Improve sjson package fuzzing #3071
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
Previously, only FuzzDocument was reaching the marshal+unmarshal steps due to the document type assertion. If it wasn't a documentType, the rest of the fuzzing function was skipped. Since data validation is available only via `types.Document.ValidateData`, a helper `isValidDocumentData` function is introduced. It may go away in the future. But for now, it allows all sjson types to pass the validation. The other change is related to the way we fuzz sjson. Instead of generating both values and schemes at the same time, focus on one thing at the time: a fixed scheme against the generated and fixed documents against the generated schemes. This approach can be further improved by the increased number of relevant corpus seed entries, but that's a different issue (#3067). To answer the question is it any better: yes, it is. The old approach skipped any cases where either of those two was invalid; it was a very rare occasion for a fuzz runner to generate a valid JSON for both document and scheme. It's even more rare to get both of them even remotely compatible. Updates #1273
Codecov Report
Additional details and impacted files@@ Coverage Diff @@
## main #3071 +/- ##
==========================================
- Coverage 76.48% 76.44% -0.05%
==========================================
Files 386 386
Lines 21165 21165
==========================================
- Hits 16188 16179 -9
- Misses 4050 4060 +10
+ Partials 927 926 -1
... and 12 files with indirect coverage changes
Flags with carried forward coverage won't be shown. Click here to find out more. |
quasilyte
changed the title
AlekSi
approved these changes
Jul 19, 2023
| - go test -run=XXX -fuzz=FuzzArrayWithFixedSchemas -fuzztime={{.FUZZ_TIME}} ./internal/handlers/sjson/ | ||
| - go test -run=XXX -fuzz=FuzzArrayWithFixedDocuments -fuzztime={{.FUZZ_TIME}} ./internal/handlers/sjson/ | ||
| - go test -run=XXX -fuzz=FuzzDocumentWithFixedSchemas -fuzztime={{.FUZZ_TIME}} ./internal/handlers/sjson/ | ||
| - go test -run=XXX -fuzz=FuzzDocumentWithFixedDocuments -fuzztime={{.FUZZ_TIME}} ./internal/handlers/sjson/ |
There was a problem hiding this comment.
They were sorted alphabetically, please move them back
| func addRecordedFuzzDocs(f *testing.F, needDocument, needSchema bool) int { | ||
| // We're trying to use that corpus with our hopes set high, | ||
| // but chances are, it will still be 0 extra documents. | ||
| // See #3067 for more details. |
There was a problem hiding this comment.
We always use full-issue URLs and TODO markers for such cases
There was a problem hiding this comment.
It is still like that in the main branch: https://github.com/FerretDB/FerretDB/blob/main/internal/handlers/sjson/sjson_test.go#L196
Comment on lines
+296
to
+297
| // if there was no error, | ||
| // check that the documents match each other |
There was a problem hiding this comment.
That comment duplicates the code, we should improve it:
Lines 226 to 229 in fbe7748
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.
Description
Previously, only FuzzDocument was reaching the marshal+unmarshal steps due to the document type assertion. If it wasn't a documentType, the rest of the fuzzing function was skipped.
Since data validation is available only via
types.Document.ValidateData, a helperisValidDocumentDatafunction is introduced. It may go away in the future. But for now, it allows all sjson types to pass the validation.The other change is related to the way we fuzz sjson. Instead of generating both values and schemes at the same time, focus on one thing at the time: a fixed scheme against the generated and fixed documents against the generated schemes.
This approach can be further improved by the increased number of relevant corpus seed entries, but that's a different issue (#3067).
To answer the question is it any better: yes, it is. The old approach skipped any cases where either of those two was invalid; it was a very rare occasion for a fuzz runner to generate a valid JSON for both document and scheme. It's even more rare to get both of them even remotely compatible.
Closes #1273
Readiness checklist
task all, and it passed.@FerretDB/core), Labels, Project and project's Sprint fields.