19913d85-f662-4518-a9a1-7adc1793dfd4.png){kind=small}
Channels
#announcements
- s
schrockn
11/14/2018, 12:32 PMright - s
schrockn
11/14/2018, 12:32 PMso - f
freiksenet
11/14/2018, 12:32 PMeg we might have fully gui config - s
schrockn
11/14/2018, 12:32 PMya - f
freiksenet
11/14/2018, 12:32 PMand then it's stupid to convert that back and forth through yaml - s
schrockn
11/14/2018, 12:32 PMjson, yaml, future format, UI - s
schrockn
11/14/2018, 12:32 PMetc - s
schrockn
11/14/2018, 12:33 PMso - s
schrockn
11/14/2018, 12:33 PMif we project into graphql - s
schrockn
11/14/2018, 12:33 PMwe have an interesting problem - s
schrockn
11/14/2018, 12:33 PMhilariously related to input unions again - f
freiksenet
11/14/2018, 12:33 PMoh yeah - s
schrockn
11/14/2018, 12:33 PMwithout generics - s
schrockn
11/14/2018, 12:33 PMthis is hard - s
schrockn
11/14/2018, 12:33 PMor super awkward at a minimum - f
freiksenet
11/14/2018, 12:34 PMwe can treat it as "json type" - f
freiksenet
11/14/2018, 12:34 PMat first - f
freiksenet
11/14/2018, 12:34 PMso no type level validation on graphql side - f
freiksenet
11/14/2018, 12:34 PMI think it's okay-ish - s
schrockn
11/14/2018, 12:34 PMwhat do you mean by “json type” exactly - f
freiksenet
11/14/2018, 12:34 PMthere is this popular graphql pattern that I don't like, but might be okay here - f
freiksenet
11/14/2018, 12:35 PM - f
freiksenet
11/14/2018, 12:36 PMthis is future compatible if we do project our type system to graphql - s
schrockn
11/14/2018, 12:36 PMso all that does is move the parsing into the GraphQL type system - f
freiksenet
11/14/2018, 12:36 PMyeah - s
schrockn
11/14/2018, 12:36 PMrather than have that dumb function in execution.py - f
freiksenet
11/14/2018, 12:36 PMyeah - f
freiksenet
11/14/2018, 12:37 PMfor now, yes - f
freiksenet
11/14/2018, 12:37 PMbut it's like, json - s
schrockn
11/14/2018, 12:37 PMjson also easily to deal with within tools