fix client package #1027
fix client package #1027
Conversation
turbocrime
force-pushed
the
turbocrime/fix-client-package
branch
2 times, most recently
from
55a240c to
90a2364
Compare
|
working demo with this client package here https://github.com/turbocrime/nextjs-penumbra-client-example |
| @@ -5,7 +5,7 @@ export enum PenumbraRequestFailure { | |||
|
|
|||
| export const PenumbraSymbol = Symbol.for('penumbra'); | |||
|
|
|||
| export interface PenumbraProvider { | |||
| export interface PenumbraInjection { | |||
There was a problem hiding this comment.
What's the thinking behind the rename? Is this standard with other wallets? I feel like the Provider naming is more common in things I've encountered.
There was a problem hiding this comment.
Provider naming is common for react components that provide something to child components, which felt awkward while writing the demo and using other Provider-named things, so i renamed this.
| "prebuild": "$npm_execpath run clean", | ||
| "prepack": "$npm_execpath run build" |
There was a problem hiding this comment.
What's $npm_execpath do here?
There was a problem hiding this comment.
$npm_execpath will refer to npm/pnpm/yarn/whatever was used to call the script. it seems like this is the only cross-package-manager way for one package script to call another. these scripts probably don't matter because a consumer won't call them outside of our monorepo, but in general when writing a package for publication it's better to not name an explicit npm/pnpm/yarn in scripts
|
i'm going to add a changeset patch bump to bech32m to calm down the peer dependency, and then merge this |
turbocrime
force-pushed
the
turbocrime/fix-client-package
branch
from
5070eaa to
d6d65f1
Compare
some more client package work in another PR but this is enough for the demo