Greg-Hawkridge
commented
5 months ago - fixes #2
Closed Greg-Hawkridge closed 5 months ago
Greg-Hawkridge
commented
5 months ago 
tobyaharris
commented
5 months ago Testing:
dev was done first. System was stopped, pruned and dir deleted before directly cloning the bugfix branch. |
Computer | operation | branch & time taken |
|---|---|---|---|
| Pi5 8GB | ./setup_keys.sh |
99.5s on dev 102s on bugfix/trim-packages |
|
| Pi5 8GB | docker compose build |
147.4s on dev 155.6s on bugfix/trim-packages |
|
| Pi5 8GB | ./start.sh |
78.3s on dev 76.0s on bugfix/trim-packages |
|
| Pi4 2GB | ./setup_keys.sh |
?s on dev ?s on bugfix/trim-packages |
|
| Pi4 2GB | docker compose build |
?s on dev ?s on bugfix/trim-packages |
|
| Pi4 2GB | ./start.sh |
?s on dev ?s on bugfix/trim-packages |
1) I won't rush to say this has made a significant saving - results are within experimental noise. 2) While we're editing the NPM dependencies, there are an awful lot of depreciation warnings.
Greg-Hawkridge
commented
5 months ago The deprecation warnings are related to "create-react-app" which creates a template application and aren't a worry because none of them are runtime dependencies for React - I've looked at alternatives - but don't see a reason for the migration effort just yet
Greg-Hawkridge
commented
5 months ago Also I didn't think there would be much of a difference and we're probably within margin of error on these tests
tobyaharris
commented
5 months ago Regardless of by how much this reduces the build time / is there further work to do, this is in no way a bad thing. Approving.