GH-48545: [C++][Parquet][CI] Add more encodings to fuzzing seed corpus #48546
Conversation
github-actions
bot
added
Component: Parquet
Component: C++
awaiting review
pitrou
force-pushed
the
gh48545-pq-fuzz-encodings
branch
2 times, most recently
from
6712d9b to
34e293c
Compare
|
@github-actions crossbow submit fuzz |
|
Revision: 34e293c Submitted crossbow builds: ursacomputing/crossbow @ actions-4036a960d4
|
|
@adamreeve Would you like to take a look at this PR? |
github-actions
bot
added
awaiting committer review
pitrou
force-pushed
the
gh48545-pq-fuzz-encodings
branch
from
34e293c to
894ec95
Compare
|
@github-actions crossbow submit fuzz |
|
@adamreeve I've added INT96 data and also made a cosmetic improvement on generated filenames. |
|
Revision: 894ec95 Submitted crossbow builds: ursacomputing/crossbow @ actions-af8817413c
|
|
I'll probably merge this by the end of the day (Paris time) if there are no more comments, as I'm on vacation for two weeks after that and it would be nice to get the expanded seed corpus on OSS-Fuzz. |
wgtmac
left a comment
wgtmac
left a comment
There was a problem hiding this comment.
+1
Enjoy your coming vacation!
pitrou
force-pushed
the
gh48545-pq-fuzz-encodings
branch
from
894ec95 to
1619218
Compare
|
@github-actions crossbow submit fuzz |
|
Revision: 1619218 Submitted crossbow builds: ursacomputing/crossbow @ actions-8bf2d54a7b
|
|
After merging your PR, Conbench analyzed the 3 benchmarking runs that have been run so far on merge-commit 713c57a. There were no benchmark performance regressions. 🎉 The full Conbench report has more details. It also includes information about 98 possible false positives for unstable benchmarks that are known to sometimes produce them. |
3 participants
Rationale for this change
Ensure that the seed corpus contains all supported combinations of physical types and encodings, with varying amounts of nulls and distributions of values.
Also give more meaningful names to seed corpus files, to ease manual inspection (this won't affect the fuzzer since we rename those files to the sha256 of their contents).
Are these changes tested?
Manually and later by OSS-Fuzz.
Are there any user-facing changes?
No.