Matt Rakow
Matt Rakow
Latest includes an aqueduct-style container runtime factory to facilitate the pattern.
I had a chance to try this pattern further with the SharedWorker driver project, and I feel it's a fairly positive direction. Since this change would only make changes in...
> @ChumpChief any idea why this causes a bundle size increase? I think it's wrong about the baseline size? Running the bundle-size-tests locally I see the same sizes both with...
> > @ChumpChief any idea why this causes a bundle size increase? > > I think it's wrong about the baseline size? Running the bundle-size-tests locally I see the same...
In particular, this looks to be the problematic bit. Presumably needs to differentiate which comparison commit based on target branch. https://github.com/microsoft/FluidFramework/blob/main/tools/bundle-size-tools/src/utilities/gitCommands.ts#L12
Filed [AB#1570](https://dev.azure.com/fluidframework/235294da-091d-4c29-84fc-cdfc3d90890b/_workitems/edit/1570)
Looks like it's saving ~40s (~14%) on `npm run ci:build` in the PR validation pipeline as well.
I believe the most recent build errors are due to a reliance on webpack being part of the build process for the property-inspector-table package. I'm pausing this PR to investigate...
Incremental build seems to handle these correctly now so I'm less concerned about this. Would still likely be a nice improvement but I don't think I'll pick this back up.
The large changes to package-lock.json and lerna-package-lock.json seem probably not intentional - probably drop those from the PR