[Bug]: Clone fails when git clone exceeds a fixed 120s timeout

[eshatalov]

Area

apps/server (SourceControlRepositoryService)

Summary

SourceControlRepositoryService.cloneRepository runs git clone with a hardcoded timeoutMs: 120_000. Any clone that takes longer than two minutes, whether because the repository is large or the link is slow, is killed mid-download and surfaces as the generic toast:

Clone failed. Source control repository operation cloneRepository failed for unknown: The source control operation could not be completed.

The trace log shows the real cause: GitCommandError: ... Git command timed out.

Clone duration scales with repository size and network speed, so any fixed budget strands someone. Every push variant in GitVcsDriverCore already runs with timeoutMs: null for this reason. Clone should follow the same precedent.

Steps to reproduce

1. Pick any repository whose clone takes more than two minutes on the server's connection.
2. Clone it via "Add project from remote" in the web app.

Expected

The clone completes, or fails with the actual git error.

Actual

At 120s the clone is killed and the generic "could not be completed" toast appears. Nothing indicates a timeout.

Notes

• The 120s value dates from Add remote repository publish and discovery support #2482 with no recorded rationale.
• [Bug]: Windows: hardcoded 15s timeout in removeWorktree is unreachable for large monorepos, stranding worktree directories #5614 / Fix worktree removal timing out on large install trees #3902 fixed the same shape of problem for removeWorktree.
• fix(source-control): explain clone failures #5005 (open) improves clone error messages, touches the same call, but keeps the 120s budget.