fix: update BlobReadChannelV2 handling to correctly restart for decompressed object #1867
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Outdated suggestions cannot be applied.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.
When downloading bytes from gcs and decompressing them, a restart of the stream needs to pickup from the offset within the compressed bytes not the decompressed bytes.
Prior to this change http-client was automatically applying gzip decompression to the stream it returns to us thereby causing our tracking to be off.
This change updates our interaction with http client to always request the raw bytes without any transform applied to them, we then at a higher level can handle whether gzip decompression needs to be plumbed in.
Fix up a couple channel closed state subtleties when buffering is used with decompression.
Add a new test leveraging the testbench which forces a broken stream on some compressed object bytes. NOTE This test depends on the next release of testbench and will be failing until we get that release.