fix: The hash_cache fails with filenames including surrogates #492
Conversation
This file contains 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
mwiebe
force-pushed
the
hash-cache-surrogates
branch
3 times, most recently
from
755112c to
a4c1a41
Compare
Cases encountered in practice show that filenames can include unicode surrogates. This changes the hash_cache functionality to preserve them by using the "surrogatepass" handler for them. Another issue encountered is that S3 metadata only allows ASCII, but the code was placing unicode paths in it. * Modify the hash_cache sqlite schema to use a BLOB instead of TEXT for the file_path, and then explicitly encode/decode it using the 'utf-8' encoding and 'surrogatepass' error handler. * In the canonical json code, use the 'surrogatepass' error handler in the canonical file path sorting. Add the new test case data to the test manifest file. * In the error logging of the exception that revealed this bug, include the exception context to improve debugging of future similar issues. * Add a test case for the hash_cache based on an example that was failing in practice. * Add the "asset-root-json" metadata containing the path encoded to ASCII via a JSON string. If the file path is not ASCII, it adds this metadata instead. Signed-off-by: Mark Wiebe <[email protected]>
mwiebe
force-pushed
the
hash-cache-surrogates
branch
from
a4c1a41 to
7e8626d
Compare
|
epmog
reviewed
Nov 4, 2024
leongdl
reviewed
Nov 5, 2024
| return path.path.encode("utf-16_be") | ||
| # Use the "surrogatepass" error handler because filenames encountered in the wild | ||
| # include surrogates. | ||
| return path.path.encode("utf-16_be", errors="surrogatepass") |
There was a problem hiding this comment.
And I thought this was some magic keyword :)
https://docs.python.org/3/library/codecs.html
leongdl
reviewed
Nov 5, 2024
| @@ -527,7 +528,16 @@ def _upload_output_manifest_to_s3( | |||
| full_output_prefix, | |||
| f"{manifest_name_prefix}_output", | |||
| ) | |||
| metadata = {"Metadata": {"asset-root": root_path}} | |||
| metadata = {"Metadata": {"asset-root": json.dumps(root_path, ensure_ascii=True)}} | |||
There was a problem hiding this comment.
I will need to amend this to the manifest upload CLI since it has similar mechanisms.
leongdl
reviewed
Nov 5, 2024
| @@ -670,7 +671,10 @@ def _get_asset_root_from_s3( | |||
| except Exception as e: | |||
| raise AssetSyncError(e) from e | |||
|
|
|||
| return head["Metadata"].get("asset-root", None) | |||
| if "asset-root-json" in head["Metadata"]: | |||
There was a problem hiding this comment.
Same here, I need to port this to manifest download. (Will do after this PR is merged)
leongdl
approved these changes
Nov 5, 2024
epmog
approved these changes
Nov 5, 2024
This was referenced Nov 14, 2024
Closed
Merged
2 tasks
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.
Fixes: #481
What was the problem/requirement? (What/Why)
Cases encountered in practice show that filenames can include unicode surrogates. This changes the hash_cache functionality to preserve them by using the "surrogatepass" handler for them.
What was the solution? (How)
What is the impact of this change?
Jobs that could not be submitted, because their filenames include surrogates, will work now.
How was this change tested?
See DEVELOPMENT.md for information on running tests.
Yes
No
downloadorasset_syncmodules? If so, then it is highly recommendedthat you ensure that the docker-based unit tests pass.
Yes, but have not run these tests.
Was this change documented?
N/A
Does this PR introduce new dependencies?
Is this a breaking change?
No. In increases the table version of the sqlite hash cache, but that does not break compatibility.
Does this change impact security?
No
By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.