mattp- / dbix-class-filtercolumn-bytype Goto Github PK

[13:07:11] <ribasushi> TBSliver: in other words - mattp_  needs to fix FC::BT to not apply to anything that is part of a PK/UC
[13:07:27] <ribasushi> and that would be a breaking change, but there is literally nothing else that can be done :(
[13:07:46] <TBSliver> ribasushi: ta for letting me know :)
[13:07:50] <mattp_> ribasushi: you gave me a gist once
[13:08:04] <ribasushi> mattp_: I probably did, I remember this coming up
[13:08:42] <ribasushi> TBSliver: nah, I should have caught the problem at review-time, way before we got into this mess
[13:10:54] <mattp_> ribasushi: I can push a patch to skip pk/uc 
[13:11:01] <mattp_> should that be done though?
[13:13:19] <mattp_> i didnt think anyone was using the module outside of the $job I quit two years ago :) 
[13:39:16] <ribasushi> mattp_: given dbic will continue to throw exceptions in that case - I'd hazard a guess that, yes, it should be done
[13:40:17] <mattp_> ok
[13:43:20] <mattp_> ribasushi: i seem to remember you had something crazy with end_of_scope or something in your patch. do you remember why?
[13:44:05] <ribasushi> mattp_: because add_columns almost invariably happens before pks/ucs are defined
[13:44:21] <mattp_> ah right
[13:46:37] <mattp_> i.. hmm
[13:47:38] <mattp_> I guess the add_columns wrapper needs to be completely removed/deferred instead then. ill try to take a look tonight
[13:48:55] <mattp_> ribasushi: might poke you for a quick review later if thats ok