Improve safety of BIO api #751
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
The bio extension methods _erroneously_ assume that the buffer is an input-buffer; this is blatantly false with delimited input buffers, and also alternative implementations of buffers if someone is so inclined. This fixes the problem by checking the exact type and only apply the dangerous stuff if it is an exact instance; otherwise fallback to generic methods. This applies for the inline methods and also read-char/peek-char. This in turn necessitated the introduction of a `put-back` method to the BufferedReader interface, otherwise generic read-char and peek-char implementations are impossible to implement correctly. I reckon this will be generally useful for parsers, so it's not abstraction leakage. It also fixes the issue with get-buffer-output-* not flushing the buffer before retrieving the output.
So that we can't cast from one to the other, which would be a disaster as the file is open in different directions.
fare
reviewed
Aug 5, 2023
- add notes for put-back behavior. - add thread safety notes for general Reader/Writer behavior.
fare
approved these changes
Aug 5, 2023
| (bio-read-u8 bio)))))) | ||
| (if (is-input-buffer? reader) | ||
| (let () | ||
| (declare (not interrupts-enabled)) |
There was a problem hiding this comment.
maybe a without-interrupts macro would be useful?
There was a problem hiding this comment.
nah, i prefer to see the declaration.
There was a problem hiding this comment.
I dont want to encourage people to disable interrupts right and left.
| (begin | ||
| (put-back! buf 0 previous-input) | ||
| (set! (&input-buffer-rhi bio) prevlen) | ||
| (void)))))))) |
There was a problem hiding this comment.
That makes consecutive put-backs expensive. Maybe worth noting in the documentation?
There was a problem hiding this comment.
yeah, thats why the list.
Maybe worth adding a note.
There was a problem hiding this comment.
nah, it is not really deterministic on how much it will cost; it is usually cheap, but sometimes it can be expensive.
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 bio extension methods erroneously assume that the buffer is an input-buffer; this is blatantly false with delimited input buffers, and also alternative implementations of buffers if someone is so inclined.
This fixes the problem by checking the exact type and only apply the dangerous stuff if it is an exact instance; otherwise fallback to generic methods. This applies for the inline methods and also read-char/peek-char.
This in turn necessitated the introduction of a
put-backmethod to the BufferedReader interface, otherwise generic read-char and peek-char implementations are impossible to implement correctly. I reckon this will be generally useful for parsers, so it's not abstraction leakage.It also fixes the issue with get-buffer-output-* not flushing the buffer before retrieving the output.
And while at it, it separates input from output file io so that we can't cast one to the other; the file is open in different directions.