| kfox1111 | docaedo: any further progress on the json infra patch? | 03:20 |
|---|---|---|
| kfox1111 | We've got a coulpe of outstanding additions to the catalog that probably need to wait until the patch goes through. | 03:21 |
| kfox1111 | I also have 3 patches to the app catalog waiting. | 03:21 |
| kfox1111 | also want to make another pypi release to do more work on rpms/debs which is blocked too. | 03:22 |
| kfox1111 | docaedo: gota head out... thanks for looking into it. | 03:42 |
| *** kfox1111 is now known as kfox1111_away | 03:45 | |
| *** ig0r_ has quit IRC | 04:59 | |
| *** ig0r_ has joined #openstack-app-catalog | 05:01 | |
| *** kzaitsev_mb has joined #openstack-app-catalog | 09:18 | |
| *** kzaitsev_mb has quit IRC | 09:24 | |
| *** kzaitsev_mb has joined #openstack-app-catalog | 09:26 | |
| *** ig0r__ has joined #openstack-app-catalog | 09:38 | |
| *** ig0r_ has quit IRC | 09:39 | |
| *** kzaitsev_mb has quit IRC | 09:47 | |
| *** kzaitsev_mb has joined #openstack-app-catalog | 10:44 | |
| *** ig0r_ has joined #openstack-app-catalog | 12:04 | |
| *** ig0r__ has quit IRC | 12:05 | |
| *** kzaitsev_mb has quit IRC | 13:06 | |
| *** kzaitsev_mb has joined #openstack-app-catalog | 14:27 | |
| *** kzaitsev_mb has quit IRC | 14:32 | |
| kfox1111_away | docaedo: you around? | 16:22 |
| *** kfox1111_away is now known as kfox1111 | 16:22 | |
| docaedo | kfox1111: I am | 16:22 |
| kfox1111 | plan b... | 16:22 |
| kfox1111 | I'm updating the app catalog to have a tool to refresh the assets json file if changed, and create it if it doesnt exist. | 16:23 |
| kfox1111 | if we could change the infra patch to just exec it, that might be easier to approve, and gives us the ability to change things more easily in the future. | 16:23 |
| kfox1111 | also, we need to update the CORS patch to work in /api/v1/assets | 16:23 |
| docaedo | oh yeah I forgot about the cors patch, I will update that and make it depend on the infra json patch | 16:24 |
| docaedo | but I can get it pushed through either today or tomorrow (the one I put it) | 16:24 |
| docaedo | how would the tool you're talking about work? What would trigger it? | 16:25 |
| openstackgerrit | Kevin Fox proposed stackforge/apps-catalog: Break out the code to update assets. https://review.openstack.org/218670 | 16:25 |
| kfox1111 | just exec tools/update_assets.sh. No triggers or anything. just always call it. | 16:25 |
| kfox1111 | it will update things only if they changed. | 16:25 |
| docaedo | "just always call it" - you mean every puppet run? | 16:26 |
| kfox1111 | yes. | 16:27 |
| kfox1111 | then we can take care of the logic on our end. | 16:27 |
| kfox1111 | if we add or move a file, then we just update the shell script in our repo rather then having to tweak infra. | 16:28 |
| docaedo | I can almost guarantee that won't fly, infra wants all things that happen on their servers to be controlled by their servers. landing a patch like this (that would then let us do literally anything we want later by changing the shell script would get a lot of pushback, because they would say "why can't you do this in puppet, on infra?" | 16:31 |
| kfox1111 | I'd push back and say its a build script. | 16:32 |
| docaedo | what is? the shell script is 218670? | 16:33 |
| kfox1111 | most projects are allowed to control how their stuff is built. | 16:33 |
| kfox1111 | yea. 218670 | 16:33 |
| docaedo | yes true, but projects that are hosted on infra are controlled through puppet scripts in infra repo | 16:34 |
| docaedo | thing is, infra would have zero control over what that script changes to later | 16:35 |
| * kfox1111 shrugs. | 16:35 | |
| docaedo | so right now, it's a build script | 16:35 |
| kfox1111 | they're being slow to fix things. so I'm suggesting a faster way in the futre. if htey want to commit to faster reviews, then we can keep doing it the other way. | 16:36 |
| docaedo | some day later, I go crazy and turn it into a back door that let's me log in to that server and start running portscans, or mining bitcoins :) | 16:36 |
| kfox1111 | that could be true if we were shipping a foo.py that was a cgi, | 16:36 |
| kfox1111 | and we updated that. | 16:36 |
| kfox1111 | at the end of the day, if its running code from another sourcd, they have to trust somewhat, and remediate if the actor goes bad. | 16:37 |
| docaedo | yes, true, there are other cases where they trust external source and risk someone doing something bad, but in this case trying to switch to this method would guarantee a whole lot more scrutiny, so would not further the cause | 16:38 |
| docaedo | instead, I'll keep the conversation going on the infra channel today, I promise, and will rattle all the cages I can | 16:39 |
| kfox1111 | k. | 16:39 |
| docaedo | at this point we will get a +2 from core as soon as nibs responds (should be today) | 16:40 |
| docaedo | in the mean time I'll look at the other stuff that's outstanding today and give notes or plusses | 16:41 |
| kfox1111 | k. I still won't withdraw the update patch though. it makes it so that it doesn't rebuild assets if it hasn't changed, so its still a useful thing even if infra doesn't call it. | 16:42 |
| kfox1111 | if you can start the cors one soon though, that would be great. I'm expecting that one wont be very controvertial. | 16:43 |
| kfox1111 | with the way reviews have been going lately, I'm not so postive about the assets one getting resolved today. :/ | 16:44 |
| docaedo | the way the infra puppet one works right now, the "build json file" is subscribed to the repo check job, so the only time it runs is if the repo is updated, so I think it's pretty close to only rebuilding assets when they change | 16:49 |
| docaedo | well, not exactly that I guess - rebuilds the assets file upon commit anyway | 16:50 |
| kfox1111 | hmm.. k. | 16:51 |
| kfox1111 | not sure what nib's concern was then. | 16:54 |
| kfox1111 | still think, the moment we want to enhance things a bit, we're going to be right back here again. | 16:56 |
| kfox1111 | like, when we add initial stars support, we're going to probably want to merge stars stored in a db back into the generated assets file. | 16:57 |
| kfox1111 | so the puppet job cant just be what it is, and the update code should probably live in our repo. | 16:57 |
| docaedo | nibs concern was misplaced, he realized it after we chatted on IRC but then did not get back to remove the -1 unfortunately | 16:58 |
| kfox1111 | ah. ok. | 16:58 |
| docaedo | but I think adding stars/rating etc. is going to be a much bigger thing that I want to put up in concert with a lot of other new stuff | 16:59 |
| docaedo | so .. no matter what, it's going to need/get lets of review and looking at before it goes in | 16:59 |
| docaedo | this week I will put together that etherpad page I promised (I'll shoot for doing that tomorrow) | 17:00 |
| docaedo | where we can outline all the functions/features we want the site to have | 17:00 |
| kfox1111 | kk. | 17:01 |
| docaedo | and in the same place we can start noting what frameworks would solve this, and balance something like django against writing from scratch (or else something in between, lighter/smaller than django, but still giving some common methods for the things we want the site to do) | 17:01 |
| kfox1111 | The more I play with angular, the more I'm for it. | 17:02 |
| kfox1111 | I'm kind of hoping we can share most of the code between the website and the plugin. | 17:02 |
| docaedo | big reason is that I fear if we start doing it in really small steps (like DB backend for holding stars for ratings), we will end up stacking up tons of little bits on top of some simple initial decisions, and likely end up with a horror show that we have to start from fresh with anyway | 17:02 |
| docaedo | sharing code between plugin and website for the presentation layer would be AOK with me | 17:03 |
| kfox1111 | yeah, could be. though often you kind of have to assume your going to throw the first one away anyway (mythical man month), so that might not be too bad. | 17:04 |
| kfox1111 | I'm kind of thinking maybe we just do pecan for the rest api like most of the other services use, | 17:10 |
| kfox1111 | and keep doing the website staatic on the server side, and use angular to do the dynamic bits. | 17:10 |
| docaedo | not a bad idea, and ++++ on sticking with pecan as that seems to be the favorite | 17:14 |
| *** kzaitsev_mb has joined #openstack-app-catalog | 17:30 | |
| kfox1111 | kzaitsev_mb: I think the murano job that's failing is not due to the patch? | 17:34 |
| kzaitsev_mb | kfox1111: CI guys were meddling with the servers last week — adding 2d job | 17:36 |
| kzaitsev_mb | looks like that might be related | 17:36 |
| kfox1111 | ah. ok. | 17:36 |
| kfox1111 | any more movement on the other patch? it seems to have quited down. :/ | 17:37 |
| kzaitsev_mb | on the other hand my test empty patch was fine | 17:37 |
| kzaitsev_mb | https://review.openstack.org/#/c/218592/ | 17:37 |
| kfox1111 | still think it might make it in in time? | 17:37 |
| kzaitsev_mb | still think so, yes | 17:37 |
| kfox1111 | I think its something slightly more subtile... | 17:37 |
| kfox1111 | I'm adding a javascript file, so the cache needs rebuilding. | 17:37 |
| kfox1111 | and its not automatically rebuilding the cache. | 17:37 |
| kfox1111 | so so long as you don't add a javascript file, it works. | 17:37 |
| kfox1111 | I'm guessing if we disabled the inclusion of the javascript file, it would pass. | 17:41 |
| *** kzaitsev_mb has quit IRC | 19:34 | |
| *** kzaitsev_mb has joined #openstack-app-catalog | 19:38 | |
| *** kzaitsev_mb has quit IRC | 19:48 | |
| *** kzaitsev_mb has joined #openstack-app-catalog | 20:04 | |
| *** kzaitsev_mb has quit IRC | 20:23 | |
| *** kzaitsev_mb has joined #openstack-app-catalog | 20:40 | |
| *** kzaitsev_mb has quit IRC | 20:55 | |
| *** kzaitsev_mb has joined #openstack-app-catalog | 21:22 | |
| *** kzaitsev_mb has quit IRC | 22:38 | |
| *** kzaitsev_mb has joined #openstack-app-catalog | 22:44 | |
| *** kzaitsev_mb has quit IRC | 22:50 | |
| *** kzaitsev_mb has joined #openstack-app-catalog | 22:56 | |
| *** kzaitsev_mb has quit IRC | 23:22 | |
| *** kzaitsev_mb has joined #openstack-app-catalog | 23:54 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!