| *** openstack has joined #openstack-app-catalog | 15:46 | |
| *** rhagarty has joined #openstack-app-catalog | 16:27 | |
| kfox1111 | morning | 16:34 |
|---|---|---|
| kfox1111 | docaedo: good catch. :) | 16:34 |
| kfox1111 | docaedo: you around? | 16:47 |
| kfox1111 | https://bugs.launchpad.net/horizon/+bug/1503396 since you have devstack, can you please have a look at that? | 16:48 |
| openstack | Launchpad bug 1503396 in OpenStack Dashboard (Horizon) "glance wizard form var passing broken" [Undecided,New] | 16:48 |
| *** kzaitsev_mb has quit IRC | 17:30 | |
| docaedo | kfox1111: hi, yes, in and out for a bit but around generally - should be able to take a look at that bug today too | 17:59 |
| *** reed has quit IRC | 18:24 | |
| *** reed has joined #openstack-app-catalog | 18:28 | |
| *** hogepodge has joined #openstack-app-catalog | 18:37 | |
| kfox1111 | cool. thanks. | 18:59 |
| docaedo | no prob - I just added a note to the bug with my local.conf, not sure if there's anything else I can do to help with it? | 19:09 |
| docaedo | (and I did verify it works with devstack as-is) | 19:09 |
| docaedo | I suspect he had a working app-catalog-ui but was looking for the murano part | 19:10 |
| docaedo | Of course right after I unstack the environment I look at the bug again and see "Could someone who can reproduce this problem please take a screenshot of the Javascript console with the error that's causing the problem?" | 19:40 |
| docaedo | so re-stacking, and will try to trigger that error | 19:40 |
| docaedo | Pretty sure this is the error - at least it's the only one I see popping up when trying to populate the glance import: http://paste.openstack.org/show/480420/ | 19:53 |
| *** kzaitsev_mb has joined #openstack-app-catalog | 20:01 | |
| *** kzaitsev_mb has quit IRC | 20:27 | |
| kfox1111 | looking... | 20:33 |
| kfox1111 | not enough info. can you please switch DEBUG=true and try again? | 20:34 |
| kfox1111 | I'm kind of surprised devstack isn't defaulting to that. | 20:35 |
| docaedo | kfox1111: DEBUG=true where? | 20:35 |
| kfox1111 | horizon config. | 20:35 |
| docaedo | 'k, will find the horizon screen and config, and try that :) | 20:36 |
| kfox1111 | k. thx. | 20:37 |
| docaedo | np, perfect timing, should only take me a minute | 20:37 |
| docaedo | so changed DEBUG to true in /opt/stack/horizon/openstack_dashboard/settings.py and bounced apache .. lots of DEBUG lines on devstack console now, but nothing that looks like it would help here | 20:44 |
| docaedo | errors we need to see would be on javascript console right? | 20:44 |
| kfox1111 | it should change the javascript code to not be something like: /static/dashboard/js/6e84d1677058.js | 20:45 |
| kfox1111 | but instead a bunch of little js files with better names. | 20:45 |
| kfox1111 | oh.... unless there is an explicit COMPRESS_ENABLED=true somewhere... | 20:45 |
| kfox1111 | or whatever the exact setting is. | 20:45 |
| docaedo | ah yeah I wonder/suspect .. | 20:49 |
| docaedo | probably need to do more than just bounce apache | 20:50 |
| docaedo | COMPRESS_ENABLED = True in settings.py | 20:50 |
| kfox1111 | ah. there it is. | 20:50 |
| kfox1111 | try disabling it. | 20:51 |
| docaedo | 'k | 20:52 |
| docaedo | ok compress is turned off, debug is on, but not sure what to look for | 21:06 |
| kfox1111 | what does the javascript console look like now? | 21:11 |
| docaedo | heh, looks like a big ugly mess! | 21:19 |
| docaedo | will try to pastebin it if I can grab everything | 21:20 |
| kfox1111 | k. thanks. :) | 21:21 |
| docaedo | I suspect this will not be easily read - http://paste.openstack.org/show/480429/ | 21:22 |
| kfox1111 | thats pretty odd looking... | 21:24 |
| kfox1111 | can you see if you get any errors if the plugin's disabled? | 21:24 |
| kfox1111 | seems like there should be more to it then that too... | 21:25 |
| kfox1111 | never seen an error in just angular.js before, without some other piece of code being in the stack trace too. | 21:25 |
| kfox1111 | should be able to set the plugin to disabled in the enabled directory and restarting the horizon service. | 21:26 |
| docaedo | ok that error comes up when rendering the catalog page - no errors come up after that when clicking on a glance image and then clicking import | 21:28 |
| kfox1111 | k. I'm not expecting any errors from that activity. I never saw errors in the past from horizon when data failed to pass to the modals. | 21:29 |
| kfox1111 | in the past, they were just passed via the links and the ajax-modal class on the link. | 21:30 |
| kfox1111 | horzion on the other side would grab the data and then fill in the field in the modal. | 21:30 |
| kfox1111 | but for some reason, it failed to work after they patched horizon. I think I referenced the patch. | 21:31 |
| kfox1111 | So, they either need to revert that patch, or figure out why the patch had the effect. :/ | 21:31 |
| kfox1111 | also sounds like that js error message is unrelated. | 21:32 |
| docaedo | yeah I think the JS error doesn't have anything to do with it (but was the only error I was seeing) | 21:32 |
| docaedo | maybe Richard will be able to get the plugin running in his devstack now though and be able to make sense of it? I looked at the patch you referenced and didn't see any obvious reason why this broke (BUT I don't grok the horizon bits well enough for that opinion to be worth anything) | 21:34 |
| kfox1111 | I suspect it is because they reordered when event's were happening in angular land. | 21:37 |
| kfox1111 | the events fire off too early now I think and the code to copy in the field values happen before they are initally populated. | 21:38 |
| kfox1111 | but I've not tested that theory at all. | 21:38 |
| kfox1111 | can you specify a commit to use for horizon for devstack? | 21:41 |
| docaedo | um, yes, in theory. I've had mixed luck getting that right TBH but I can try | 21:44 |
| kfox1111 | if it would work, it might be a quick way to show commit X works, commit X+1 fails... | 21:45 |
| kfox1111 | rather then try and prove it plus all of the changes horizon has kept making since. | 21:45 |
| docaedo | sure .. lemme see if I can just manually grab that without redoing the whole stack/unstack thing | 21:48 |
| docaedo | oh wait, duh, I don't want to check out the one the broke it :) | 21:50 |
| kfox1111 | it would be interesting from the standpoint of knowing for sure if it spits out that javascript error. (I don't remember it doing that) | 21:51 |
| kfox1111 | but yeah, the one right before too. :) | 21:51 |
| docaedo | I'm missing something obvious (poor sleep, not enough caffeine?) if the parent was a merge, how do I find/specify that? | 21:58 |
| kfox1111 | git log to show the history, then get the change id from the previous one? | 21:58 |
| kfox1111 | bbiab. | 21:58 |
| docaedo | ah, my confusion was that I was looking for a specific patch/review, but several previews commits are from jenkins | 22:05 |
| docaedo | well all I have confirmed so far is that checking out previous commits just breaks the catalog (catalog pane is empty), then checkout master and restart apache, all is well | 22:13 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!