| [09:38:49] | <symbioquine[m]> | ACTION uploaded a video: (354KiB) < https://libera.ems.host/_matrix/media/v3/download/matrix.org/GkJWXefRbGx... > |
| [09:46:26] | <symbioquine[m]> | Explanation: Field A is already located in Field C which is already located in Field B. Attempting to move Field B to Field A thus throws a circular location constraint violation when the update is sent to the server. The update gets retried a few times then moved to an internal DLQ whence users can choose to retry or delete the update. |
| [09:48:19] | <symbioquine[m]> | Ideally, I should also enforce the constraint locally so that it works offline too. |
| [09:53:56] | <mstenta[m]> | Cool! |
| [09:54:18] | <mstenta[m]> | Video doesn't play on my phone but sounds great! :-) |
| [09:55:22] | <symbioquine[m]> | Oh, weird. Guess that's a downside to posting those in WebM vs GIF. |
| [09:55:57] | <mstenta[m]> | Your vids have worked in the past for me so maybe it's something on my end |
| [09:56:59] | <mstenta[m]> | I do see the first frame 🤷 |
| [10:53:24] | <mstenta[m]> | Sketched this up for consideration... wasn't too hard with the new datetime form element: https://github.com/farmOS/farmOS/pull/637 |
| [10:53:41] | <symbioquine[m]> | Yeah, I was just looking at that... |
| [10:54:11] | <mstenta[m]> | Just a sketch - probably could be improved, but curious to hear thoughts at least |
| [10:55:10] | <symbioquine[m]> | I'm trying to understand the consequences of https://github.com/farmOS/farmOS/pull/635 and https://github.com/farmOS/farmOS/pull/637 together in terms of the ordering of things like movements... |
| [10:56:01] | <symbioquine[m]> | It seems like although the ordering would still work (by falling back on the log ids), it would end up being less intuitive because it users wouldn't see the obvious "this log occurs after this other log" relationship as clearly. |
| [10:56:59] | <symbioquine[m]> | I think it is still valuable to set the time in many scenarios even when the user mostly "doesn't care" about it. |
| [10:58:49] | <symbioquine[m]> | Another problematic scenario is when the user selects a date in the past for a new log - because it is probably unintuitive for that log to (by default) "come before" other logs on that date. |
| [10:59:18] | <symbioquine[m]> | (All this is assuming I'm understanding the effects of those changes correctly.) |
| [11:03:13] | <symbioquine[m]> | * It seems like although the ordering would still work (by [falling back on the order of the log ids](https://github.com/farmOS/farmOS/blob/ac72de72bac7d2ff03ddd7207a872522e0...)), it would end up being less intuitive because it users wouldn't see the obvious "this log occurs after this other log" relationship as clearly. |
| [15:48:34] | <mstenta[m]> | symbioquine: yea we should consider all the different contexts... |
| [15:49:46] | <mstenta[m]> | Not attached to it... just something that all of a sudden became easy to try :-) |
| [15:50:47] | <mstenta[m]> | Maybe it could be an optional JS lib too... that could be added to select forms and made available as an option |
| [15:51:02] | <mstenta[m]> | Current PR just applies it everywhere |
| [15:55:39] | <mstenta[m]> | > I think it is still valuable to set the time in many scenarios even when the user mostly "doesn't care" about it. |
| [15:55:39] | <mstenta[m]> | Existing behavior doesn't change (if i'm understanding your thoughts), the time field just gets hidden with a link to show it. The default value of the time is the same with all these changes. |