Adds a runtime variable __webpack_require__.dp to generate a dynamic path

[arkapratimc]

fixes #18895

What kind of change does this PR introduce? It just introduces one variable __webpack_require__.dp which generates a dynamic path. Users can write __webpack_dynamic_public_path__ = (filename, publicPath) => .... to generate too.

Did you add tests for your changes? Yes

Does this PR introduce a breaking change? No

What needs to be documented once your changes are merged? A note of what __webpack_dynamic_public_path__ does in https://webpack.js.org/api/module-variables/

[arkapratimc]

Hey @alexander-akait , I just have one question, what’d be the signature of the function of __webpack_public_path__ ? For example, will it be like this →

__webpack_public_path__ = (filename) => "hey/" + filename;

[alexander-akait]

@arkapratimc Yes, you are right

[Jack-Works]

Does this work for chunk loading?

[alexander-akait]

@Jack-Works no, missing in this PR, anyway it is just __webpack_public_path__ improvement, in the original issue we need to allow define custom public path function in build time, not runtime

[arkapratimc]

hey @alexander-akait, test passes on one ubuntu machine https://dev.azure.com/webpack/webpack/_build/results?buildId=20771&view=logs&j=059dd913-30b4-5dab-74c6-1d726b9b84c8&t=0f11fa17-217a-5071-18f9-b54ec876c796&l=54 , but fails on another https://github.com/webpack/webpack/actions/runs/14623185311/job/41028317052?pr=19238#step:12:8 . Any idea on whats causing this ?

[hai-x]

hey @alexander-akait, test passes on one ubuntu machine dev.azure.com/webpack/webpack/_build/results?buildId=20771&view=logs&j=059dd913-30b4-5dab-74c6-1d726b9b84c8&t=0f11fa17-217a-5071-18f9-b54ec876c796&l=54 , but fails on another webpack/webpack/actions/runs/14623185311/job/41028317052?pr=19238#step:12:8 . Any idea on whats causing this ?

You need check the filename type. It can be { expr: string }. So the stringified (${filename}) result will be [object Object].

[codspeed-hq[bot]]

CodSpeed Performance Report

Merging #19238 will not alter performance

_{Comparing arkapratimc:issue-18895 (1863436) with main (ebeace3)}

Summary

✅ 33 untouched benchmarks
🆕 12 new benchmarks

Benchmarks breakdown

	Benchmark	BASE	HEAD	Change
🆕	benchmark "future-defaults", scenario '{"name":"mode-development","mode":"development"}'	N/A	286.1 ms	N/A
🆕	benchmark "future-defaults", scenario '{"name":"mode-development-rebuild","mode":"development","watch":true}'	N/A	34.9 ms	N/A
🆕	benchmark "future-defaults", scenario '{"name":"mode-production","mode":"production"}'	N/A	2.4 s	N/A
🆕	benchmark "many-chunks-commonjs", scenario '{"name":"mode-development","mode":"development"}'	N/A	220.7 ms	N/A
🆕	benchmark "many-chunks-commonjs", scenario '{"name":"mode-development-rebuild","mode":"development","watch":true}'	N/A	42.8 ms	N/A
🆕	benchmark "many-chunks-commonjs", scenario '{"name":"mode-production","mode":"production"}'	N/A	1.9 s	N/A
🆕	benchmark "many-chunks-esm", scenario '{"name":"mode-development","mode":"development"}'	N/A	213.8 ms	N/A
🆕	benchmark "many-chunks-esm", scenario '{"name":"mode-development-rebuild","mode":"development","watch":true}'	N/A	42.1 ms	N/A
🆕	benchmark "many-chunks-esm", scenario '{"name":"mode-production","mode":"production"}'	N/A	2.2 s	N/A
🆕	benchmark "typescript-long", scenario '{"name":"mode-development","mode":"development"}'	N/A	30.9 s	N/A
🆕	benchmark "typescript-long", scenario '{"name":"mode-development-rebuild","mode":"development","watch":true}'	N/A	27.7 ms	N/A
🆕	benchmark "typescript-long", scenario '{"name":"mode-production","mode":"production"}'	N/A	30.9 s	N/A

[arkapratimc]

@hai-x hi, thanks, done - https://github.com/webpack/webpack/pull/19238/files#diff-07626cfb7da83ee25692aef62aec9dc6a32d6f0511b00e6b342121c5a87b5c75R596

Tests on the same machine are still failing ... (1) https://dev.azure.com/webpack/webpack/_build/results?buildId=20859&view=logs&j=059dd913-30b4-5dab-74c6-1d726b9b84c8&t=0f11fa17-217a-5071-18f9-b54ec876c796&l=54 (2) https://github.com/webpack/webpack/actions/runs/14811503127/job/41586744979?pr=19238#step:12:8

do you think, that making a separate PR can solve this, since it was created 3 months ago ?