14:31:26 <asmacdo> #startmeeting Pulp Triage 2019-05-03
14:31:27 <asmacdo> !start
14:31:27 <asmacdo> #info asmacdo has joined triage
14:31:27 <pulpbot> Meeting started Fri May  3 14:31:26 2019 UTC.  The chair is asmacdo. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:31:27 <pulpbot> Useful Commands: #action #agreed #help #info #idea #link #topic.
14:31:27 <pulpbot> The meeting name has been set to 'pulp_triage_2019-05-03'
14:31:27 <pulpbot> asmacdo: asmacdo has joined triage
14:31:32 <dalley> that solves the issue we're having with name overlaps on the content pointer, and it solves the "no bulk_create" problem
14:31:42 <asmacdo> dalley: lets come back to this at open floor
14:31:48 <dalley> downsides: change all the things :(
14:31:49 <dkliban> +1
14:32:08 <bmbouter> #info bmbouter has joined triage
14:32:08 <bmbouter> !here
14:32:08 <pulpbot> bmbouter: bmbouter has joined triage
14:32:17 <dkliban> #info dkliban has joined triage
14:32:17 <dkliban> !here
14:32:18 <pulpbot> dkliban: dkliban has joined triage
14:32:28 <dawalker> #info dawalker has joined triage
14:32:28 <dawalker> !here
14:32:28 <pulpbot> dawalker: dawalker has joined triage
14:32:51 <asmacdo> !next
14:32:52 <pulpbot> asmacdo: 8 issues left to triage: 3345, 3785, 4749, 4750, 4759, 4760, 4780, 4782
14:32:53 <asmacdo> #topic https://pulp.plan.io/issues/3345
14:32:53 <pulpbot> RM 3345 - amacdona@redhat.com - NEW - Add Exporter template
14:32:54 <pulpbot> https://pulp.plan.io/issues/3345
14:33:02 <asmacdo> #idea Proposed for #3345: change to story
14:33:02 <asmacdo> !propose other change to story
14:33:03 <pulpbot> asmacdo: Proposed for #3345: change to story
14:33:10 <dkliban> accept
14:33:15 <dkliban> and add to sprin
14:33:24 <asmacdo> do we need to add this to a sprint?
14:33:39 <dkliban> i want to keep plugin_template up to date
14:33:42 <asmacdo> i think we can leave this as a stub until a plugin needs an exporter
14:33:42 <bmbouter> yes
14:33:53 <bmbouter> nice tag use
14:34:22 <dkliban> i don't care too much ... let's just accept
14:34:28 <dkliban> and we'll use the tag to find it later
14:34:41 <asmacdo> #agreed change to story
14:34:41 <asmacdo> !accept
14:34:41 <pulpbot> asmacdo: Current proposal accepted: change to story
14:34:42 <asmacdo> #topic https://pulp.plan.io/issues/3785
14:34:42 <pulpbot> asmacdo: 7 issues left to triage: 3785, 4749, 4750, 4759, 4760, 4780, 4782
14:34:43 <pulpbot> RM 3785 - muattiyah - NEW - Document that plugin implementors will have to call super().validate() on their serializers.
14:34:44 <pulpbot> https://pulp.plan.io/issues/3785
14:35:47 <asmacdo> #idea Proposed for #3785: accept, add GA blocker, add to sprint
14:35:47 <asmacdo> !propose other accept, add GA blocker, add to sprint
14:35:47 <pulpbot> asmacdo: Proposed for #3785: accept, add GA blocker, add to sprint
14:36:13 <dkliban> +1
14:36:14 <ttereshc> #info ttereshc has joined triage
14:36:14 <ttereshc> !here
14:36:14 <pulpbot> ttereshc: ttereshc has joined triage
14:36:15 <mikedep333> #info mikedep333 has joined triage
14:36:15 <mikedep333> !here
14:36:16 <pulpbot> mikedep333: mikedep333 has joined triage
14:36:16 <ttereshc> +1
14:36:58 <asmacdo> #agreed accept, add GA blocker, add to sprint
14:36:58 <asmacdo> !accept
14:36:58 <pulpbot> asmacdo: Current proposal accepted: accept, add GA blocker, add to sprint
14:36:59 <asmacdo> #topic https://pulp.plan.io/issues/4749
14:36:59 <pulpbot> asmacdo: 6 issues left to triage: 4749, 4750, 4759, 4760, 4780, 4782
14:37:00 <pulpbot> RM 4749 - gmbnomis - NEW - "repository" field on Publication is broken
14:37:01 <pulpbot> https://pulp.plan.io/issues/4749
14:37:03 <ppicka> #info ppicka has joined triage
14:37:03 <ppicka> !here
14:37:03 <pulpbot> ppicka: ppicka has joined triage
14:37:40 <asmacdo> i was looking at this one-- i reporoduced wit hthe python plugin
14:38:15 <ipanova> #info ipanova has joined triage
14:38:15 <ipanova> !here
14:38:15 <pulpbot> ipanova: ipanova has joined triage
14:38:18 <ttereshc> accept and add to the sprint? GA blocker?
14:38:18 <bmbouter> we should accept and fix the url at least and then let the Master/Detail issue resolve it (I think)
14:38:26 <bmbouter> sure we can take action on it
14:38:43 <dkliban> +1
14:38:46 <asmacdo> bmbouter: publications are already master detail
14:38:58 <ipanova> +1
14:39:19 <dawalker> +1
14:39:27 <asmacdo> #idea Proposed for #4749: Leave the issue as-is, accepting its current state.
14:39:27 <asmacdo> !propose accept
14:39:27 <pulpbot> asmacdo: Proposed for #4749: Leave the issue as-is, accepting its current state.
14:39:59 <asmacdo> #agreed Leave the issue as-is, accepting its current state.
14:39:59 <asmacdo> !accept
14:39:59 <pulpbot> asmacdo: Current proposal accepted: Leave the issue as-is, accepting its current state.
14:40:00 <asmacdo> #topic https://pulp.plan.io/issues/4750
14:40:00 <pulpbot> asmacdo: 5 issues left to triage: 4750, 4759, 4760, 4780, 4782
14:40:01 <pulpbot> RM 4750 - kersom - NEW - Repository Version docs claims support for PATCH/PUT
14:40:02 <pulpbot> https://pulp.plan.io/issues/4750
14:40:25 <dkliban> accept and add to sprint
14:40:32 <dkliban> repo versions are immutable
14:40:34 <kersom> I wrote tests for this one already
14:40:47 <asmacdo> +1
14:40:49 <kersom> just a docs issue
14:40:57 <ppicka> +1
14:41:00 <dkliban> OpenAPI schema issue
14:41:08 <kersom> ok
14:41:19 <asmacdo> it would be nice if we could make this change before the rc2
14:41:21 <bmbouter> asmacdo: oh I thought it was the Distribution.repository field
14:41:30 <ttereshc> +1 to add to sprint
14:41:42 <asmacdo> #idea Proposed for #4750: accept and add to sprint
14:41:42 <asmacdo> !propose other accept and add to sprint
14:41:42 <pulpbot> asmacdo: Proposed for #4750: accept and add to sprint
14:42:17 <ipanova> +1
14:42:20 <asmacdo> dkliban: could you click triaged on this one, and add a comment that its a schema problem, not an API problem?
14:42:29 <dkliban> sure
14:42:30 <asmacdo> #agreed accept and add to sprint
14:42:30 <asmacdo> !accept
14:42:30 <pulpbot> asmacdo: Current proposal accepted: accept and add to sprint
14:42:32 <asmacdo> #topic https://pulp.plan.io/issues/4759
14:42:32 <pulpbot> asmacdo: 4 issues left to triage: 4759, 4760, 4780, 4782
14:42:33 <pulpbot> RM 4759 - kersom - NEW - REST API doc shows duplicate content for uploads endpoint
14:42:34 <pulpbot> https://pulp.plan.io/issues/4759
14:42:44 <dkliban> same thing here
14:42:49 <dkliban> OpenAPI schema problem
14:42:52 <dkliban> accept and add to sprint
14:42:59 <ttereshc> +1
14:43:00 <asmacdo> #idea Proposed for #4759: accept and add to sprint
14:43:00 <asmacdo> !propose other accept and add to sprint
14:43:00 <pulpbot> asmacdo: Proposed for #4759: accept and add to sprint
14:43:11 <asmacdo> #agreed accept and add to sprint
14:43:11 <asmacdo> !accept
14:43:11 <pulpbot> asmacdo: Current proposal accepted: accept and add to sprint
14:43:12 <asmacdo> #topic https://pulp.plan.io/issues/4760
14:43:12 <pulpbot> asmacdo: 3 issues left to triage: 4760, 4780, 4782
14:43:14 <pulpbot> RM 4760 - kersom - NEW - Discrepancy in the docs about HTTP verb used to start a chuncked upload
14:43:15 <pulpbot> https://pulp.plan.io/issues/4760
14:43:41 <asmacdo> so glad to see all these docs issues rolling in kersom, thanks so much!!!
14:43:43 <dkliban> i think this is the same category
14:44:00 <kersom> asmacdo, np.
14:44:04 <asmacdo> #idea Proposed for #4760: accept and add to sprint
14:44:04 <asmacdo> !propose other accept and add to sprint
14:44:04 <pulpbot> asmacdo: Proposed for #4760: accept and add to sprint
14:44:29 <kersom> I ask for docs so much, I have to use them.
14:44:45 <kersom> :P
14:44:48 <asmacdo> #agreed accept and add to sprint
14:44:48 <asmacdo> !accept
14:44:48 <pulpbot> asmacdo: Current proposal accepted: accept and add to sprint
14:44:49 <asmacdo> #topic https://pulp.plan.io/issues/4780
14:44:49 <pulpbot> asmacdo: 2 issues left to triage: 4780, 4782
14:44:50 <pulpbot> RM 4780 - rochacbruno - NEW - Task search by invalid field should raise meaningful error
14:44:51 <pulpbot> https://pulp.plan.io/issues/4780
14:44:55 <asmacdo> :)
14:45:43 <rochacbruno> I just opened this issue
14:45:54 <asmacdo> i dont remember if this is something we can fix or if this is a DRF issue
14:46:12 <asmacdo> but lets accept anyway, the proposed behavior is clearly better
14:46:16 <asmacdo> #idea Proposed for #4780: Leave the issue as-is, accepting its current state.
14:46:16 <asmacdo> !propose accept
14:46:16 <pulpbot> asmacdo: Proposed for #4780: Leave the issue as-is, accepting its current state.
14:46:38 <dkliban> accept
14:46:45 <ttereshc> works for me
14:46:49 <dkliban> i already commented that we need to use another project to handle that
14:46:53 <asmacdo> #agreed Leave the issue as-is, accepting its current state.
14:46:53 <asmacdo> !accept
14:46:53 <pulpbot> asmacdo: Current proposal accepted: Leave the issue as-is, accepting its current state.
14:46:53 <asmacdo> #topic https://pulp.plan.io/issues/4782
14:46:54 <pulpbot> asmacdo: 1 issues left to triage: 4782
14:46:55 <pulpbot> RM 4782 - amacdona@redhat.com - NEW - Plugin Writer guide mentions of publish task (publisher vs publication) are out of date
14:46:56 <pulpbot> https://pulp.plan.io/issues/4782
14:47:16 <asmacdo> dkliban: i don't see your comment on there
14:47:39 <dkliban> asmacdo: on the PR that he links at the top of the issue
14:48:16 <asmacdo> ah thanks
14:48:33 <asmacdo> #idea Proposed for #4782: accept and add to sprint
14:48:33 <asmacdo> !propose other accept and add to sprint
14:48:33 <pulpbot> asmacdo: Proposed for #4782: accept and add to sprint
14:48:48 <ttereshc> add docs tag
14:48:49 <dkliban> +1
14:48:55 <ppicka> +1 to accept
14:49:07 <asmacdo> lol cant believe i of all people forgot the docs tag
14:49:16 <ttereshc> haha
14:49:23 <asmacdo> #idea Proposed for #4782: accept, +documentation, add to sprint
14:49:23 <asmacdo> !propose other accept, +documentation, add to sprint
14:49:23 <pulpbot> asmacdo: Proposed for #4782: accept, +documentation, add to sprint
14:49:32 <ttereshc> you had too many of them recently :)
14:49:55 <asmacdo> #agreed accept, +documentation, add to sprint
14:49:55 <asmacdo> !accept
14:49:55 <pulpbot> asmacdo: Current proposal accepted: accept, +documentation, add to sprint
14:49:56 <pulpbot> asmacdo: No issues to triage.
14:50:01 <asmacdo> open floor
14:50:28 <ttereshc> asmacdo, you should add !openfloor command to pulpbot ;0
14:50:29 <ttereshc> ;)
14:50:39 <asmacdo> !friday
14:50:39 <pulpbot> ♪ It's Friday, Friday, gotta get down on Friday ♪
14:50:47 <dalley> lol that's new
14:50:47 <bmbouter> yeyah!
14:51:08 <bmbouter> so I'm about to start on this PR after a call to pulp-dev for final grooming https://pulp.plan.io/issues/4745#note-10
14:51:16 <bmbouter> https://pulp.plan.io/issues/4745 rather
14:51:56 <asmacdo> bmbouter: does this include master/detail distributions?
14:52:10 <bmbouter> it does not
14:52:27 <bmbouter> before triage we were discussing with gmbnomis and that will be done separately
14:53:11 <asmacdo> ok. i have some finer points im interested in, but they can wait for a PR
14:53:25 <asmacdo> i can groom this
14:53:54 <dkliban> dalley: to go back to your suggestion of using Abstract models instead of Master/Details
14:54:18 <dkliban> we introduced Master/Detail to handle multiple types of content being associated with a repository version
14:54:59 <dkliban> so you could do a single query to see what all content belongs to a repo version
14:55:02 <dalley> yeah, so, tbh it normally would be more of a Pulp 4 concept, but if we're talking about breaking things anyway
14:55:13 <dalley> dkliban, that's true, but there's two counterpoints
14:56:20 <dalley> 1: to my knowledge (could be wrong), Katello has no plans to mix content types in one repo, and it's something that we may even discourage because of how badly it interacts with mirror mode
14:56:52 <dkliban> dalley: what about RpmPackage and UpdateRecord
14:56:59 <dkliban> or Errata ... however we call it
14:57:41 <dalley> 2: for plugins that have multiple content types, it would be two+ queries instead of one, but that's not a huge deal
14:58:49 <dalley> especially when you consider what follows from that... which is that it would immediately enable bulk_create() which would give a huge speedup
14:59:47 <dkliban> yeah
15:00:04 <dkliban> so which stages can't use bulk create right now?
15:00:09 <dalley> the main issue though, is that it would be a huge change
15:00:27 <dalley> at that point the whole concept of "RC" would be thrown out the window
15:00:39 <dalley> dkliban, ContentAssociation
15:00:55 <dalley> er, creation
15:00:59 <dalley> not association
15:01:44 <dkliban> gotcha
15:01:56 <asmacdo> ok, so abstract models pros: 1) mixing types is weird 2) performance improvement
15:02:49 <asmacdo> im not sure that those 2 reasons are enough to set us back so much
15:02:55 <dalley> the con, every plugin would need to boilerplate a couple of things that they don't currently have to touch
15:03:04 <dalley> ContentArtifact, RepositoryVersion
15:03:10 <dalley> RepositoryContent
15:04:19 <dalley> it's a big con, admittedly
15:04:20 <asmacdo> can we shift to the master/detail distributions idea to compare?
15:04:47 <dalley> they wouldn't have to copy and paste, metaprogramming would get you 95% of the way there I think
15:04:50 <dalley> sure asmacdo
15:05:13 * asmacdo shivers at metaprogramming
15:06:04 <dalley> I only mention it at all because we're already breaking/changing a bunch of things
15:06:36 <dkliban> right now we have a BaseDistribution that has some fields and it's abstract
15:06:47 <asmacdo> totally fair. im glad you brought it up-- i doubt we will go that direction if we can avoid it, but its still good to put our options on the table IMO
15:07:32 <asmacdo> what are the weaknesses of distributions as they are now?
15:08:02 <asmacdo> for 1 (minor, IMO) plugins that need custom distributions have unexpected urls
15:08:49 <asmacdo> the bigger issue is that users of multiple plugins (especially if they are scripting) will have to do extra stuff for those plugins
15:09:20 <asmacdo> are there correctness issues as well dkliban?
15:09:31 <dkliban> asmacdo: not sure what you mean
15:09:50 <dkliban> i think the biggest problem with Distributions as they ar enow is that they are different from every other part of the plugin api
15:10:12 <asmacdo> so they work fine, but they are awkward
15:10:30 <dkliban> every other object in the plugin api is either concrete and used directly or is a Master that requires a detail  class
15:10:48 <dkliban> it's the only thing that is abstract
15:11:07 <dkliban> gmbnomis: is that why you want to make them master/detail?
15:17:50 <asmacdo> #endmeeting
15:17:50 <asmacdo> !end