16:00:45 #startmeeting Horizon 16:00:46 Meeting started Tue Feb 11 16:00:45 2014 UTC and is due to finish in 60 minutes. The chair is david-lyle. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:47 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:00:49 The meeting name has been set to 'horizon' 16:00:57 Hello Horizon folks 16:01:03 who's around? 16:01:03 Hi o/ 16:01:05 hi all 16:01:05 hey 16:01:13 hello 16:01:19 Hello! o/ 16:01:22 hey folks 16:01:34 I was checking openstack-meeting and waiting :-D 16:02:43 hi all 16:02:44 I don't have any general announcements other than a reminder of the feature freeze on Feb 18, we'll dig into i-3 later 16:02:50 Hello 16:03:03 #topic Less vs. Sass 16:03:22 I think this is a no match 16:03:36 MaxV: no match? 16:03:47 MaxV, waht do you mean? 16:03:52 Sass provides much more features, 16:04:01 it's easier to integrate 16:04:09 the syntax is quite the same 16:04:41 we do not use the feature specific to less in horizon.less 16:04:50 so teh translation is easy 16:04:58 My worry is that this is a cosmetic change and skirts the real issue of needing a better CSS and JS toolchain 16:05:14 and we can get out of our dead end on bootstrapV3 16:05:40 we still ultimately need node in the toolchain at some point 16:05:46 for me the real question is do we use a pythonic compiler or use the official gem 16:06:01 david-lyle, /me thinks the same, I still feel trapped in the toolchain 16:06:26 Are the python tools for compiling sass more mature / robust than lesscpy? 16:06:36 and I'd vote against introduction of ruby etc. etc for solving the same things 16:06:43 that's why I'm against this solution ;) 16:06:53 jpich, are there python tools for sass? 16:06:59 I mean really? 16:07:11 hi all 16:07:29 I know it is not perfect but it's still better than node 16:07:35 storyboard has a good story for pulling in node at build time, but the story around a deployed server is still lacking 16:07:42 mrunge: I don't know, I think there were some mentioned in the email? The issue with bootstrap v3 was due to lesscpy so I'm trying to understand if we have a better python toolchain to deal with it if moving to sass 16:07:45 o/ hey all 16:07:58 jpich: agree with you 16:08:00 but maybe I misunderstood why we want to move to sass if that's not the case 16:09:22 I vote for anything that works :-) And I like Ruby, so.. :-D 16:10:28 with the separation between horizon and dashboard maybe we can use multiple languages 16:10:59 and keep a clean pythonic solution where there will be no javascript nor css 16:11:21 MaxV: yep 16:11:37 maybe a stupid question: why do we use less or sass at all? 16:11:37 I think we will want find a way to bring node in for the JS anyway then we are no longer blocked by bootstrap etc 16:11:47 MaxV: that makes sense 16:11:48 if that causes us that pita? 16:12:05 mrunge: +1 16:12:06 mrunge: well, CSS is old you know :-) 16:12:10 mrunge: certainly nicer to write CSS in than CSS 16:12:38 mrunge: pure css is the straight way to hell, less/sass makes it much much more easier to do styling 16:12:43 yes, but sass and less gets translated into css. 16:12:53 easier and cleaner 16:12:56 david-lyle, one more time, what is the problem with node? 16:13:08 well, proposing to introduce that huge dep chain is the same... 16:13:08 david-lyle: so it is either Node or SASS right? 16:13:26 enykeev, no-one sane wants node on a server 16:13:45 it is for development not for serving pages 16:13:48 and they are still pretty good in breaking compatibility 16:13:56 mrunge, but that is current web development 16:14:25 david-lyle, yes. but that doesn't mean, I have to like it... 16:14:54 hehe 16:14:59 mrunge agreed, but it may not be a matter of want 16:14:59 so, it's basically about who likes what? =) 16:15:22 well... 16:15:24 enykeev, supporting in a stable distribution is the main concern 16:15:35 with node we can get out of selenium which is not a real open source solution 16:15:56 so in another hand ... 16:16:06 MaxV, honestly? I don't see that right now 16:16:11 yeah 16:16:16 * jpich will be bringing more Selenium in the integration tests 16:16:17 I have a POC for this 16:16:41 MaxV, does your poc render pages? 16:16:42 but I think MaxV pointed out the best reasons for Sass anyway 16:16:42 webdriver + phantomjs + protractor (angular testing framework) 16:16:46 lsmola: I don't think they are mutually exclusive, but we seem to be sidestepping the larger issue by moving to Sass 16:16:48 * lsmola wrote one selenium test 16:17:04 mrunge: phantomjs allows you to make screenshots 16:17:17 in an automatic way 16:17:27 and trying to deliver a stable release makes me hessitant to keep replacing all of the core technologies we build on 16:17:37 and I think there is a binding in protractor 16:17:45 and anyone that has extended Horizon will have to move to Sass, which is a burden 16:18:07 MaxV, what about the webdriver? is that such a bundle like in selenium? 16:18:07 david-lyle: when do you expect the node.js appear again? 16:18:18 if yes, it's basically the same pain 16:18:38 mrunge: I have to check but I think it's a specification to allow remote connection to a browser 16:18:59 mrunge: It's not a software 16:19:07 jcoufal, I think if when we split out the JS and CSS, around then 16:19:17 MaxV, yes, and that part of selenium is the bundled binary in selenium 16:19:43 we could make it a build time only dependency and require CSS additions be written in CSS only 16:19:54 david-lyle: I just hope it is soon enough for us to update to Bootstrap 3 and not to be running the old version for the whole next release cycle 16:19:59 mrunge: let me take some additionnal informations 16:20:19 MaxV, agreed. Let's look at that e.g tomorrow in the morning 16:20:30 and deliver the compiled LESS along with the source in the horizon-web package 16:21:04 sigh, sorry guys, I need to leave right now. See you later. 16:21:35 david-lyle: then you will actually force everybody with extensions to Horizon to do it the same way... 16:22:07 or they can choose to use put node on their servers 16:22:41 david-lyle: well I understand the concern of switching to sass less then month before feature freeze 16:22:59 so what if we bring this issue to the summit and discuss it properly there? 16:23:03 I think this is a much larger issue than we can solve right here, but my concern is moving to sass is not a real solution 16:23:11 just a bandaid 16:23:40 I think the summit is the correct place to work through this 16:24:19 +1 for summit discussion 16:24:34 Regardless, I would like to hold off on Sass until post Icehouse, if we want to move that way 16:24:51 we are 7 days from feature freeze 16:25:55 #topic JavaScript requirements 16:26:50 mrunge is no longer present but it concerns the bundling of javascript libraries in Horizon 16:27:23 the idea is to bundle javascript in django packages 16:27:26 I think the summary here is to deliver the JS in external python packages of JS rather than static copies of the JS with Horizon 16:27:29 like this one https://bitbucket.org/massimilianoravelli/django-jquery 16:28:50 maybe a blueprint can be written 16:28:55 I think this makes sense 16:29:07 david-lyle - that's a great goal. How does the JS get into python packages? Do we need JS library authors to contribute them there? 16:29:24 the idea is to maintain our packages 16:29:26 MaxV, there is one: https://blueprints.launchpad.net/horizon/+spec/remove-javascript-bundling 16:29:42 I think we'll need to pin versions and qualify version updates before bumping the upper bounds 16:29:51 it is just changing the js and make a flag 16:30:01 so it is easy to maintain 16:30:06 so we can package the external js into python packages? 16:30:12 yes 16:30:17 okay got it. Thx! 16:30:26 with django binding 16:30:27 several already exist 16:30:46 MaxV: so only the external JS will move to that JS package? 16:30:55 yes 16:30:57 david-lyle, which ones? 16:30:58 libraries 16:31:09 external package per library 16:31:16 Sounds good to me, manage dependencies more clearly, like for our python deps 16:31:26 we are not necessarily the maintainers either 16:31:26 MaxV: cool, I like that. 16:31:31 hopefully we are not 16:31:52 then the Horizon specific JS and CSS will end up in a package that we do maintain 16:32:09 but that would not bundle jquery, angularjs, etc 16:32:12 david-lyle: some exists but if we can have the control of the versions and avoid third part code 16:32:23 MaxV 16:32:25 sure 16:32:50 I think this is a great goal to work toward 16:33:04 it is not really different than changing js by hand in the folder 16:33:10 but it's cleaner 16:33:26 much cleaner :) 16:33:28 MaxV agreed 16:33:42 I don't see a downside 16:33:59 let's make it happen 16:34:02 how should we deal with outdated packages like Angular? 16:34:30 +1 on moving to external JS package 16:34:36 enykeev? outdated how? our version 16:34:39 outdated? 16:34:51 version = static copy 16:34:59 we just have to make a flag 16:35:00 i mean like https://pypi.python.org/pypi/django-angularjs/1.0.2 from BP 16:35:20 we will maitain our package 16:35:28 the package should be published to pypi and versioned as it is updated 16:35:35 ^ 16:36:12 we can then pin the version and this should make validating updates a lot easier 16:37:01 anyone opposed? 16:37:28 Is this something we are going to attempt for icehouse? 16:37:48 I think we need to wait for Juno, just for stability 16:37:56 super. 16:38:22 there is some infrastructure set up, if we are going to publish these packages via the openstack bild infra 16:38:42 the plus side of that is some things are automated 16:38:53 +1 to external JS package 16:39:04 Doing it in icehouse seemed risky to me. I was just thinking through any concerns I might have about it. Seems like a great idea. 16:39:10 like all tags get published 16:39:42 #topic Icehouse-3 16:39:53 https://launchpad.net/horizon/+milestone/icehouse-3 16:40:24 Lots of blueprints have code that need review 16:40:50 I think the Heat items are actually up for review as well 16:40:58 and bug fixes too :p 16:40:58 just not put in the correct stat 16:41:00 e 16:41:07 but yes.. will help out reviews 16:41:47 I have one of my own out and some more in the line :) 16:41:55 I've been working on django 1.6, but we have about 30 broken tests when running against 1.6 16:41:58 thanks for the quick review and approve yesterday! 16:42:02 so I have some more work to do there 16:42:19 so there will be several small changes there 16:43:00 any concerns re: i-3 targets? 16:43:31 david-lyle: https://blueprints.launchpad.net/horizon/+spec/project-overview-page-ceilometer can be moved out 16:43:33 david-lyle: aside from test failing, any other issues running django 1.6? 16:44:08 david-lyle: the ceilometer work is targeted to I3, but it probably won't be in 7 days, so I will have no time to implement it here 16:44:28 lsmola: moved 16:45:01 lcheng_: I haven't actually made it that far yet :o 16:45:24 just ran tests and started scrambling 16:45:43 as that should really make icehouse 16:46:31 any other items for i-3? 16:46:39 david-lyle: okay :-) Let me know if you need on that. 16:46:40 https://blueprints.launchpad.net/horizon/+spec/launch-instance-ux-enhancement 16:47:06 we'd like to have feedback or approval for this bp, if possible 16:47:23 david-lyle: this sits there for a while and it should be ready https://review.openstack.org/#/c/64448/ 16:47:32 david-lyle: sorry, did i screw up the blueprint status on some of the heat items? 16:47:38 david-lyle: there are 3 up for review right now 16:48:07 Just make sure the blueprint status matches reality 16:49:00 tshritman: I'll take a look. Certainly any improvement is needed 16:49:06 david-lyle: I got the heat rbac up for review. I see a couple of heat patches that will be impacted (need to add rbac stuff as well). Would be nice if we can get the heat rbac reviewed soon, so that the pending heat patches can be updated to include rbac too. 16:49:23 jpich: i marked the bps as "needs review" 16:49:38 jomara: sounds correct to me :) 16:49:44 jomara: the heat patches might be yours :-) 16:49:44 so regarding the BP I brought up last week - https://blueprints.launchpad.net/horizon/+spec/neutron-subnet-mode-support 16:49:56 I've updated the whiteboard and work items 16:50:03 lcheng_, good point, we just need to track what order they merge 16:50:12 and make the appropriate changes 16:50:18 david-lyle: thanks, the work that have been done on wizard and that is being done on angular is helpful for this bp, and we are working on a test implementation on our side, so any feedback is useful :) 16:50:27 the neutron code is still in review.. but on the Horizon side it really should be a pretty simple and straightforward implementation 16:50:51 jomara: I'll review the pending heat patches today, and add notes as to what to put for the rbac stuff. 16:50:52 the subnet workflows needs to be updated to include two new IPv6 fields 16:51:00 lcheng_: thanks! 16:51:11 absubram: may warrant a FFE, but let's see how fast the neutron changes get in 16:52:20 FFE? 16:52:23 david-lyle: ok.. what's FFE? I'm proceeding to implement and test in my private setup with the neutron code.. so that I can be ready on my end as soon as neutron is taken carte of 16:52:28 doug-fish: Feature Freeze Exception 16:52:32 thx! 16:52:38 jpich: thanks! 16:53:37 #topic Open 16:53:54 not a lot of time left, what else do people want to discuss 16:54:12 Maybe a general clarification on dates since it can get confusing: for Icehouse, code proposal deadline is Feb 18th (code for a blueprint should be up for review on gerrit), Feature Freeze is March 4th (no more blueprints approved unless exception explicitly granted) and bugs fixes are welcome until the first RC on March 27th ( https://wiki.openstack.org/wiki/Icehouse_Release_Schedule ) 16:54:36 jpich: thanks 16:55:01 currently working on reimplementing all the js libraries in angular, I will repatch jasmine testing for jasmine 2.0 16:55:02 Hi all, I'm applying as a student to GSoC and I'm willing to work on a feature for Horizon 16:55:06 I'm looking for mentorship 16:55:16 jpich: what happens to code for blueprints approved between feb18 and mar4? 16:55:17 If someone is able to, please let me know 16:55:27 but I will not maitain old testing 16:55:45 vkmc: do you have a link to the requirements? 16:56:03 jomara: If the code was up on gerrit before the 18th, should be good, there should be plenty of times for reviews 16:56:29 jpich: oh, so its for when theres already code but no blueprint? 16:56:49 This is one the features I had in mind https://blueprints.launchpad.net/horizon/+spec/improve-user-feedback-and-error-handling, I've seen a lot of bugs related to it 16:57:02 jomara: the first freeze is an attempt to stabilize for the end of cycle gate mess 16:57:07 https://wiki.openstack.org/wiki/GSoC2014 - note that OpenStack is applying as an organisation but we haven't been accepted into GSoC yet (and don't know if we will, our application was declined twice before) 16:57:09 And, about GSoC requirements, we are gathering all the information here https://wiki.openstack.org/wiki/GSoC2014 16:57:28 Thanks jpich 16:58:00 jomara: If the blueprint is targeted to icehouse-3 but there is no sign of the code before the 18th, the blueprint will be moved to Juno and the code won't be approved for this cycle 16:58:23 of course there's always some wiggle room / flexibility but that's the idea 16:58:27 gotcha 16:58:49 yeah.. good to know.. :) 16:59:14 ...that's my understanding anyhow :) 17:00:07 jpich: mine too :P 17:00:23 I guess I should have said "no more code for blueprints (as in, no more features) approved in gerrit", sorry for adding to the confusion when I was hoping to reduce it 17:00:25 jpich: mrunge just told me today we were accepted 17:00:26 david-lyle: Good to know :P 17:00:37 jpich: we as the Openstack 17:00:52 lsmola: excellent 17:00:57 lsmola: The deadline for organisations to apply is later this week, I don't think Google has annoucned anything yet 17:01:06 People have volunteered to admin just last week 17:01:29 lsmola, The deadline for organizations to apply is Feb 14, currently we don't know if we will get accepted 17:01:33 * david-lyle watching tennis, back and forth 17:01:35 jpich: hm ok :-) 17:01:48 Accepted list will be announced on the 24th - http://www.google-melange.com/gsoc/events/google/gsoc2014 17:02:01 jpich: he was looking for the mentor, telling me we was accepted :-), now I am confuse 17:02:28 Looks like we're over time. Thanks everyone! Let's review the code that's up, and get the rest up for review. Have a great week! 17:02:29 lsmola, Probably he misunderstood 17:02:33 thanks david-lyle 17:02:34 #endmeeting