Repair Quarkus Vert.x benchmark#10736
Conversation
|
The following frameworks were updated, pinging maintainers: |
|
Not only that, they need to calculate the length in each request !!! EDIT: or will be marked as stripped !!! |
|
Another big problem is the use of constant for the length header that is not permitted. |
Ongres SCRAM was missing (required for authentication). Signed-off-by: Thomas Segismont <tsegismo@ibm.com>
d6d048a to
cf47945
Compare
|
@joanhey thanks for taking a look as well Regarding your comments about the length header, I'm not a maintainer of this benchmark, consequently not in a position to discuss whether this should be changed. The scope of this PR is repairing a broken benchmark, so not related to your issue anyway. |
|
Without the header change, this PR will not be merged. |
|
Except for plaintext in quarkus-reactive-routes-pgclient and quarkus-vertx, no Quarkus benchmark computes the content length beforehand. My understanding is that this is compliant with the rule as explained in #10611 (comment) @joanhey @msmith-techempower if I didn't get the statement wrong, can we get this PR merged after one of the maintainers has approved the PR? |
|
@tsegismont I have to agree with @joanhey - we have been really pushing to be more equitable about the stripped implementations. Without the header changes or marking the implementation as 'Stripped', we won't merge. I recommend simply ensuring that the headers are computed on every request. |
|
Thanks for your answer @msmith-techempower
So, this is no longer true? |
|
If you want to play in the bench, you need to be a fair player !! If you find another fw than make the same, please mark it as Stripped. |
If this is a general statement, I fully agree.
Sorry I'm not comfortable with sending PRs to other participants marking their implementation as stripped. I don't believe this would cultivate a positive competitive spirit. |
Follows-up on #10715
The Ongres SCRAM dependency is required for authentication with the database.