Log of the #fcrepo channel on chat.freenode.net

Using timezone: Eastern Standard Time
* dwilcox joins07:53
* jrgriffiniii joins07:59
* mhwood joins08:18
* tolloid joins08:39
* tolloid leaves08:40
* tolloid joins08:41
* dwilcox leaves08:52
* dwilcox joins08:59
* osmandin joins09:01
* ajs6f joins09:04
* mhwood leaves09:14
<ajs6f>awoods: That Spring stuf in the migrator?09:15
Here is what it is doing: https://github.com/fcrepo4-labs/fcrepo4-upgrade-utils/blob/tech-md-migrator/src/main/resources/spring/master.xml
He's not loading the repo Spring. He's loading hisown.
<awoods>ajs6f: albeit a small master.xml09:16
<ajs6f>awoods: Yes, which is okay for him, because he just wants a JCR repo. I want that, plus a translator chain, plus a ContainerService. The firs and last I could rewire in Spring (bleh) but this utility cannot know what people have used for a translator chain.09:17
I have to have the real repo Spring wiring.
<awoods>ajsf6: What about this: https://github.com/fcrepo4-labs/fcrepo4-upgrade-utils/blob/tech-md-migrator/src/main/java/org/fcrepo/upgrade/utils/TechnicalMetadataMigrator.java#L6309:18
<ajs6f>awoods: Yeah, he's getting that from the code via autocreation, that woul do for the ContainerService. Like I said, that's not the issue. The issue is the translator chain.
<awoods>ajs6f: so what is your plan?09:19
<ajs6f>awoods: Maybe you can just put the WAR on the classpath. It is a JAR, after all.
<awoods>ajs6f: And by "putting the war on the classpath" you simple mean bundling the war into an executable one-jar utility?09:20
ajs6f: And by "putting the war on the classpath" you simply mean bundling the war into an executable one-jar utility?09:21
<ajs6f>awoods: No, that would miss the point. It's the _real_ repo war I need, because we have foolishly put out configuration into the classpath.
<awoods>ajs6f: how do you plan on executing your utility?09:22
<ajs6f>awoods: java -jar or a script with that. What else is there? I'm not going to ask people to install Scala for this.
<awoods>ajs6f: exactly. And typically, the jar your create will have all of its dependencies bundled into it.09:23
<ajs6f>awoods: Yes. So?
<awoods>ajs6f: So by "putting the war on the classpath", I assume you mean bundling the war into an executable one-jar utility?09:25
ajs6f: or are you thinking something different?
<ajs6f>awoods: This is not a dependency. It is _configuration_. I should be able to simply let the user use a CLI flag to point at the repo config of interest, but I can't do that because _we shoved the config into SPring XML and shoved THAT into the classpath_.09:26
awoods: We packed config into a container that should only hold code.
(Or immutable resources.)
<awoods>ajs6f: If you only need the spring.xml, it would probably be best to copy the appropriate file(s) into your project.09:27
<ajs6f>awoods: UUUURGGG. I do not need the Spring XML we pack with the artifact. I need the version of it that is actually _in use_. If people change the translation chain, I have to use the same one they are actually using, or I will mung their repo.09:28
<awoods>ajs6f: So you plan on allowing the user of this utility to pass in a path to their deployed war?09:29
<ajs6f>awoods: Something like that. But that's not going to be enough, because suppose they have an unpacked folder from which they are running the app (very common). Then the Spring XML in question will be found inside that.09:31
awoods: Maybe it would better to just write long detailed instructions about how to c+p the appropriate XML into a config file for this utility. This is really gross.09:32
<awoods>ajs6f: Can you detail again why this is an awful scenario for considering the REST API?09:33
<ajs6f>awoods: Because the whole purpose of this exercise is speedy bulk loading. 1) HTTP will require protocol. 2) It should not normally be used for multi-resource operations unless there is no alternative. 3) Fault handling in HTTP is awful for this sort of thing. The user will get no feedback until everything is done and all of the feedback has to be coagulated into a single response.09:36
<awoods>ajs6f: I was taking a slightly different view: This utility (against the REST API) would offer an easy way to performance test F4 with large sets of real RDF. We need to know where F4 capacity limits are! It would also act as an example for users of how resources can be created from existing RDF. I do not think there are many users in the wild that have bulk RDF sets at this point.09:41
<ajs6f>awoods: Now I don't understand at all what the point of this is.09:42
awoods: What is wrong with the current performance testing scripts?09:44
<awoods>ajs6f: they are not real.09:45
ajs6f: which ones do you have in mind?
<ajs6f>awoods: Ah. So fasseg and the folks at VA Tech who did that performance testing were just drug-addled? Or lying?
<awoods>ajs6f: the benchtool turned out to just be wrong (i.e. buggy in its reporting). VA-Tech only did cluster testing.
awoods: Then let's fix what's wrong with what we have. I thought this was supposed to me a user-facing load tool.09:47
<awoods>ajs6f: Fleshing out the Grinder tool makes sense. So does trying to load some real data, like: http://www.theeuropeanlibrary.org/tel4/access/data/lod09:48
<ajs6f>awoods: Do you want to edit that ticket to so say?09:49
Or should I?
<awoods>ajs6f: Feel free.
<awoods>ajs6f: The performance work was the one piece from the Technical Working Group that was never completed.
<ajs6f>awoods: You may wish to advertise that. Institutions that cannot contribute dev time may contribute time for that.
<awoods>ajs6f: My understanding of Grinder is that it is good at creating test requests. Do you think it could effectively be used to ingest an existing dataset?09:52
<ajs6f>awoods: I do not know. Why don't you ask Dan Davis on-list?09:53
afk bbl09:54
* ajs6f leaves
* jrgriffiniii leaves09:57
* ajs6f joins10:02
<awoods>welcome back10:03
<ajs6f>to the tar pit
<awoods>of love?
<ajs6f>Isn't that a B52s song? Tar Pit of Love?
Should be.10:04
* dwilcox leaves10:15
* awead joins10:21
* dwilcox joins
* awead leaves10:36
* dwilcox_ joins10:37
* dwilcox leaves10:46
* ajs6f leaves11:12
* ajs6f joins11:16
* tolloid leaves11:20
* tolloid joins11:24
* tolloid leaves11:30
* osmandin leaves
* tolloid joins
* ksclarke joins11:54
* dwilcox_ leaves11:55
* dwilcox joins12:01
* chadmills joins12:13
* osmandin joins13:06
* dwilcox leaves13:08
* dwilcox joins13:10
<ajs6f>awoods: Do you know who Wei Shi is?14:06
<awoods>ajs6f: I do not
<ajs6f>awoods: Something strikes me as odd about what he is doing.
<awoods>ajs6f: Maybe someone needs to introduce him to wordpress?14:07
<awoods>chadmills: how are things?14:11
<chadmills>awoods: Sorry. Pretty good. Fiscal year just ended so that's a good thing, about to start tinkering soon. How about you?14:18
<awoods>chadmills: not bad on this side... enjoying not being on travel and getting in some keyboard time.14:19
<chadmills>awoods: Does feel good. I am finishing up some grant business as well. Less conference rooms, calls, etc. the better for a while.14:21
<awoods>ajs6f: your java8 refactoring is really great.14:23
* ajs6f blushes prettily
awoods: I14:27
will send a PR for the auth module in a bit14:28
<awoods>ajs6f: thanks
* github-ff joins14:40
[fcrepo-module-auth-xacml] ajs6f created Cleanup (+1 new commit): http://git.io/vtbYK
fcrepo-module-auth-xacml/Cleanup 245d6ce ajs6f: Cleanup into Java 8
* github-ff leaves
* github-ff joins14:42
[fcrepo-module-auth-xacml] ajs6f opened pull request #36: Cleanup into Java 8 (master...Cleanup) http://git.io/vtbOF
* github-ff leaves
* travis-ci joins15:00
fcrepo4/fcrepo-module-auth-xacml#65 (Cleanup - 245d6ce : ajs6f): The build has errored.
Change view : https://github.com/fcrepo4/fcrepo-module-auth-xacml/commit/245d6cea87b4
Build details : https://travis-ci.org/fcrepo4/fcrepo-module-auth-xacml/builds/69180646
* travis-ci leaves
* dwilcox leaves15:07
* ksclarke leaves15:18
* osmandin leaves15:56
* dwilcox joins16:01
* github-ff joins
[fcrepo4] awoods closed pull request #822: Java8 with eggs (master...Java8WithEggs) http://git.io/vt3Hv
* github-ff leaves
* github-ff joins16:02
[fcrepo-module-auth-xacml] awoods pushed 1 new commit to master: http://git.io/vtb5i
fcrepo-module-auth-xacml/master 377491a ajs6f: Cleanup into Java 8...
* github-ff leaves
* github-ff joins
[fcrepo-module-auth-xacml] awoods deleted Cleanup at 245d6ce: http://git.io/vtb5F
* github-ff leaves
* travis-ci joins16:08
fcrepo4/fcrepo-module-auth-xacml#67 (master - 377491a : ajs6f): The build has errored.
Change view : https://github.com/fcrepo4/fcrepo-module-auth-xacml/compare/29a2c16c15de...377491a4d852
Build details : https://travis-ci.org/fcrepo4/fcrepo-module-auth-xacml/builds/69190726
* travis-ci leaves
* travis-ci joins16:30
fcrepo4/fcrepo4#3789 (master - ac3cb12 : Andrew Woods): The build passed.
Change view : https://github.com/fcrepo4/fcrepo4/compare/1a4cf678bf4a...ac3cb12f3da2
Build details : https://travis-ci.org/fcrepo4/fcrepo4/builds/69190679
* travis-ci leaves
* dwilcox leaves
* ajs6f leaves17:18
* github-ff joins18:36
[fcrepo4-swordserver] awoods opened pull request #7: Add SCM section to pom.xml (master...fcrepo-1558) http://git.io/vtNda
* github-ff leaves
* github-ff joins
[fcrepo-audit] awoods opened pull request #15: Add SCM section to pom.xml (master...fcrepo-1558) http://git.io/vtNdD
* github-ff leaves
* github-ff joins18:37
[fcrepo4-oaiprovider] awoods opened pull request #19: Add SCM section to pom.xml (master...fcrepo-1558) http://git.io/vtNd9
* github-ff leaves
* ksclarke joins19:12
* ksclarke leaves19:33
* ksclarke joins19:47
* ferret__ joins20:37
* ferret__ leaves20:46
* github-ff joins21:07
[fcrepo4] awoods opened pull request #827: Eliminate the publishing of UUID triples (master...fcrepo-1539) http://git.io/vtAad
* github-ff leaves
* ksclarke leaves21:32
* ksclarke joins21:50
* ksclarke leaves21:53
* ksclarke joins21:57
* ksclarke leaves22:32
* ksclarke joins22:39
* ksclarke leaves22:43
* ksclarke joins23:03
* ksclarke leaves23:05
* ff_logger leaves23:22
* f4jenkins leaves23:23
* ff_logger__ leaves
* f4jenkins joins
* ff_logger_ joins23:42
* ff_logger__ joins23:43
* tolloid leaves23:56

Generated by Sualtam