|From:||Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>|
|To:||Victor Drobny <v(dot)drobny(at)postgrespro(dot)ru>|
|Cc:||Thomas Munro <thomas(dot)munro(at)enterprisedb(dot)com>, Aleksander Alekseev <a(dot)alekseev(at)postgrespro(dot)ru>, Pgsql Hackers <pgsql-hackers(at)postgresql(dot)org>|
|Subject:||Re: Red-Black tree traversal tests|
|Views:||Raw Message | Whole Thread | Download mbox|
> In the meantime, here's my version. Notable changes:
I went ahead and pushed this, with the removal of the preorder/postorder
code, so we can see if the buildfarm finds out anything interesting.
Feel free to continue to submit improvements though.
One thing that occurred to me is that as-is, this is entirely black-box
testing. It doesn't try to check that the tree actually satisfies the
RB invariants, which is something that is interesting for performance
reasons. (That is, the code could pass these tests even though it
produces an unbalanced tree with horrible performance.) Is it worth
adding logic for that? Not sure. It's not like we are actively
changing the RB code or have reason to think it is buggy.
regards, tom lane
|Next Message||Tom Lane||2017-09-10 18:39:40||relation mapping file checksum failure|
|Previous Message||Beena Emerson||2017-09-10 17:04:21||Re: increasing the default WAL segment size|