Chore: Backport recent refactorings #1648
Conversation
🦋 Changeset detectedLatest commit: ab0a238 The changes in this PR will be included in the next version bump. This PR includes changesets to release 1 package
Not sure what this means? Click here to learn what changesets are. Click here if you're a maintainer who wants to add another changeset to this PR |
| compiler.hooks.emit.tapAsync('CliDevPlugin', (compilation, callback) => { | ||
| let missingDeps = compilation.missingDependencies; | ||
| let nodeModulesPath = resolve(__dirname, '../../../node_modules'); | ||
|
|
||
| // ...tell webpack to watch node_modules recursively until they appear. | ||
| if ( | ||
| Array.from(missingDeps).some( | ||
| file => file.indexOf(nodeModulesPath) !== -1 | ||
| ) | ||
| ) { | ||
| compilation.contextDependencies.push(nodeModulesPath); | ||
| } | ||
|
|
||
| callback(); | ||
| }); | ||
|
|
There was a problem hiding this comment.
This was adopted from CRA's WatchMissingNodeModulesPlugin, however, it seems to have been broken for a few years. This path that it watches resolves to <project-name>/node_modules/preact-cli/node_modules which is pretty pointless, as nothing will ever exist/be added to that.
Rather than correct it, I've elected to remove it after seeing it was recently cut from CRA in v5. It apparently was having a rather large performance impact which they either couldn't solve or decided it wasn't worth solving. Not entirely sure which.
I figure there's probably no reason to doubt the performance issues claims, so I cut it.
rschristian
force-pushed
the
chore/backport
branch
from
8deb57d
to
d4452f8
Compare
What kind of change does this PR introduce?
Minor refactors
Did you add tests for your changes?
Existing cover
Summary
Backport of misc changes made in #1466/#1608 and #1645. I don't want those PRs to get completely out-of-hand and impossible to review, so I'm extracting out some of the more less-focused changes.
jsconfig.jsonfor better intellisense. Seems like a pretty handy addition, caught a bunch of minor issues already.publicPath = http://some-url.com/foo->http://some-url.com/foosw.jswithout this, which is no good.Does this PR introduce a breaking change?
No