-ci joins22:39
[travis-ci] futures/fcrepo4#225 (master - ff0e6d6 : Chris Beer): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/b22a74cb3681...ff0e6d60c317
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/5489140
* travis-ci leaves
* nbanks joins22:59
<cbeer>hm. not so much23:01
* nbanks leaves23:10
<cbeer>hi eddies23:55
<eddies>i don't get how rubydora was working before if it was never sending that accept header23:56
<cbeer>eddies: it was getting the xml content before
<eddies>oi. so fcrepo4 changed in the interval?
<cbeer>working theory.23:57
i'm checking under jetty now though, just to be sure
<eddies>hey, so i was thinking about boston for april
for our next face2face
reasons being...
i) wgbh (video, large files use case)
ii) boston public library, dpla project (large scale repo)23:58
although i have no idea what they're using for software
<cbeer>for dpla?
<eddies>iii) dspace committers at harvard and mit
<cbeer>when is the hydra meeting?23:59
<cbeer>(not that i mind boston. i think i even suggested it at one point.)00:00
<eddies>i don't have it in my calendar
<cbeer>oh, june.
early june
that's kinda annoying.00:01
<eddies>to go to boston twice or to have a meeting there in early june?
<cbeer>twice, and then OR
because i'm complaining to you about air travel :P
<cbeer>(yeah, i think fcrepo4 changed in the interval. going to kill the html response to be sure.)
anyway, yes, i think boston has a lot of merit. and not inconvenient for our europeans either.00:03
<eddies>i have a hard time trying to make a strong argument even to myself about copenhagen =)
and this time, i think we could essentially pick anywhere we like in boston area00:04
maybe not south boston00:05
<cbeer>hm. what's the cape like in april?
<eddies>oh! now that's using your noggin00:06
<cbeer>or, kcariani has a house on the cape
<eddies>i'll let you ask her ;-)
actually, my friend's dad does too, but i think he's actually there every weekend00:07
<cbeer>mind you, i don't know how we'd get there from BOS, but..
hm. i don't understand jax-rs content negotiation at all00:08
<eddies>1:35min bus ride from south station
or rent a car
i think i'd go for the rental. as long as we don't let barmintor pick the rental00:09
<cbeer>oh, yeah, they do rent cars, don't they.
<eddies>since he likes to pick really bad rates from hertz ;-)
hrmph. airbnb rentals in cape code == not so cheap00:11
oh wait. nm. not so bad. i forgot to divide by the number of ppl. haven't had any coffee yet00:12
<cbeer>If the client request does not have an Accept HTTP header, then by default the XML format is returned.00:14
i guess there's nothing stopping me from setting the right header in rubydora00:15
:accept => '*/*; q=0.5, application/xml'
that's the default header
<eddies>application/xml wasn't working for me00:16
<cbeer>ooh. we have produces(text_xml
* github-ff joins00:20
[fcrepo4] cbeer pushed 1 new commit to master: http://git.io/zSydzQ
fcrepo4/master 488fa87 Chris Beer: TEXT_XML and APPLICATION_XML are different, and i guess we should serve both.
* github-ff leaves
<eddies>ok. have to take the dog out and run some errands. have a good night.00:22
* travis-ci joins00:36
[travis-ci] futures/fcrepo4#226 (master - 488fa87 : Chris Beer): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/ff0e6d60c317...488fa8701b73
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/5490387
* travis-ci leaves
* jcoyne leaves00:57

Generated by Sualtam