| [11:58:28] | <mstenta[m]> | farmOS weekly dev call starting soon - all are welcome! https://meet.jit.si/farmos-dev |
| [13:08:19] | <paul121[m]> | thx for reviewing timeline stuff mstenta symbioquine ! |
| [13:09:09] | <paul121[m]> | I want to ping the maintainer and see if they can get a new release so we can clean up at least one of those things... |
| [13:09:15] | <symbioquine[m]> | The minimal NGINX config that sets the headers correctly is pretty small - wouldn't be surprising if you have it set mstenta;... (full message at <https://matrix.org/_matrix/media/v3/download/matrix.org/ysTMRpLeUGDGsSgv...) |
| [13:10:24] | <symbioquine[m]> | (I recommend also applying this strategy to make it a bit more reliable: https://sandro-keil.de/blog/let-nginx-start-if-upstream-host-is-unavaila...) |
| [13:10:35] | <paul121[m]> | Briefly mentioned, but very curious in having a much more abstracted API for logs and handling their popups... eg timeline.addLog. Maybe noodle on it and we revisit later :-) |
| [13:10:38] | <mstenta[m]> | symbioquine: Yes - looking at it now, I do have that |
| [13:10:59] | <mstenta[m]> | So I'll have to retest to see what went wrong earlier for me with http |
| [13:12:03] | <mstenta[m]> | symbioquine: Oh! I have that on my list for Farmier hosting to look into (using `set $upstream`) |
| [13:12:08] | <mstenta[m]> | thanks for the link |
| [13:12:23] | <mstenta[m]> | i've run into that issue before |
| [13:13:39] | <paul121[m]> | DDEV configures https locally for me. I'm curious how hard it would be to make a local instance under a subpath to test this all |
| [13:13:55] | <symbioquine[m]> | So my actual configuration for each service is more like this;... (full message at <https://matrix.org/_matrix/media/v3/download/matrix.org/yiXSFWokpDYyWQmn...) |
| [13:13:56] | <mstenta[m]> | Oh that would be great to have automated tests for |
| [13:14:28] | <mstenta[m]> | Might be tricky to test everything though, because a lot of it is client side testing I think |
| [13:14:35] | <symbioquine[m]> | paul121[m]: https://github.com/symbioquine/farm-faux-cloud hosts instances under subpaths |
| [13:14:44] | <mstenta[m]> | (juts based on my experience with bugs related to subdirectory installations) |
| [13:14:54] | <symbioquine[m]> | I've been building some tests on top of that. |
| [13:15:16] | <mstenta[m]> | (they aren't caught by our current kernel/functional tests i don't think, in other words) |
| [13:16:00] | <symbioquine[m]> | https://github.com/symbioquine/farmOS_asset_link/blob/95484b590e9ab5140a... |
| [13:17:24] | <symbioquine[m]> | symbioquine[m]: TL;DR; I'm testing that JSON:API works correctly against those ephemeral farmOS instances (which requires that - at least for JSON:API - the generated paths are correct with subdirectory installations) |
| [13:20:49] | <mstenta[m]> | Yea a lot of the bugs I've fixed (in the past) with regard to subdirectory installations have been issues with custom JavaScript code that wasn't properly using Drupal's base path, and was hard coded to assume a non-subdirectory install |
| [13:21:06] | <mstenta[m]> | And we don't have a lot of test coverage for JS in farmOS right now |
| [13:21:10] | <mstenta[m]> | (if any?) |
| [13:25:31] | <symbioquine[m]> | mstenta[m]: I haven't tried spinning up farm-faux-cloud as part of a GH action, but it should be possible. |
| [13:26:01] | <mstenta[m]> | Ah yea! Should be |
| [13:26:24] | <mstenta[m]> | And technically it is possible to test some client side stuff via PHP (BrowserTestBase) |
| [13:27:20] | <paul121[m]> | So it sounds like we should be pretty safe using non-absolute URLs when rendered server side using methods like Url::fromRoute(...)? |
| [13:40:39] | <paul121[m]> | I was curious what happens when you render a relative path in text, like for a status message and want to "copy" the link. A relative link wouldn't be too useful. But looks like the browser expands it to be an absolute link which makes sense :-) |
| [13:45:00] | <mstenta[m]> | paul121: I think that's correct (re: server side URLs) |
| [13:45:31] | <mstenta[m]> | Ah yea... makes sense that browsers expand it too |
| [14:13:44] | <mstenta[m]> | @room Anyone want to provide one more approval for https://github.com/farmOS/farmOS-community-blog/pull/28 so we can publish it? 😁 |
| [14:34:09] | <mstenta[m]> | paul121: FYI now that other issues are resolved, I'm not seeing the `http` vs `https` issue I saw earlier, so `setAbsolute()` seems to be working fine after all |
| [15:58:04] | <symbioquine[m]> | In this context I can't really think of an argument for doing it one way or the other except that `setAbsolute` makes the code (and payload) slightly larger. |
| [16:40:08] | <paul121[m]> | Yeah, I think I'm convinced it isn't necessary! |