On Tue, Apr 23, 2024 at 8:05 AM Robert Haas <robertmhaas@gmail.com> wrote:
> I reworked the test cases so that they don't (I think) rely on
> symlinks working as they do on normal platforms.
Cool.
(It will remain a mystery for now why perl readlink() can't read the
junction points that PostgreSQL creates (IIUC), but the OS can follow
them and PostgreSQL itself can read them with apparently similar code.
I find myself wondering if symlinks should go on the list of "things
we pretended Windows had out of convenience, that turned out to be
more inconvenient than we expected, and we'd do better to tackle
head-on with a more portable idea". Perhaps we could just use a
tablespace map file instead to do our own path construction, or
something like that. I suspect that would be one of those changes
that is technically easy, but community-wise hard as it affects a
load of backup tools and procedures...)