ci: gate Claude Code install to Linux server-tests
Installing claude-code globally on the Windows server-tests runner exposed 12 pre-existing Windows-specific test failures that were silently skipped before (rmdir EBUSY during cleanup, .CMD spawn EINVAL via the SDK, plus the launch-regression test that asserts claude is *not* on PATH). The Linux e2e tests (model-catalog, live-preferences, models) are the only ones that needed the binary to flip isAvailable() true; none of those run on Windows. Skip the install on Windows so those Windows compatibility gaps stay surfaced separately and don't block the SDK upgrade.
This commit is contained in:
@@ -88,7 +88,8 @@ jobs:
|
||||
- name: Install dependencies
|
||||
run: npm install
|
||||
|
||||
- name: Install Claude Code CLI for provider tests
|
||||
- name: Install Claude Code CLI for provider tests (Linux only)
|
||||
if: runner.os == 'Linux'
|
||||
run: npm install -g @anthropic-ai/claude-code
|
||||
|
||||
- name: Build highlight dependency
|
||||
|
||||
Reference in New Issue
Block a user