Tom Lane wrote: > Alvaro Herrera <alvhe...@2ndquadrant.com> writes: > > Tom Lane wrote: > >> All of the MSVC critters are failing at "make check". > > > Yeah, I noticed that, thanks. As far as I can see the only way to fix > > it is to install dummy_seclabel to run the core seclabel test. That > > doesn't seem smart; I think it'd be better to remove that part of the > > core seclabel test, and move the rest of the test to a new test in the > > dummy_seclabel module. > > Sounds good to me. The other parts of the core tests that depend on > contrib modules aren't exactly good models to follow.
Pushed; tests pass for me, let's see what buildfarm says. I think the input/ and output/ stuff is rather annoying. I tried to make dummy_seclabel an extension instead of a bare library, so that CREATE EXTENSION inside the .sql loads it easily. The problem there is that dummy_seclabel doesn't have any sql-callable function, so the module doesn't ever load. I guess we could create a dummy function which is there only to cause the module to get loaded ... I am in touch with Andrew about adding a new stage to buildfarm client so that animals will build src/test/modules by default. It should work fine everywhere except MSVC, I think. The issue with MSVC continues to be that Install.pm and Mkvcbuild.pm need to be in sync regarding what to build and what to install. I could use help from some MSVC-enabled developer there ... In the meantime, I'm going to rebase commit_ts to src/test/modules and get one bug I found there fixed, and will push that shortly too. -- Álvaro Herrera http://www.2ndQuadrant.com/ PostgreSQL Development, 24x7 Support, Training & Services -- Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-hackers