PR #153 has contradictory test expectations for BuildPhase2 when the SSR command is not found:
Old tests (expect error):
- Line 159:
Expect(err).To(HaveOccurred(), "Phase 2 must attempt to invoke ssr.command")
- Line 229:
Expect(err).To(HaveOccurred(), "BuildPhase2 must return error when ssr.command is not found")
New tests (#164, expect NO error):
- Line 368:
Expect(err).NotTo(HaveOccurred(), "BuildPhase2 must not return a fatal error when SSR fails")
- Line 387:
Expect(err).NotTo(HaveOccurred(), "SSR failure for one page must not abort the build")
Both use a nonexistent command with pages containing custom elements. The old tests expect BuildPhase2 to return an error. The new tests expect it to skip failed pages and return a result with no error. These cannot both be satisfied simultaneously.
One resolution: the old tests could be updated to check that BuildPhase2 attempts the command (proven by the result containing original HTML) rather than requiring an error return.
PR #153 has contradictory test expectations for BuildPhase2 when the SSR command is not found:
Old tests (expect error):
Expect(err).To(HaveOccurred(), "Phase 2 must attempt to invoke ssr.command")Expect(err).To(HaveOccurred(), "BuildPhase2 must return error when ssr.command is not found")New tests (#164, expect NO error):
Expect(err).NotTo(HaveOccurred(), "BuildPhase2 must not return a fatal error when SSR fails")Expect(err).NotTo(HaveOccurred(), "SSR failure for one page must not abort the build")Both use a nonexistent command with pages containing custom elements. The old tests expect BuildPhase2 to return an error. The new tests expect it to skip failed pages and return a result with no error. These cannot both be satisfied simultaneously.
One resolution: the old tests could be updated to check that BuildPhase2 attempts the command (proven by the result containing original HTML) rather than requiring an error return.