Re: PATCH: Unlogged tables re-initialization tests - Mailing list pgsql-hackers

From Michael Paquier
Subject Re: PATCH: Unlogged tables re-initialization tests
Date
Msg-id 20180302045947.GG2259@paquier.xyz
Whole thread Raw
In response to Re: PATCH: Unlogged tables re-initialization tests  (David Steele <david@pgmasters.net>)
Responses Re: PATCH: Unlogged tables re-initialization tests  (David Steele <david@pgmasters.net>)
List pgsql-hackers
On Thu, Mar 01, 2018 at 01:12:19PM -0500, David Steele wrote:
> But your point is well-taken.  No symlinks are used in this test so it
> *should* work.
>
> Michael, what do you think?

Perl's symlink() does not work on Windows.  It does not fly higher than
that, and that's the reason why a good chunk of the tests are skipped
for pg_basebackup.  If perl was to have a version of symlink() which
works, say with junction points, or if Windows was to have a sane
symlink implementation (or with [1]?), or if it was possible to create
junction points using an in-core implementation in perl, then those
tests could not be skipped. But it seems that none of those scenarios
have happened yet.

From what I read in your patch, it seems to me that this test should
work.  If they don't for whatever reason, your patch then does not give
a correct justification explaining why they should be skipped.

[1]: https://blogs.windows.com/buildingapps/2016/12/02/symlinks-windows-10/
--
Michael

Attachment

pgsql-hackers by date:

Previous
From: Amit Langote
Date:
Subject: Re: 2018-03 Commitfest Summary (Andres #3)
Next
From: Yoshimi Ichiyanagi
Date:
Subject: Re: [HACKERS][PATCH] Applying PMDK to WAL operations for persistent memory