build: development server builds for solid-js and @solidjs/web; uniform <entry>.dev.js naming

[ryansolid]

Summary

Step one of the server observability track (plan: documentation/plans/server-dev-build-plan.md, P0). Three commits.

1. Development server builds (80ff52e1)

SSR had no dev build. The only server artifact was built with _SOLID_DEV_ stripped, so the 26 dev gates in @solidjs/web's src/server.ts (head/preload descriptor validation, useHead warnings, the committed-response header guard) never ran outside the test suite — and both server entries hard-coded their public dev flags (solid-js export const DEV = undefined, @solidjs/web export const isDev = false).

• solid-js: dist/server.dev.{js,cjs}; replaceDev(false) added to the prod server build (no gates in src/server/ yet — insurance so the first one can't constant-fold into the dev branch in prod, the Production SSR bundle throws on post-commit header writes #2982 failure @solidjs/web shipped with).
• @solidjs/web: dist/server.dev.* and frames/dist/server.dev.*.
• exports: development nested under worker/deno/node for every server entry, including the nested frames/package.json stub. Nesting is required — those conditions precede the top-level development key, so a top-level entry never matches on a server. Same shape ./server-functions already used.
• Server entries now gate DEV / isDev on the same _SOLID_DEV_ replace as their internals. solid-js's server DEV is @solidjs/signals' DEV object in the dev artifact (so DEV.diagnostics.subscribe/capture work server-side — this is the channel P1 emits into) and undefined in prod. The export-parity test asserted the old hard-coded undefined; it now asserts identity with the client's DEV.

2. Uniform dev artifact naming (bd22ac83)

Every entry added after the originals used <entry>.dev.js (refresh, server, frames/client, frames/server, server-functions/server); only the three original client entries were a bare dist/dev.js. With dist/server.dev.js shipping, dev.js no longer said which entry it was the dev build of. Renamed: solid.dev.*, web.dev.*, universal.dev.*. File names are private to the exports map; only deep imports of dist/dev.js are affected. Signals keeps dev.js vs prod/ (chunked dir for the mangle pass — a restructure, not a rename).

3. Signals development CJS build (c07a044e)

The require branch of @solidjs/signals' exports had only the prod dist/node.cjs (__DEV__ false). A CJS host resolving solid-js's dist/server.dev.cjs under development would require a prod signals and get DEV === undefined from a dev artifact — a lie that stays invisible until P1 makes the server emit into DEV.diagnostics. Added dist/node.dev.cjs (flat, unmangled twin of dist/dev.js) behind require.development; prod CJS unchanged. Also: solid-js#test now depends on solid-js#build in turbo.json — it had no dist-based tests until this branch, so it was the only package whose test task didn't wait for its own build.

Behavior change

Dev SSR hosts that pass the development condition (Vite dev does by default: ssr.resolve.conditions is ['module', 'node', 'development|production']) now get the dev artifacts. A header write after the response has committed throws with the offending header named, where the production artifact continues to console.error and drop the write. Called out in the changeset.

Tests

• web/test/server/dist-server-dev-artifact.spec.tsx — dev artifact throws on late header write; isDev === true. Twin of the existing prod spec, which gains isDev === false. A string scan can't pin this (babel's folding erases the marker either way); the behavior does.
• solid/test/server/dist-server-artifact.spec.ts — DEV undefined in prod, is signals' object in dev, same export surface.
• web/test/server/exports-server-conditions.spec.tsx — spawns Node with --conditions and reads import.meta.resolve / createRequire().resolve, so Node's real resolver (not a reimplementation) pins that all five server entries flip to .dev together under node, worker, deno ± development; that browser ± development pairs the client artifacts; and that the CJS hops @solidjs/web → solid-js → @solidjs/signals flip together. Catches the key-ordering trap, the rename, and a prod-signals-under-dev-solid mix permanently.
• signals/tests/dist-cjs-artifacts.test.ts — DEV undefined in node.cjs, live with diagnostics in node.dev.cjs, same export surface.

Verified: @solidjs/signals 1590, solid-js 588, @solidjs/web 712 / 766 / 165, universal 43, h 61, html 192, element 10, type tests, prettier. @solidjs/web dist/server.dev.js is +5.5 KB over prod; solid-js's differs only in DEV.

Not in this PR

• P1–P4 of the plan (server diagnostics channel, converting the existing warnings to codes, hook surface, @solidjs/diagnostics server scenario). P1 is unblocked by this PR plus feat(diagnostics): responsiveness attribution — holds, provenance, feedback tables, effect-sync and identity diagnostics #3302.
• documentation/proposals/production-observability-sketch.md rides along as the "why" for the track. It is a proposal, not something this PR implements; skip it if you're reviewing the code.

---

Authored with Claude via Cursor.

[coveralls]

Coverage Report for CI Build 34212340599

Coverage remained the same at 71.892%

Details

• Coverage remained the same as the base build.
• Patch coverage: No coverable lines changed in this PR.
• No coverage regressions found.

Uncovered Changes

No uncovered changes found.

Coverage Regressions

No coverage regressions found.

---

Coverage Stats

Relevant Lines:	1002
Covered Lines:	768
Line Coverage:	76.65%
Relevant Branches:	784
Covered Branches:	516
Branch Coverage:	65.82%
Branches in Coverage %:	Yes
Coverage Strength:	14.99 hits per line

---

💛 - Coveralls

[codspeed-hq]

Merging this PR will degrade performance by 29.9%

⚠️ Different runtime environments detected

Some benchmarks with significant performance changes were compared across different runtime environments,
which may affect the accuracy of the results.

Open the report in CodSpeed to investigate

❌ 2 regressed benchmarks
✅ 140 untouched benchmarks

Warning

Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Benchmark	BASE	HEAD	Efficiency
❌	merge	186.8 µs	293.4 µs	-36.35%
❌	merge	266.4 µs	345.1 µs	-22.81%

Tip

Investigate this regression by commenting @codspeedbot fix this regression on this PR, or directly use the CodSpeed MCP with your agent.

---

_{Comparing server-dev-build (c07a044) with next (94fe5b4)}

[changeset-bot]

🦋 Changeset detected

Latest commit: c07a044

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 11 packages

Name	Type
solid-js	Patch
@solidjs/web	Patch
@solidjs/universal	Patch
@solidjs/signals	Patch
@solidjs/element	Patch
@solidjs/h	Patch
@solidjs/html	Patch
test-integration	Patch
@solidjs/babel-plugin	Patch
@solidjs/compiler	Patch
@solidjs/diagnostics	Patch

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