IRC logs for #farmOS, 2022-04-18 (GMT)

[20:56:17]* farmBOT has joined #farmos
[01:48:15]* paul121[m] has quit (*.net *.split)
[01:48:15]* mstenta[m] has quit (*.net *.split)
[01:48:15]* calbasi[m] has quit (*.net *.split)
[01:48:15]* dazinism[m] has quit (*.net *.split)
[01:49:36]* paul121[m] has joined #farmos
[01:55:10]* dazinism[m] has joined #farmos
[01:55:58]* calbasi[m] has joined #farmos
[01:56:09]* mstenta[m] has joined #farmos
[05:03:49]* wybpip[m] has joined #farmos
[05:03:50]* wybpip[m] has left #farmos ()
[08:39:12]* GuilhermePerotta has joined #farmos
[15:10:29]<mstenta[m]>paul121: FYI you may have seen already but I started an issue to debug the automated test failures issue
[15:11:32]<mstenta[m]>it's actually two different things i think... scheduled runs are failing to build images, and branch-push runs fail during MariaDB tests (but maybe not all the time? still need to look closer)
[15:12:18]<mstenta[m]>i thought i had actually narrowed it down to the fact that we are now branching from `build` to both `sniff` and `test` (removing `sniff` caused `test` to pass) ... but then it started acting inconsistently, so i need to double check that
[15:13:08]<paul121[m]>yes! saw that bug for the scheduled runs. seems like setting a default `FARMOS_VERSION` env to `2.x` would fix it?
[15:13:19]<mstenta[m]>yea i think so
[15:13:53]<paul121[m]>haven't had a chance to grok the rest
[15:14:27]<mstenta[m]>if i can't isolate it i might ask for your and symbioquine's eyes on the weekly dev call :-)
[15:27:18]<mstenta[m]>ludwa6: I'm a bit tangled up regarding the 2.0.0-beta4 release on Farmier :-/
[15:27:58]<mstenta[m]>May need to wait until we fix this Google Maps issue:
[15:28:54]<mstenta[m]>If I roll out the update right now it will break maps for anyone who installed the Google module :-(
[15:31:16]<mstenta[m]>I may have to uninstall it from all instances to unblock the upgrade - and just explain to folks why they no longer have Google layers available if they ask (at least until a fix is available)
[15:31:28]<mstenta[m]>Weighing options...
[15:52:16]<symbioquine[m]>I can probably work on a fix for the GM issue this evening if that helps.
[15:52:43]<mstenta[m]>thanks symbioquine! but i don't want to put you out - it's my problem not yours :-)
[15:53:10]<mstenta[m]>you know my opinion on google map layers :-P
[15:54:12]<symbioquine[m]>Should be a simple matter of bumping the OL version, pushing a new release to NPM then pointing farmOS-map at that version of olgm then making a new farmOS-map release.
[15:54:24]<mstenta[m]>oh ok that's good :-)
[15:55:20]<mstenta[m]>it sounded like it was going to require work in `farmos-olgm` (and/or `ol3-google-maps` upstream) to add an `unrender` method
[15:56:16]<symbioquine[m]>mstenta[m]: Either way :)
[15:56:39]<mstenta[m]>> then making a new farmOS-map release
[15:56:39]<mstenta[m]>ah so it will require a new farmOS release too then
[15:56:45]<mstenta[m]>(to reference the new version of farmOS-map)
[15:58:00]<mstenta[m]>i think i will uninstall the module from the instances i host, and send an email to their owners letting them know why, and that it will be fixed in the next version...
[15:58:12]<mstenta[m]>just to keep things simple
[15:58:16]<symbioquine[m]>Yeah, though individual projects could probably also override the farmOS-map version in their composer files.
[15:58:34]<symbioquine[m]>mstenta[m]: Fair enough
[16:04:42]<symbioquine[m]>Interestingly, this `unrender` change wasn't mentioned in Openlayers' [upgrade guidance] ( - though there are other things we should maybe track a bit closer when doing version bumps...
[16:05:23]<mstenta[m]>i did a quick search in the `ol3-google-maps` issue queue but didn't see any mention of it there...
[16:05:57]<mstenta[m]>but i'm not sure how many are actually using that... and if they've tried upgrading yet :-)
[16:06:09]<mstenta[m]>google maps... always a headache!
[16:06:52]<symbioquine[m]>Yeah it's like an ABI incompatibility issue. Only occurs when parts of an application are built against two different versions of OL.
[16:07:10]<mstenta[m]>ahh ok
[16:07:16]<mstenta[m]>gotcha - that makes sense
[16:07:42]<symbioquine[m]>Tbf it would require a stricter definition of "backwards compatibility" to avoid this.