Skip to content

Latest commit

 

History

History
56 lines (32 loc) · 2.21 KB

0013-remove-extract-class-and-convert-to-pure-component-refactorings.md

File metadata and controls

56 lines (32 loc) · 2.21 KB
Raw

13. Remove Extract Class and Convert to Pure Component refactorings

Date: 2023-04-13

Status

Accepted

Amends 8. Don't propose Quick Fix for React Convert to Pure Component

Context

Among all refactorings we implemented, 2 are quite different. They are not using the regular process, but instead import their own packages:

  • "Extract Class" uses ts-morph
  • "Convert to Pure Component" uses jscodeshift and react-codemod

While these were cool demonstrations of the plug-and-play architecture of Abracadabra, it comes at a cost.

In particular, the final bundle contained twice the typescript.js file, which is really unefficient for a single refactoring:

The resulting production bundle is ~40 MB (3.5 MB gzipped).

It also created some technical debt, since:

  • we wanted to refactor the code to retrofit the style of other refactorings (consistency)
  • the react-codemod source code has code that generated noise in our tests
  • "Convert to Pure Component" is specific to React and was causing performance issues

Finally, these are edge-case refactorings that are not "core" ones. Even though there is no telemetry to have an objective view on this, these aren't as common as "Extract Variable", say.

Decision

The following refactorings aren't part of Abracadabra anymore (until re-implemented):

  • "Extract Class"
  • "Convert to Pure Component"

Consequences

The production bundle is much smaller:

Before After
38.49 MB 26.1 MB (-32%)
3.48 MB (gzipped) 2.42 MB (-30%)

It should also result in faster bundles and more consistency in the codebase.

On the other hand, we have lost 2 refactorings. These should be re-implemented following what Abracadabra uses today—or at least, ensures the ROI is worth it.