> [snip] I believe this is fixed as of 8.2 --- can you duplicate it
> there? (No, backpatching the fix is not practical.)
No, I was not able to duplicate it on 8.2, so I think it's fixed (given
that on 8.1 the errors are triggered almost 100% of the runs).
> How sure are you about that uninterlocked getChildTableName() thing?
> It's possible to get a failure complaining about duplicate type name
> instead of duplicate relation name during CREATE TABLE, if the timing
> is just right.
Not sure at all (I left it deliberately unsynchronized to go as fast as
it can even if it errors sometimes on duplicate tables), so that might
be an explanation. The error is a bit misleading though, or better
inconsistent: if I would have to detect the duplicate table error
condition in my code so that I can take corrective steps I would need to
look for 2 error types instead of 1 - if I only knew that I have to.
And BTW, I have seen something similar while creating temporary tables
which should not conflict even with the same table name I think...
Cheers,
Csaba.