Page tree
Skip to end of metadata
Go to start of metadata

Developers Meeting on Weds, April 13, 2016


Today's Meeting Times


Ongoing Reminders

Discussion Topics

  1. DSpace 6.0 Status
    1. Proposed 6.0 Timeline:
      1. Release Candidate 1 (RC1): ~April 21
      2. Testathon : April 25 through May 6 (two weeks)
      3. RC2: May 13
      4. Final Release: Around May 26
    2. Current 6.0 status:
      1. One outstanding feature PR:  (Final decision this week)
      2. Only two outstanding "Blocker" bugs:
        1. One of which is delayed until RC2.
      3. Eight (8) outstanding "Improvement" PRs:
      4. Six (6) outstanding "code task" PRs:
    3. All the PRs flagged for 6.0 (some may need rescheduling):
  2. Other topics?

Meeting Notes

Meeting Transcript

  • Partial IRC Transcript is available at -  (Unfortunately DuraLogBot crashed before the meeting).

  • Full IRC Transcript (from Tim) is pasted below (all times are in CDT, so 3:00PM = 20:00 UTC):

    (3:00:48 PM) tdonohue: Hi all, it's time for our weekly DSpace DevMtg. Agenda for today at:
    (3:00:49 PM) kompewter: [ DevMtg 2016-04-13 - DSpace - DuraSpace Wiki ] -
    (3:01:52 PM) tdonohue: The main topic for today to resolve is 6.0 release timelines. I'm getting quite worried that if we don't move forward with Testathon quickly, we may risk not releasing 6.0 until *after* OR16 (and we really should strive to release prior to OR16)
    (3:03:04 PM) roelandatmire [] entered the room.
    (3:03:30 PM) tdonohue: I've proposed a possible timeline at the top of the agenda... I think this seems doable, but will require us to move *rapidly* towards an RC1 (and finalize whether our one outstanding feature is ready or not -- we'll get to that shortly)
    (3:04:32 PM) roelandatmire left the room (quit: *.net *.split).
    (3:04:33 PM) tdonohue: Before we get into the discussion about that last feature, I want to see if this timeline *seems* reasonable to others (assuming we can meet the first deadline of RC1 by April 21, i.e. one week)
    (3:04:34 PM) rdillen is now known as roelandatmire
    (3:05:01 PM) mhwood: What do we want to accomplish by the RC1 deadline?
    (3:05:32 PM) tdonohue: As you'll also notice, if we *don't hit* that first deadline (April 21), it seems less likely that we'll see a Final Release prior to early June (OR16), which is unfortunate
    (3:06:26 PM) tdonohue: mhwood: for an RC1, we need all features finalized, testathon scheduled & announced, and all blockers closed (or resolutions in progress that we can clearly document to testers) 
    (3:07:05 PM) tdonohue: (Ideally, we'd also get many/most code tasks & improvements decided upon as well...but, minor ones might still be acceptable post-RC1)
    (3:08:02 PM) tdonohue: (and I believe *most* of our remaining improvements / code tasks are minor...and don't really significantly affect database or configs)
    (3:08:59 PM) tdonohue: So, in summary, RC1 is really supposed to be a "mostly ready" 6.0... It should have all our features, be semi-stable, and we should *hopefully* have "mostly finalized" database migrations / configurations.
    (3:09:33 PM) hpottinger: I think the 21st is do-able
    (3:09:47 PM) tdonohue: I, personally, think we are nearly to RC1. But, the big question mark is the final outstanding feature (which we need to finalize this week)
    (3:10:28 PM) tdonohue: But, if anyone *disagrees* with that analysis, it'd be good to know soon, so that we can tackle any other remaining major tickets/PRs rapidly
    (3:10:34 PM) mhwood: At the moment I see one blocker, not worrying.
    (3:11:04 PM) hpottinger: DS-3125 is brand new, awaiting a volunteer
    (3:11:05 PM) kompewter: [ ] - [DS-3125] Submitters cannot delete bistreams of workspaceitems - DuraSpace JIRA
    (3:11:27 PM) luizsan_ [d178abe3@gateway/web/freenode/ip.] entered the room.
    (3:11:35 PM) tdonohue: yes, 3125 is a brand new blocker. It does need resolution before RC1, but seems "fixable" quickly. Just needs a volunteer
    (3:11:41 PM) hpottinger: wow, this meeting has a few attendees...
    (3:11:49 PM) KevinVdV: I’ll take DS-3125
    (3:11:50 PM) kompewter: [ ] - [DS-3125] Submitters cannot delete bistreams of workspaceitems - DuraSpace JIRA
    (3:12:01 PM) tdonohue: Thanks KevinVdV!  That'd be wonderful
    (3:12:01 PM) KevinVdV: Clearly some sort of migration issue
    (3:12:18 PM) tdonohue: yes, it seems like an issue between 5.x API and new Service (6.x) API
    (3:12:33 PM) hpottinger: thanks, KevinVdV, call out for testers when you have a PR?
    (3:12:55 PM) KevinVdV: Will do, shouldn’t be to hard (I hope)
    (3:13:39 PM) tdonohue: So, now that the remaining blocker has a volunteer (and we'll work on getting testers ASAP too)...we can move along to the other major ticket / PR... DSPR#1162 / DS-2877
    (3:13:41 PM) kompewter: [ ] - [DS-2877] Import of ScienceDirect metadata including embargo and linking to or embedding of the final version - DuraSpace JIRA
    (3:13:45 PM) kompewter: [ ] - DS-2877 Import of ScienceDirect metadata including embargo and linking to or embedding of the final version by LetitiaMukherjee
    (3:15:00 PM) tdonohue: As mentioned on our Committer list, I had several major concerns with the PR.  @mire staff though have responded to the most significant ones, and offered other plausible options for 6.x.  We can go through those more today (if we felt it worthwhile), but I also wonder what folks think about this PR in general for 6.0?
    (3:16:02 PM) hpottinger: I'd like to hear about these plausible options?
    (3:17:11 PM) hpottinger: looking at my email
    (3:17:14 PM) tdonohue: hpottinger: have you caught up on the -commit email thread?  I'm talking about the response that we can do away with the "elsevier.*" schema...put everything into "dc" instead (not necessarily ideal but plausible enough for now).  Also the explanation for the usage of the new "workflow.fileaccess" fields.
    (3:18:59 PM) tdonohue: My personal opinion is none of this is's a shame we didn't have these discussions earlier, as we likely could have come up with better solutions.  But, we're now stuck with (mostly) what we have, and either have to "live with it" for 6.0, or say "sorry, we wish we could accept it, but we'd like to see this improved for 7.0 inclusion"
    (3:21:18 PM) mhwood: My immediate concern about the metadata remapping is:  can we figure out which is which after multiple fields are folded into one?  Even if it does require parsing the values all the time.
    (3:21:46 PM) tdonohue: To me, the new metadata schema was a "no go". But, that being resolved, the rest of the code is messy / less than ideal, but maybe something we could "live with".  At the same time, I do realize though we've spent a TON of time/effort on this one PR for so very long that it's frustrating that we still are talking about whether to include it 4 months after the initial review.
    (3:22:41 PM) tdonohue: mhwood: good question. Not sure I know whether we'll be able to easily tell all the new "dc.rights" or "dc.identifier" fields apart once they get into DSpace
    (3:22:43 PM) hpottinger: caught up on the e-mail thread...
    (3:23:39 PM) mhwood: What we need is a mapping to some namespace with more fine-grained concepts.  But we are out of time.
    (3:23:48 PM) tdonohue: mhwood++
    (3:23:53 PM) hpottinger: so... here's the process (as of now) new code needs 3 +1s, and no vetos, right?
    (3:24:48 PM) tdonohue: hpottinger: correct, that is the formalized voting process for features:
    (3:24:48 PM) kompewter: [ Developer Voting Procedures - DSpace - DuraSpace Wiki ] -
    (3:26:13 PM) tdonohue: I will admit, we often do simplify that to 2 +1 votes (and no major objections) when we review PRs (just to try and get non-controversial things in more quickly)
    (3:26:47 PM) hpottinger: we generally assume, in the case of a submission from a committer, that *they* will vote +1
    (3:26:48 PM) tdonohue: But this particular PR is obviously one that is at least semi-controversial (otherwise it would have been in long ago)
    (3:26:58 PM) tdonohue: hpottinger: correct
    (3:27:33 PM) hpottinger: I see no +1s on 1162 at the moment
    (3:29:17 PM) hpottinger: I don't want to be a stickler about the rules and all... but, that, in nutshell, is the issue with 1162
    (3:29:48 PM) tdonohue: So, our real options today are... (A) do those in attendance today wish to (individually or together) veto this PR (in which case discussion is really over, unless the veto is later withdrawn)?  or (B) do we call for a very rapid, formal vote of Committers to see if we can achieve 3 +1 votes, and no one else vetoes it?
    (3:30:47 PM) tdonohue: We also *do* have a few @mire staff members here (KevinVdV and roelandatmire), and I'm assuming they could answer questions or add clarifications.
    (3:31:45 PM) hpottinger: I'm not sure about my ability to support this feature if it were released, honestly.
    (3:32:29 PM) KevinVdV: I can only speak for myself but I wasn’t involved in this feature… but from what I understand if the configuration of the fields is a problem, this can easily be altered
    (3:32:52 PM) mhwood: Stupid idea:  dc.identifier entries could have text_lang = the namespace URI, so there is a fixed string that tells us what text_value means.  (Maybe authority instead -- I don't know much about that column.)  That doesn't help any of the other mappings, though.  I just had to get rid of this idea.
    (3:34:22 PM) tdonohue: My "gut" is that, at best, I'm really a 0 ("unsure" / on the fence).  I honestly appreciate all the hard work that @mire and Elsevier have put into this, and I *see* it's potential as a feature, but the code/implementation is less than ideal / messy, and seems like there's a lot left to figure out / clean up.
    (3:34:31 PM) mhwood: The problem is:  to what do we alter them?  Nobody really likes any of the specific choices already before us.
    (3:36:36 PM) tdonohue: right, the metadata issues are the core problem, but neither solution is "great" and it seems like we are stuck either going with something "messy-ish", or delaying until we can make proper metadata decisions (or changing the implementation to not pull down all this info, and just grab very *basics*)
    (3:36:40 PM) hpottinger: It's probably obvious, but I'm also a 0 at this point
    (3:38:18 PM) KevinVdV: Given the low attendence, perhaps a vote by mail would be better suited ? Then we can move on to other issues ?
    (3:38:31 PM) hpottinger: +1 mail list vote
    (3:40:57 PM) tdonohue: Ok, I'll call a mailing list vote after this meeting then. It's going to *have* to be a very tight turnaround though (unfortunately), otherwise I don't think we will achieve RC1 by next week. This PR either needs almost immediate merger, or it needs delaying
    (3:41:19 PM) tdonohue: So, I'll likely give folks the weekend, and close voting on Monday
    (3:42:45 PM) hpottinger: thanks, tdonohue!
    (3:43:42 PM) mhwood1 [] entered the room.
    (3:44:46 PM) tdonohue: Ok. so, other 6.0 topics of note... We *really* need to get Testathon scheduled ASAP (so people can block it off on their schedules and actually help test).  If we can achieve RC1 by next week, it is *plausible* to hold Testathon starting April 25.  Otherwise, we cannot really start it until early May.
    (3:44:47 PM) mhwood1: Argh, it looks like we lost the logbot
    (3:45:28 PM) tdonohue: yikes. seriously?  Ugh. what a bad day to lose the log bot. I'll paste in my logs after the meeting
    (3:46:13 PM) tdonohue: mhwood: We decided to call a mailing list vote after this meeting
    (3:46:31 PM) mhwood1: Thanks.  That was my position, in case it didn't come through.
    (3:47:04 PM) DuraLogBot [] entered the room.
    (3:47:04 PM) DuraLogBot: (notice) This channel is logged -
    (3:47:07 PM) mhwood left the room (quit: *.net *.split).
    (3:48:02 PM) tdonohue: So, regarding Testathon dates... any preference from folks on April 25 start vs May 2 start?  (April 25 is tight, but gives us more time before OR16)
    (3:48:35 PM) mhwood1: No preference.
    (3:48:37 PM) hpottinger: might as well say it's going to be April 25
    (3:49:15 PM) terry-b: I assume the risk of the earlier date would be the possibility of moving it?  Otherwise, earlier sounds better
    (3:50:00 PM) tdonohue: terry-b: correct. to achieve April 25, we *have* to release RC1 next week (no real excuses). While, I think that's doable, there's some risk there
    (3:51:18 PM) hpottinger: So... other than possibly bringing in 1162, we are frozen?
    (3:51:30 PM) tdonohue: So, I'm hearing April 25. So, I'll work on getting an announcement out this week. I'd announce for a two week testathon (April 25 through May 6)
    (3:52:24 PM) tdonohue: hpottinger: frozen for features, yes.  But, I still think improvements and code tasks (refactors) are welcome, provided they are minor and given sanity checks/tests
    (3:53:04 PM) tdonohue: And bug fixes are obviously *always* welcome. But, if there are any bug fixes to *database migrations*, we really should strive to get them in prior to RC1, if possible.
    (3:53:49 PM) tdonohue: Unfortunately, we are almost out of time for today, so we'll have to do work on these three categories offline (ad hoc). But, here's what remains in each category
    (3:54:11 PM) tdonohue: Improvements (most look minor at a glance):
    (3:54:12 PM) kompewter: [ Pull Requests · DSpace/DSpace · GitHub ] -
    (3:54:28 PM) tdonohue: Code Tasks / Refactors (again, most look minor):
    (3:54:29 PM) kompewter: [ Pull Requests · DSpace/DSpace · GitHub ] -
    (3:55:00 PM) tdonohue: Bug fixes (there's a lot left, we need to look for any that still involve database migrations):
    (3:55:01 PM) kompewter: [ Pull Requests · DSpace/DSpace · GitHub ] -
    (3:56:01 PM) tdonohue: I could *really* use help combing through these and testing / merging. Thanks to everyone who has already helped some, but we need help making this final push to RC1
    (3:56:36 PM) tdonohue: I will warn that any that are flagged "work in progress" need to either get that flag removed soon, or risk being delayed
    (3:58:08 PM) hpottinger: tdonohue: I am sure kshepherd reads these logs, if you have a list of PRs you need testing, I can't guarantee he'd make time to test, but... he's been in testing mode for a few days, it would be worth pinging him
    (3:58:34 PM) tdonohue: So, please, at a minimum, review PRs you've created (or are assigned to you). If possible, go through those lists above and claim a few to test
    (3:58:35 PM) KevinVdV: I need to run, until next week
    (3:58:36 PM) KevinVdV left the room (quit: Quit: KevinVdV).
    (3:59:04 PM) mhwood1: I have a list right here....
    (3:59:56 PM) tdonohue: hpottinger: unfortunately, the list of PRs is *all of the above links* (improvements, code tasks, bugs fixes).  I'm not finding myself having much time these days to review all these myself (with balancing 6.0 and new UI work, both of which need some resolution or decisions by OR16)
    (4:01:00 PM) tdonohue: And if these decisions start to come down to me... be aware I'm likely going to immediately delay anything marked "work in progress" or "merge conflict". I just cannot review all these myself and have to cull it down quickly
    (4:01:14 PM) mhwood1: Indeed.
    (4:02:22 PM) tdonohue: Ok, since we are now at the top of the hour, I'm going to close up today's meeting. In summary though (and for those reading the logs since DuraLogBot was rebooted)...
    (4:02:44 PM) tdonohue: 1. We will have a vote on the outstanding feature (PR 1162) of Committers. I will call it shortly on -commit list
    (4:03:20 PM) tdonohue: 2. The tentative timeline in the agenda will be announced later this week and we'll schedule Testathon for April 25:
    (4:03:21 PM) kompewter: [ DevMtg 2016-04-13 - DSpace - DuraSpace Wiki ] -
    (4:03:50 PM) tdonohue: 3. As part of #2, RC1 will be next week (no exceptions)
    (4:04:38 PM) tdonohue: 4. Finally, we could REALLY use help reviewing / merging PRs (especially improvements / code tasks / and major bug fixes) PRIOR to RC1.  Any help is welcome! If you find an important PR that needs more attention, let us know immediately
    (4:05:24 PM) tdonohue: That's it. I'll also get my IRC logs copied into the agenda so folks can read through the entirety of the logs
    (4:05:25 PM) hpottinger: and DS-3125 needs a volunteer
    (4:05:26 PM) kompewter: [ ] - [DS-3125] Submitters cannot delete bistreams of workspaceitems - DuraSpace JIRA
    (4:05:32 PM) hpottinger: (I think)
    (4:05:41 PM) tdonohue: hpottinger: KevinVdV volunteered today
    (4:05:48 PM) hpottinger: oh, wait, yeah, that's right, woohoo
    (4:05:51 PM) tdonohue: (and assigned it to himself)
    (4:06:11 PM) hpottinger: Last minute reminder: OR16 early bird registration ends *today*
    (4:06:20 PM) tdonohue: With that, I'm closing up the meeting. I have emails to send out to get this voting started, etc!
    (4:06:41 PM) tdonohue: (YES, do go register today for OR16, if you are going and want to save some $$)
    (4:07:27 PM) mhwood1: Looks like I *will* be going :-) , but *won't* get the administrative stuff done today. :-(
    (4:07:28 PM) hpottinger: official accommodations are all booked, but there are still plenty of alternatives
    (4:08:20 PM) hpottinger: Oh, that reminds me, my funding has also been approved, I'm going
    (4:09:11 PM) mhwood1: Good news.
    (4:09:37 PM) mhwood1: Must dash.
    (4:09:42 PM) mhwood1: 'bye all.
    (4:09:49 PM) mhwood1 left the room.
    (4:12:01 PM) hpottinger left the room (quit: Quit: Leaving, later taterz!).

Action Items

(Action items go here, if any)