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

Using timezone: Eastern Standard Time
* jcoyne leaves01:09
* kaarefc joins02:17
* kaarefc leaves02:24
* kaarefc joins02:25
* kaarefc leaves02:29
* kaarefc joins04:08
* eddies leaves04:49
* eddies joins06:00
* eddies leaves07:41
* eddies joins07:47
* eddies leaves
* eddies joins
* kaarefc leaves08:08
* VincentNG joins08:29
* jcoyne joins08:44
<VincentNG>anyone on IRC yet? how can i open a port on futures1?08:53
* jcoyne leaves09:08
* barmintor leaves09:19
* kaarefc joins09:26
* ap2972 joins09:34
* ap2972_ joins10:07
* ap2972_ leaves10:09
* ap2972 leaves10:16
* kaarefc leaves10:20
<eddies>vincentng: i'd ask anusha. send an email to ff-internal10:21
<VincentNG>ok, thanks
* ap2972 joins10:28
* jcoyne joins10:29
<eddies>https://plus.google.com/hangouts/_/46ed3fe3f2d8606cbc4c4056cc8bb4cce9176ebc?authuser=0&hl=en10:32
* barmintor joins10:43
I like the idea of leveraging GSoC against the modeshape committers :)
<cbeer>iRODS!10:46
"committed"10:59
<pivotal-bot>Chris Beer added "Provide patches to the Metrics project to make integration with our stack.. sane." https://www.pivotaltracker.com/story/show/4662880511:07
Edwin Shin added "Set up CNAME for sonar.fcrepo.org" https://www.pivotaltracker.com/story/show/4662966111:12
Edwin Shin started "Set up CNAME for sonar.fcrepo.org" https://www.pivotaltracker.com/story/show/46629661
Edwin Shin accepted "Add Metrics Servlet to kitchen-sink" https://www.pivotaltracker.com/story/show/4654283711:13
Aleksey Pokalyukhin started "Add automated code-quality checking" https://www.pivotaltracker.com/story/show/4654336711:32
Edwin Shin finished "Set up CNAME for sonar.fcrepo.org" https://www.pivotaltracker.com/story/show/4662966111:33
<eddies>^ vincentng: sonar.fcrepo.org CNAME is ready
<cbeer>VincentNG: you know how to do the proxy thing11:34
?
<VincentNG>cbeer: not off the top of my head.
<cbeer>ok, i'll do it. port 9000, right?11:35
<VincentNG>yup, thanks
also, can you create a database called: sonar with a sonar user?
<cbeer>oh, right. i remember now. they created these machines from some existing vm image with an existing password11:37
hm. somehow I got in before. oh well.11:40
nothing important there anyway
http://sonar.fcrepo.org/11:41
<VincentNG>is the mysql port open? i thought it was11:44
3306 right?
<cbeer>VincentNG: localhost.
* ap2972 leaves11:48
* ap2972 joins12:00
* kaarefc joins12:16
<pivotal-bot>Aleksey Pokalyukhin added comment: "I think the good cursory overview is provided in: ""12:21
http://www.sonarsource.org/what-makes-checkstyle-..." https://www.pivotaltracker.com/story/show/46543367
* VincentNG leaves12:42
* ap2972 leaves12:51
* ap2972 joins13:01
<barmintor>how would people feel about moving the repo size updates out of the API classes (there's a dup of it in fcrepo-http-api and legacy-api) and into the DatastreamService and ObjectService classes?13:09
<cbeer>+1
there's a trick though
well, two tricks.
1) in order to do the datastream size update (as we're doing it now), we have to save the session first.. and i think it's bad form to save the session deep in the stack.13:10
so, somehow, you need to get around that.
(oh, and the reason you have to save the session is we've decided object size means the size of the datastreams and their jcr properties, some of which are generated on save)13:11
(i'd be happy to decide datastreams only count their binary content)
<barmintor>yeah, I'm a lttle hesitant about the session save (though it is only one layer deep in the stack)
<cbeer>2) i don't think the pattern we have is that great in the first place. it'd be so much better to calculate the size from the index we have
i think jasondgi may have done some work on that front, but i'm not sure it was merged in/was performant13:12
<barmintor>hmm13:13
<cbeer>i think, on 1), if we decide binary size is a good enough measurement
we're storing the binary size as a property now (for fixity reasons)
we just need to make it indexed.. and figure out how to get the info we want back out of lucene
maybe that's a better thing for me to work on today13:15
<barmintor>another option would be to have the service proxies be the ones that handle loggin into the repo13:16
<pivotal-bot>Jonathan Green edited "Update islandora to point at new, tomcat location for fcrepo4" https://www.pivotaltracker.com/story/show/4614227513:17
<barmintor>ie, the transactional methods get a new session for their transaction
and the api classes just initiate transactions and serialize the results
<pivotal-bot>Jonathan Green edited "Update islandora to point at new, tomcat location for fcrepo4" https://www.pivotaltracker.com/story/show/46142275
<barmintor>maybe I'll for the unit-testing branch and hack that together to see what it would look like when I get over to CDRS13:18
<cbeer>ok. i'm not sold, but who knows.
* github-ff joins13:24
[fcrepo4] barmintor pushed 1 new commit to unit-testing: http://git.io/h6Jh0Q
fcrepo4/unit-testing 08322b8 Benjamin Armintor: example unit test, with a mocking helper to support continued JCR interactions at API level
* github-ff leaves
<bljenkins>Project fcrepo4 build #237: UNSTABLE in 3 min 49 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo4/237/13:27
* VincentNG joins13:30
<barmintor>ap2972 I'm headed over to CDRS. I assume I can work in the conference area in the back?13:32
<ap2972>will check the conference room
I think nobody use it now
<eddies>barmintor, cbeer: while i'm sure there are folks who might argue otherwise, i'm certainly happy limiting the size calculations to binary size (i.e. not including jcr properties)13:36
<cbeer>cool. then i'll try to familiarize myself with the jcr query syntax.. or the modeshape lucene index13:45
makes me long for a REPL console.
* ap2972 leaves13:47
* ap2972 joins13:49
* jcoyne leaves13:56
* jcoyne joins13:57
* barmintor_ joins14:07
eddies: Do you feel comfortable enough with that breakdown ap2972 linked to start moving forward with Sonar?14:14
<eddies>barmintor_: you mean to use pmd, checkstyle, findbugs in concert?14:15
<barmintor_>yeah
<eddies>yeah
<barmintor_>ok, so we need to have a supporting db and tomcat running on one fo the futures machines14:16
futures6 is where the jenkins build is running, right? Is there a reason not to put a sonar db on futures4 or futures5?14:21
<cbeer>what's the tomcat for?14:22
(or, i'm wondering if we should use gluck.cul instead.. if this is really CI machinery)
<barmintor_>sonar runs these code analysis tools, stores the results in some form in a Db, and then has a tomcat webapp for viewing the outputted reports14:23
the other option is to run the different tools separately, and try to collect HTML output where possible as github pages (I think eddies raised this as a possibility)14:24
that would work for PMD14:26
<cbeer>ok.. so.. sonar is a standalone app.. and rather than publishing through there, has a webapp component too?14:29
i hate java developers sometimes.
<barmintor_>a wise old sysadmin told me once that the only way to understand any software is to remember that the developers hate you.14:31
<cbeer>anyway. i might still vote gluck.cul
it has tomcat (obviously)
and isn't doing very much
and wouldn't be called on to be part of our fcrepo 4 cluster
<barmintor_>I think there's a question as to how much the product manager wants html artifacts as part of the deliverable for those tickets14:32
ap2972 is researching how to incorporate the tools as part of the jenkins build now
<cbeer>well.. here's the brute-force way to calculate repository size:14:55
https://gist.github.com/cbeer/73f0b810600b9d8262f6
i'd guess we can memoize that and recalculate as needed14:56
and worry about scaling later
<barmintor_>cbeer: is the existing jenkins build config viewable anywhere but gluck?15:11
<cbeer>no.15:12
<barmintor_>ok, thanks
<pivotal-bot>Vincent Nguyen added comment: "http://sonar.fcrepo.org/dashboard/index/1" https://www.pivotaltracker.com/story/show/4627457115:17
* jcoyne leaves15:42
* jcoyne joins15:43
* barmintor_ leaves15:53
* ap2972 leaves
* ap2972 joins15:59
* ap2972 leaves16:00
* kaarefc leaves16:25
* jcoyne leaves16:48
* VincentNG leaves16:49
* ap2972 joins20:21
* ap2972 leaves20:23
* jcoyne joins20:41
* ap2972 joins20:43
<pivotal-bot>Aleksey Pokalyukhin added comment: "There are Jenkins plugins for all of those tools including Cobertura. Trying to find how to use it w..." https://www.pivotaltracker.com/story/show/4654336721:06
* ap2972 leaves22:04