| [11:14:51] | * gusaus has quit (Quit: gusaus) |
| [11:19:01] | * christefano has quit (Quit: christefano) |
| [12:15:27] | <bgm> | offtopic, but pretty good video of the new features: https://about.gitlab.com/2016/12/22/gitlab-8-15-released/ |
| [13:47:36] | * shaneonabike1 has quit (Quit: Leaving.) |
| [14:31:16] | * gusaus has joined #aegir |
| [14:32:55] | * gusaus has quit (Client Quit) |
| [14:48:29] | * hefring has joined #aegir |
| [15:36:58] | * gusaus has quit (Quit: gusaus) |
| [15:40:18] | * gusaus has joined #aegir |
| [17:19:52] | * gusaus has quit (Quit: gusaus) |
| [17:23:48] | * gusaus has joined #aegir |
| [18:24:03] | * gusaus has quit (Quit: gusaus) |
| [19:06:13] | * boshtian has joined #aegir |
| [19:22:55] | * ybabel has joined #aegir |
| [23:04:26] | * hestenet has quit (Ping timeout: 256 seconds) |
| [01:15:02] | <ergonlog2c> | bgm, very cool |
| [02:20:09] | <ergonlog2c> | bgm: I believe https://www.drupal.org/node/2794645 was mostly blocked by wanting better tests. Unless I'm mistaken, I believe you've done that, no? Can we revisit the patch from #8? The one from #11 could just move to the api.php file as an example. |
| [02:20:09] | <hefring> | https://www.drupal.org/node/2794645 => Hardcode cache paths in civicrm.settings.php [#2794645] => 12 comments, 3 IRC mentions |
| [02:35:54] | * gusaus has joined #aegir |
| [02:59:16] | * boshtian has quit (Quit: boshtian) |
| [03:01:03] | <bgm> | ergonlog2c: one possible issue comes to mind: on a client (farm) we change the uploadPath. would there be a way to override? |
| [03:07:59] | <ergonlog2c> | bgm: you mean to override an existing hook implementation of the template alter? |
| [03:09:22] | <ergonlog2c> | Drush doesn't have a "weight" concept wrt hooks, so it just goes alphabetically. |
| [03:09:46] | <ergonlog2c> | so, you could implement that same hook, and further alter the template, assuming it was named correctly |
| [03:11:29] | <ergonlog2c> | actually, I don'tt think you'd need to alter the template... just the provided vars |
| [03:11:47] | <ergonlog2c> | since we're already enabling the uploadDir override |
| [03:11:56] | <bgm> | oh ok, cool |
| [03:12:35] | <ergonlog2c> | so it should just be hook_civicrm_settings_parameters_alter() with a name that comes later alphabetically |
| [03:12:58] | <ergonlog2c> | mind you, I haven't tested that scenario specifically |
| [03:13:14] | <ergonlog2c> | but that's how it ought to work |
| [03:13:34] | <bgm> | if you can PR on github to trigger tests, we can merge on pass |
| [03:13:49] | <ergonlog2c> | the $params array is passed around as a reference |
| [03:13:50] | <bgm> | otherwise will have a look when back at my computer |
| [03:14:03] | <ergonlog2c> | ok, will do. |
| [03:14:14] | <ergonlog2c> | I'm working on some of this right now anyway :) |
| [03:16:35] | <bgm> | cool thx |
| [03:23:36] | <ergonlog2c> | bgm: PR against https://github.com/mlutfy/hosting_civicrm? |
| [03:31:39] | <bgm> | ergonlog2c: yep |
| [03:59:31] | <ergonlog2c> | bgm: where's you CI running? |
| [04:03:13] | <colan> | bgm: Any chance you've run into https://gitlab.com/aegir/hosting_https/issues/24 yet and know what's going on? |
| [04:12:37] | <bgm> | ergonlog2c: ci.symbiotic.coop - it can be flaky though |
| [04:14:53] | <bgm> | colan: hmm, I think I may have encountered it, but couldn't find the cause. I had a colleague complain that "everything had disappeared" until he played with roles, which may have triggered a cache flush |
| [04:15:55] | <bgm> | re:jenkins, to clarify, it can be flaky because i'm not doing a full aegir environment reset. |
| [04:16:01] | <ergonlog2c> | as per my comment in that issue, if the view's ajax call is still going over http, that would fail, resulting in an empty response |
| [04:16:58] | <colan> | would certainly explain why "drush @hm uli" is still giving me "http://..." |
| [04:17:06] | <ergonlog2c> | yep |
| [04:17:46] | <ergonlog2c> | I've seen JS and CSS warn or fail, since Drupal still tries to server them over http too |
| [04:18:19] | <ergonlog2c> | iirc, I usually just hardcode the https:// URL into a local.settings.php, or something |
| [04:19:08] | <ergonlog2c> | but since we're already working to polish up aegir_https, we might want to add something like that to settings.php |
| [04:19:16] | <ergonlog2c> | as the local override is brittle |
| [04:19:24] | <ergonlog2c> | breaks on clones, for example |
| [04:37:53] | <colan> | as per my new comment there, it's more complicated (if that's the problem), as it didn't work on its own. |
| [04:41:05] | <colan> | ergonlog2c: maybe we should set the $base_url to https://... ? Just worried that would cause other problems. |
| [04:57:00] | * fatguylaughing_ has joined #aegir |
| [05:12:01] | * fatguylaughing_ has quit (Quit: fatguylaughing_) |
| [05:16:00] | <ergonlog2c> | colan: that was my suggestion |
| [05:16:34] | <ergonlog2c> | but yes, it will could cause a number of issues |
| [05:16:40] | <colan> | oic |
| [05:16:42] | <colan> | sorry |
| [05:17:22] | <ergonlog2c> | if it's in local.settings.php, it'll break when you clone, as the clone will think it's at the wrong url |
| [05:17:41] | <ergonlog2c> | it also might not play nice with aliases |
| [05:18:46] | <ergonlog2c> | but it sounds like the issue is *not* related to https |
| [05:19:23] | <ergonlog2c> | in which case it's likely something introduced relatively recently, since I'm not seing that behaviour on any of my aegirs |
| [05:24:51] | <colan> | right |
| [07:22:03] | <bgm> | what kind of issues would an https base_url cause? |
| [07:28:51] | * ybabel has quit (Quit: ybabel) |
| [07:42:51] | <colan> | bgm: we'd have to make sure it's set properly for all possible sites. breaks on clones according to ergonlog2c |
| [07:43:29] | <colan> | but that's not actually the problem there. |
| [07:44:22] | <bgm> | hmm ok |
| [10:21:19] | * cweagans has joined #aegir |
| [10:21:29] | <cweagans> | Hey folks, long time no see :) |
