Andres Freund <andres@anarazel.de> writes:
> On 2019-04-24 10:13:09 -0400, Tom Lane wrote:
>> I can't say that I like 0001 at all. It adds a bunch of complication and
>> new failure modes (e.g., having to panic on chdir failure) in order to do
>> what exactly? I've not been following the thread closely, but the
>> original problem is surely just a dont-do-that misconfiguration. I also
>> suspect that this is assuming way too much about the semantics of getcwd
>> --- some filesystem configurations may have funny situations like multiple
>> paths to the same place.
> I'm not at all defending the conrete patch. But I think allowing
> relative paths to tablespaces would solve a whole lot of practical
> problems, while not meaningfully increasing failure modes.
I'm not against allowing relative tablespace paths. But I did not like
the chdir and getcwd-semantics hazards --- why do we have to buy into
all that to allow relative paths?
I think it would likely be sufficient to state plainly in the docs
that a relative path had better point outside $PGDATA, and maybe
have some *simple* tests on the canonicalized form of the path to
prevent obvious mistakes. Going further than that is likely to add
more problems than it removes.
regards, tom lane