• refactor(rolldown_resolver): improve resolve method clarity and documentation #6986
• refactor(rolldown_resolver): inline unnecessary functions #6985 👈 (View in Graphite)
• refactor(rolldown_resolver): reorganize impl blocks #6984
• refactor(rolldown_resolver): extract configuration logic into separate module #6983
• refactor(rolldown_resolver): improve error messages #6982
• main

How to use the Graphite Merge Queue

Add the label graphite: merge to this PR to add it to the merge queue.

You must have a Graphite account in order to use the merge queue. Sign up using this link.

This stack of pull requests is managed by Graphite. Learn more about stacking.

Could you elaborate the word unnecessary here? On performance asepct or code clarity?

Benchmarks Rust

• target: main(8d1a387)
• pr: 11-13-refactor_rolldown_resolver_inline_unnecessary_functions(67b6599)

group                                                        pr                                     target
-----                                                        --                                     ------
bundle/bundle@multi-duplicated-top-level-symbol              1.00     61.2±2.18ms        ? ?/sec    1.03     63.2±0.98ms        ? ?/sec
bundle/bundle@multi-duplicated-top-level-symbol-sourcemap    1.00     67.6±1.19ms        ? ?/sec    1.09     73.8±2.21ms        ? ?/sec
bundle/bundle@rome_ts                                        1.00    106.1±1.43ms        ? ?/sec    1.01    106.8±1.48ms        ? ?/sec
bundle/bundle@rome_ts-sourcemap                              1.00    119.1±1.52ms        ? ?/sec    1.01    120.5±1.46ms        ? ?/sec
bundle/bundle@threejs                                        1.00     38.0±1.15ms        ? ?/sec    1.02     38.8±2.28ms        ? ?/sec
bundle/bundle@threejs-sourcemap                              1.00     41.8±0.81ms        ? ?/sec    1.01     42.1±0.43ms        ? ?/sec
bundle/bundle@threejs10x                                     1.00    386.1±4.13ms        ? ?/sec    1.03    397.4±6.51ms        ? ?/sec
bundle/bundle@threejs10x-sourcemap                           1.00    445.3±3.94ms        ? ?/sec    1.03    459.5±5.77ms        ? ?/sec
scan/scan@rome_ts                                            1.00     83.3±1.15ms        ? ?/sec    1.02     84.7±1.71ms        ? ?/sec
scan/scan@threejs                                            1.00     27.9±0.37ms        ? ?/sec    1.01     28.4±0.39ms        ? ?/sec
scan/scan@threejs10x                                         1.00    293.5±3.67ms        ? ?/sec    1.00    293.0±4.94ms        ? ?/sec

✅ Deploy Preview for rolldown-rs canceled.

Could you elaborate the word unnecessary here? On performance asepct or code clarity?

Done in PR description

A small helper function called only once that doesn't improve readability can be "unnecessary" - inlining makes the code more direct.

I think the purpose of previous code extracting the function is for better readability and restricted scope.

If there's something wrong about the behavior of infering module format, you could jump into infer_module_def_format and inspect related logic. Inlining the function will make it mixed up with irrelevant logic.

Same for fn cached_package_json too. The cache behavior is considering to be an isolated logic. It adds unnecessary burden when understanding fn resolve if inlining it..

In short, these functions decrease the complexity of fn resolve.

A small helper function called only once that doesn't improve readability can be "unnecessary" - inlining makes the code more direct.

I think the purpose of previous code extracting the function is for better readability and restricted scope.

If there's something wrong about the behavior of infering module format, you could jump into infer_module_def_format and inspect related logic. Inlining the function will make it mixed up with irrelevant logic.

Same for fn cached_package_json too. The cache behavior is considering to be an isolated logic. It adds unnecessary burden when understanding fn resolve if inlining it..

In short, these functions decrease the complexity of fn resolve.

Got it.