Implement stateful, incremental skip #156
Merged
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
…even when fetching more bytes is required
There was a problem hiding this comment.
Copilot reviewed 3 out of 3 changed files in this pull request and generated no suggestions.
Comments skipped due to low confidence (1)
src/Nerdbank.MessagePack/MessagePackStreamingReader.cs:804
- The word 'self' should be 'this'.
// We don't actually expect to ever hit self point, since every code is supported.
This is a perf optimization for streaming scenarios where a converter needs to skip a structure but the buffer may not have the entire structure in memory. This optimization records how much of the structure has been skipped already so that the reader can advance, freeing memory, and resume skipping when we get more bytes to decode. In particular, this avoids the previous behavior that skipping had to be restarted from the beginning of the structure each time new bytes are brought in until the whole structure was in memory together. While this is theoretically a memory use improvement (for skipping scenarios only), this is mostly about avoiding CPU work to parse partial structures repeatedly until we get the whole thing in memory. With this change, at most one msgpack *token* (rather than whole structure) will be decoded more than once, and that limited to only once per fetch of new bytes. Closes #154
AArnott
changed the title
MessagePackStreamingReader.TrySkip to skip the whole structure …
There was a problem hiding this comment.
Copilot reviewed 5 out of 10 changed files in this pull request and generated 1 suggestion.
Files not reviewed (5)
- src/Nerdbank.MessagePack/net8.0/PublicAPI.Unshipped.txt: Language not supported
- src/Nerdbank.MessagePack/netstandard2.0/PublicAPI.Unshipped.txt: Language not supported
- samples/CustomConverters.cs: Evaluated as low risk
- src/Nerdbank.MessagePack/Converters/ObjectMapConverter`1.cs: Evaluated as low risk
- src/Nerdbank.MessagePack/Converters/ObjectMapWithNonDefaultCtorConverter`2.cs: Evaluated as low risk
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.
This is a perf optimization for streaming scenarios where a converter needs to skip a structure but the buffer may not have the entire structure in memory. This optimization records how much of the structure has been skipped already so that the reader can advance, freeing memory, and resume skipping when we get more bytes to decode.
In particular, this avoids the previous behavior that skipping had to be restarted from the beginning of the structure each time new bytes are brought in until the whole structure was in memory together.
While this is theoretically a memory use improvement (for skipping scenarios only), this is mostly about avoiding CPU work to parse partial structures repeatedly until we get the whole thing in memory. With this change, at most one msgpack token (rather than whole structure) will be decoded more than once, and that limited to only once per fetch of new bytes.
Closes #154