fix(atlas): repoint three catalog entries, and correct what they claim - #54
Merged
Merged
Conversation
The report catalog names the test that evidences each requirement, so
renaming a test orphans its section. Three were left dangling:
U5 test_active_flash_format_is_v20 -> missing
U5b test_burned_versions_have_no_reader -> missing
U8 test_every_ladder_version_has_a_reader -> missing
They are not a mechanical rename, because all three encoded the same
false premise: that an unhandled storage version "falls out of the switch"
to a default. storage_fromFlash() has NO default case, deliberately, so
that -Werror=switch names any version nobody handled. An unlisted version
does not fall anywhere -- it fails the ARM build, which is what actually
happened on the passkeys branch.
U5 -> test_last_shipped_never_moves_backwards. The role U5 described --
"an independent witness for the number the whole gate turns on",
because the static assert compares two constants in one header
that one commit can raise together -- is now the LAST_SHIPPED
ratchet. Its old title asserted V20, which is true on 7.16 and
false on 7.15; the ratchet is true on both.
U5b -> test_burned_versions_are_dispatched_to_the_wipe_path. The label
must EXIST; what must not exist is a reader behind it.
U8 -> test_every_shipped_version_has_a_reader. Scoped to SHIPPED on
purpose: a burned format legitimately has none, so "every ladder
version has a reader" would make burning one impossible to
express. test_no_shipped_version_is_burned is what stops that
scoping becoming a loophole.
Verified every catalog reference resolves to a test that exists -- 280
entries, all green -- rather than only the three I touched.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The report catalog names the test that evidences each requirement, so renaming a test orphans its section. Three were left dangling by the storage-gate rework, and CI caught it:
Not a mechanical rename
All three encoded the same false premise: that an unhandled storage version "falls out of the switch" to a default.
storage_fromFlash()has no default case, deliberately, so that-Werror=switchnames any version nobody handled. An unlisted version doesn't fall anywhere — it fails the ARM build, which is exactly what happened on the passkeys branch.test_last_shipped_never_moves_backwardstest_burned_versions_are_dispatched_to_the_wipe_pathtest_every_shipped_version_has_a_readertest_no_shipped_version_is_burnedis what stops that scoping becoming a loophole.Verification
Checked that every catalog reference resolves to a test that exists — 280 entries, all green — not just the three touched here. The check enumerates
(section, module, test)from the catalog and greps each module fordef <test>(.Underlying test run on the branch that surfaced this: 638 passed, 30 skipped, 0 failed. The only thing red was the catalog.