| [07:04:16] | * farmBOT has joined #farmos |
| [13:27:52] | <symbioquine[m]> | I just released an updated farmOS_wfs version that works with farmOS 3.1.1. |
| [13:27:52] | <symbioquine[m]> | Interestingly, there were a bunch of geometry floating-point rounding differences between farmOS 2.x and 3.x: https://github.com/symbioquine/farmOS_wfs/commit/acaa7797f5015592f066b97... |
| [13:28:22] | <symbioquine[m]> | s/farmOS_wfs/[farmOS\_wfs](https://www.drupal.org/project/farmos_wfs/releases/2.0.1)/, s/farmOS_wfs/farmOS\_wfs/ |
| [13:31:09] | <mstenta[m]> | Oh weird |
| [13:31:37] | <symbioquine[m]> | Yeah, I wouldn't say either is especially "better" |
| [13:31:41] | <mstenta[m]> | Any chance PHP GEOS wasn't installed? |
| [13:32:36] | <symbioquine[m]> | idk, I just used the 3.1.1 image for testing... https://github.com/symbioquine/farmOS_wfs/blob/acaa7797f5015592f066b9740... |
| [13:33:14] | <mstenta[m]> | Hmm ok |
| [13:34:28] | <mstenta[m]> | That and the patches to geophp are the only things I know of that would affect those things |
| [13:35:16] | <mstenta[m]> | I don't think either of those changed between v2 and v3 |
| [13:35:26] | <mstenta[m]> | (Geos / geophp) |
| [13:36:34] | <mstenta[m]> | Oh weird it looks like they all got an extra decimal place? |
| [13:36:47] | <symbioquine[m]> | mostly |
| [13:36:59] | <mstenta[m]> | Huh |
| [13:37:42] | <mstenta[m]> | Ok so they aren't actually different values (which would suggest one of the things I said above changed)... just more precise |
| [13:37:45] | <symbioquine[m]> | -70.66046174786899 became -70.660461747869 |
| [13:38:01] | <mstenta[m]> | Oh hmm |
| [13:38:06] | <mstenta[m]> | Weird |
| [13:38:09] | <symbioquine[m]> | and -81.30167320572301 became -81.301673205723 |
| [13:40:19] | <symbioquine[m]> | But the difference is between 12 and 14 decimal places in that example and actually only 0.00000000000001 for reference, Wikipedia says that the 8th decimal place is only about 1mm of significance. |
| [13:40:32] | <symbioquine[m]> | s//`/, s//`./, s/for/For/ |
| [13:41:21] | <symbioquine[m]> | Arguably, those aren't significant figures because (basically) nobody has equipment capable of measuring at that precision. |
| [13:42:06] | <symbioquine[m]> | https://en.wikipedia.org/wiki/Decimal_degrees https://en.wikipedia.org/wiki/Significant_figures |
| [13:42:07] | <mstenta[m]> | Yea, so we're just fiddling with grass blades basically :-) |
| [13:42:14] | <mstenta[m]> | Still, curious what changed |
| [13:43:23] | <symbioquine[m]> | Yeah, if I get a chance to look I'll report back. For now the tests did their job and both showed the difference and that it wasn't majorly broken so I'm going on to upgrading the next module. |
| [13:43:39] | <mstenta[m]> | Awesome! |
| [13:43:58] | <mstenta[m]> | Thanks for the heads up. Definitely curious! |
| [13:44:36] | <mstenta[m]> | (Could it be a QGIS change and not farmOS? 🤔) |
| [13:44:59] | <symbioquine[m]> | I don't think so |
| [13:45:02] | <mstenta[m]> | (I haven't looked closely at how the tests work) |
| [13:45:31] | <symbioquine[m]> | I initially didn't update QGIS |
| [13:47:41] | <symbioquine[m]> | There weren't any rounding differences (effecting the tests anyway) between QGIS 3.26 (the version the tests were on before) and 3.34 (the version the tests were on now). |
| [13:48:20] | <symbioquine[m]> | The nice thing is that QGIS provides official release Docker images and my test harness just runs the tests within that container: https://github.com/symbioquine/farmOS_wfs/blob/acaa7797f5015592f066b9740... |
| [13:48:53] | <symbioquine[m]> | I don't think it's worth the Github actions time, but I could run a test matrix against a bunch of QGIS versions trivially. |
| [13:49:32] | <symbioquine[m]> | If lots of folks started using farmOS_wfs with QGIS it might be worth it, but AFAIK I don't have a large user base at the moment. |
| [13:49:42] | <mstenta[m]> | There were a few updates to the Geofield Drupal module over the past 6 months or so... maybe something changed there |
| [13:50:27] | <symbioquine[m]> | Seems like a good guess |
| [13:50:27] | <mstenta[m]> | Or maybe PostgreSQL version change, but thats doubtful too |
| [13:51:09] | <symbioquine[m]> | I suppose |