Repository navigation
fix(core): forward-slash provenance references on every platform - #61
Open
Apageoflove wants to merge 1 commit into
Open
Apageoflove wants to merge 1 commit into
Apageoflove wants to merge 1 commit into
Conversation
Second family of tigerless-labs#55. _store_provenance returned str(relative_to(...)), which embeds the win32 separator, and three read-side guards then refused those references: the membership gate in trace.read compares strings, and _legacy rejects any backslash outright. A store written on win32 could not trace its own provenance. Write side now stores relative_to(...).as_posix(). Read side normalises both comparison surfaces (the gate and _legacy) before validating, so stores written before the fix stay readable. The security posture is unchanged: normalisation happens before the PurePosixPath guards, so traversal via either separator is still caught by the dot-part check, pinned by a dedicated test. On real win32: test_no_provenance_and_legacy_excerpt goes red to green; the three remaining progressive_trace failures on this machine are symlink-privilege setup errors (WinError 1314), not separator issues. Full unit suite 482 passed / 21 failed against the 481/22 pre-fix state, failure-set diff exactly one: the target test.
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Fixes the second family of #55 (one correction to that issue below).
_store_provenance returned str(relative_to(...)), which embeds the win32 separator, and three read-side guards then refused those references: the membership gate in trace.read compares strings (an explicit --pointer must appear in the record's provenance), and _legacy rejects any backslash outright as part of its only-stored-excerpt-files stance. Net effect: a store written on win32 could not trace its own provenance, the CLI exiting with not a range cited by this memory or unsupported provenance depending on entry point.
The write side now stores relative_to(...).as_posix(), so new references are forward-slash on every platform. The read side normalises both comparison surfaces, the gate and _legacy, before any validation runs, so stores written before the fix stay readable too.
The security posture is unchanged. Normalisation happens before the PurePosixPath guards, so traversal attempts via either separator are still caught by the dot-part check; a dedicated test pins that archive backslash dot-dot backslash escape.md is refused. One documented edge, same as #58: a pathological POSIX filename containing a literal backslash now normalises like a separator.
Verification, on real win32: test_no_provenance_and_legacy_excerpt goes red to green (verified by stashing the fix). Full unit suite 482 passed / 21 failed against the 481/22 pre-fix state, and the failure-set diff is exactly one test, the target. New tests: forward-slash stored references, a manufactured backslash reference tracing through the CLI, and the traversal pin; the traversal pin passes on the pre-fix code too, since the old code rejected every backslash, so it guards the new surface rather than driving red-green.
Correction to #55 while here: only one of the four progressive_trace failures was the separator bug; the other three fail on win32 at symlink setup (WinError 1314, no symlink privilege), a separate environment class, not separators.