Use new.target instead of instanceof to detect if its a constructor call

[Jarred-Sumner]

What does this PR do?

Replaces usages of

if (!(this instanceof MyClass)) return new MyClass(...args)

with

if (!new.target) return new MyClass(...args)

It should be faster, though I have yet to benchmark it.

See also nodejs/node#51026

How did you verify your code works?

Existing tests should cover it

[paperdave]

stupid question but are you certain that all these classes have at least one test that calls the constructor without new?

[github-actions]

❌ @paperdave 10 files with test failures on bun-darwin-aarch64:

• test/cli/install/bunx.test.ts
• test/cli/run/env.test.ts
• test/cli/run/require-cache.test.ts
• test/js/bun/test/test-test.test.ts
• test/js/node/crypto/crypto.key-objects.test.ts
• test/js/node/util/parse_args/parse-args.test.mjs
• test/js/node/watch/fs.watchFile.test.ts
• test/js/third_party/jsonwebtoken/schema.test.js
• test/js/web/fetch/body-stream.test.ts
• test/js/web/worker.test.ts

View test output

^{#1f324fee463c9021760c8058aaca425f4ae84a36}

[github-actions]

❌ @paperdave 6 files with test failures on linux-x64-baseline:

• test/cli/install/bun-install.test.ts
• test/cli/run/env.test.ts
• test/js/node/crypto/crypto.key-objects.test.ts
• test/js/node/util/parse_args/parse-args.test.mjs
• test/js/node/v8/capture-stack-trace.test.js
• test/js/third_party/jsonwebtoken/schema.test.js

View test output

^{#1f324fee463c9021760c8058aaca425f4ae84a36}

[github-actions]

❌ @paperdave 6 files with test failures on linux-x64:

• test/cli/run/env.test.ts
• test/js/node/crypto/crypto.key-objects.test.ts
• test/js/node/util/parse_args/parse-args.test.mjs
• test/js/node/v8/capture-stack-trace.test.js
• test/js/node/watch/fs.watch.test.ts
• test/js/third_party/jsonwebtoken/schema.test.js

View test output

^{#1f324fee463c9021760c8058aaca425f4ae84a36}

[github-actions]

❌ @Jarred-Sumner 12 files with test failures on bun-darwin-x64:

• test/cli/run/env.test.ts
• test/js/bun/spawn/spawn-streaming-stdout.test.ts
• test/js/bun/util/which.test.ts
• test/js/node/crypto/crypto.key-objects.test.ts
• test/js/node/fs/fs.test.ts
• test/js/node/string_decoder/string-decoder.test.js
• test/js/node/util/parse_args/parse-args.test.mjs
• test/js/node/v8/capture-stack-trace.test.js
• test/js/node/watch/fs.watchFile.test.ts
• test/js/third_party/jsonwebtoken/schema.test.js
• test/js/third_party/webpack/webpack.test.ts
• test/js/web/timers/setTimeout.test.js

View test output

^{#2f4f3bb76ae8f660593abdad871531abeb3913fd}

[github-actions]

❌ @paperdave 13 files with test failures on bun-darwin-x64-baseline:

• test/cli/run/env.test.ts
• test/integration/next/default-pages-dir/test/next-build.test.ts
• test/js/bun/spawn/spawn-streaming-stdout.test.ts
• test/js/bun/util/filesink.test.ts
• test/js/bun/util/which.test.ts
• test/js/node/crypto/crypto.key-objects.test.ts
• test/js/node/fs/fs.test.ts
• test/js/node/util/parse_args/parse-args.test.mjs
• test/js/node/v8/capture-stack-trace.test.js
• test/js/node/watch/fs.watchFile.test.ts
• test/js/third_party/jsonwebtoken/schema.test.js
• test/js/third_party/webpack/webpack.test.ts
• test/js/web/timers/setTimeout.test.js

View test output

^{#1f324fee463c9021760c8058aaca425f4ae84a36}

[paperdave]

in the future we need to not put multiple things into the same pr. i know that is hypocritical of me considering #6938 having like 50 changes, but still.

i'm skeptical of the streams changes and would honestly want that as a separate pr. it feels like something we're going to end up breaking and have to revert, but you have valuable changes to the syntax highlighter and napi.

[paperdave]

ive moved the changes not related to new.target to another pr.

changes to node:stream must be done carefully. after looking over it a few times it looks alright, but i still wonder if there is anything that can break with using !new.target

i think this should only be merged after we manually verify that Next.js, Express, Fastify, Webpack, and some other packages that may abuse stream mechanics to verify they still work.

[paperdave]

(see prev comment)