Posted on 1 Comment

MySQL coverity reports

A user just alerted me to an article in LWN where community editions of various OSS products are compared with certified editions. MySQL’s certified edition came out with 0 defects. According to the article, the report does have a defect count for the community edition. That sounded curious to me, since the certified edition is based on the community edition of a major version, and then runs on a more conservative release cycle that only applies critical and security bugs.

Last year a Coverity report came out where a number of defects were found, which were promptly fixed in our codebase and so our community edition should be clean in this respect.

Interestingly, the reported new numbers match up with that earlier report (0.224 defects per thousand lines of code vs. 1 defect per 4000 lines of code) so what I presume happened is that the new report didn’t actually re-test the community edition but used the earlier numbers, whereas they did perform a test on a recent certified edition. And that would explain the discrepancy.

Oh, for reports, numbers and statistics… and the misconceptions they cause in the world…

Posted on 1 Comment

1 thought on “MySQL coverity reports

  1. I read it on the web…it MUST be true!

    (but it was SCIENTIFIC and junk!)

Comments are closed.