ppr: fail static generation if postponed & missing postpone data

[ztanner]

When postpone is caught by user code, this will cause PPR not to properly prerender the static parts and thus we need to fail the build. This also adds some messaging about how to fix the error.

Prior to this change, catching code that would normally trigger postpone would silently fail, but the build outputs would be incorrect as there's no postpone data available.

Relands #57477 with additional tests & fixes

[ijjk]

Stats from current PR

Default Build (Increase detected ⚠️)

General Overall increase ⚠️

	vercel/next.js canary	vercel/next.js ppr/logging-2	Change
buildDuration	10.3s	10.4s	⚠️ +117ms
buildDurationCached	6.1s	6.1s	N/A
nodeModulesSize	175 MB	175 MB	⚠️ +55.3 kB
nextStartRea..uration (ms)	402ms	403ms	N/A

Client Bundles (main, webpack)

	vercel/next.js canary	vercel/next.js ppr/logging-2	Change
199-HASH.js gzip	30 kB	30 kB	N/A
3f784ff6-HASH.js gzip	53.2 kB	53.2 kB	✓
494.HASH.js gzip	182 B	182 B	✓
framework-HASH.js gzip	45.5 kB	45.5 kB	✓
main-app-HASH.js gzip	254 B	252 B	N/A
main-HASH.js gzip	33.1 kB	33.1 kB	N/A
webpack-HASH.js gzip	1.75 kB	1.75 kB	N/A
Overall change	98.9 kB	98.9 kB	✓

Legacy Client Bundles (polyfills)

	vercel/next.js canary	vercel/next.js ppr/logging-2	Change
polyfills-HASH.js gzip	31 kB	31 kB	✓
Overall change	31 kB	31 kB	✓

Client Pages

	vercel/next.js canary	vercel/next.js ppr/logging-2	Change
_app-HASH.js gzip	205 B	205 B	✓
_error-HASH.js gzip	182 B	181 B	N/A
amp-HASH.js gzip	505 B	507 B	N/A
css-HASH.js gzip	322 B	323 B	N/A
dynamic-HASH.js gzip	2.59 kB	2.59 kB	N/A
edge-ssr-HASH.js gzip	258 B	259 B	N/A
head-HASH.js gzip	348 B	347 B	N/A
hooks-HASH.js gzip	369 B	368 B	N/A
image-HASH.js gzip	4.38 kB	4.38 kB	N/A
index-HASH.js gzip	256 B	256 B	✓
link-HASH.js gzip	2.67 kB	2.67 kB	N/A
routerDirect..HASH.js gzip	318 B	318 B	✓
script-HASH.js gzip	384 B	383 B	N/A
withRouter-HASH.js gzip	319 B	320 B	N/A
1afbb74e6ecf..834.css gzip	106 B	106 B	✓
Overall change	885 B	885 B	✓

Client Build Manifests

	vercel/next.js canary	vercel/next.js ppr/logging-2	Change
_buildManifest.js gzip	484 B	484 B	✓
Overall change	484 B	484 B	✓

Rendered Page Sizes

	vercel/next.js canary	vercel/next.js ppr/logging-2	Change
index.html gzip	528 B	527 B	N/A
link.html gzip	542 B	544 B	N/A
withRouter.html gzip	525 B	522 B	N/A
Overall change	0 B	0 B	✓

Edge SSR bundle Size Overall increase ⚠️

	vercel/next.js canary	vercel/next.js ppr/logging-2	Change
edge-ssr.js gzip	96.1 kB	96.2 kB	N/A
page.js gzip	140 kB	140 kB	⚠️ +398 B
Overall change	140 kB	140 kB	⚠️ +398 B

Middleware size

	vercel/next.js canary	vercel/next.js ppr/logging-2	Change
middleware-b..fest.js gzip	626 B	627 B	N/A
middleware-r..fest.js gzip	148 B	151 B	N/A
middleware.js gzip	24.8 kB	24.8 kB	N/A
edge-runtime..pack.js gzip	1.92 kB	1.92 kB	✓
Overall change	1.92 kB	1.92 kB	✓

Diff details

Diff for page.js

Diff too large to display

Diff for edge-ssr.js

Diff too large to display

Diff for 199-HASH.js

Diff too large to display

Commit: 427ad4c

[ijjk]

Tests Passed

[feedthejim]

are we not printing all errors?

[ztanner]

This mirrors the existing behavior for regular prerender failures below (ref) -- my reasoning was that we'll get duplicated errors from the serverComponentsRenderer and flightDataRenderer. We could add additional branching for PPR specifically to dedupe, but was thinking it's not worth the additional complexity, since these errors will still be surfaced by subsequent builds.