Set Playground runtime on reprint-pulled sites to fix startup DB error

[epeicher]

Related issues

• Related to STU-1828

How AI was used in this PR

Claude Code diagnosed the failure from the Studio + daemon logs, reproduced the root cause with the bundled PHP binary, implemented the fix and a regression test, and verified it end-to-end with a real pull-reprint. I reviewed the diagnosis and change.

Proposed Changes

Sites imported with studio pull-reprint failed to start, dying with "Error establishing a database connection."

The bug: a reprint-pulled site's files are laid out for the Playground runtime — VFS mounts place wp-content (and its SQLite db.php drop-in) under /wordpress/wp-content. But the site was registered without a runtime, so it defaulted to native-php. On that runtime PHP resolves the symlinked layout to the real WordPress core path, WP_CONTENT_DIR no longer points at the SQLite drop-in, WordPress falls back to MySQL, and the connection fails.

The fix: register reprint-pulled sites with the Playground runtime (the runtime they're already built for and the one the command uses for the auto-login URL), so they start correctly.

Testing Instructions

1. npm run cli:build
2. node apps/cli/dist/cli/main.mjs pull-reprint --url <your-site>.wpcomstaging.com -y
3. The site starts and serves HTTP 200.
4. Stop it by running studio stop --path ... and start it by running studio start --path ...
5. Verify it starts. On trunk the site does not start successfully

Pre-merge Checklist

• Have you checked for TypeScript, React or other console errors?

[wojtekn]

Is it the correct approach? What if we remove support for Playground?

[wpmobilebot]

📊 Performance Test Results

Comparing 5e96fad vs trunk

app-size

Metric	trunk	5e96fad	Diff	Change
App Size (Mac)	2341.83 MB	2341.83 MB	+0.00 MB	⚪ 0.0%

site-editor

Metric	trunk	5e96fad	Diff	Change
load	1035 ms	1080 ms	+45 ms	⚪ 0.0%

site-startup

Metric	trunk	5e96fad	Diff	Change
siteCreation	6504 ms	6492 ms	12 ms	⚪ 0.0%
siteStartup	6991 ms	7497 ms	+506 ms	🔴 7.2%

---

Results are median values from multiple test runs.

Legend: 🟢 Improvement (faster) | 🔴 Regression (slower) | ⚪ No change (<50ms diff)

[epeicher]

Is it the correct approach? What if we remove support for Playground?

Good catch! I came here to comment that this approach will be superseded by #3881 - comment here so I am closing this PR