Re: Regression tests vs existing users in an installation

<p dir="ltr"><br /> On 16 Jul 2016 12:59 pm, "Michael Paquier" <<a
href="mailto:michael.paquier@gmail.com">michael.paquier@gmail.com</a>>wrote:<br /> ><p dir="ltr">> Thanks for
doingthis.<p dir="ltr">+1<p dir="ltr">Though I might highlight this as the kind of issue that a bug tracker would help
avoidfalling through the cracks and make visible to newcomers.<br /><p dir="ltr">> I am -1 for dropping the tests.
Wecould just have a CFLAGS that adds<br /> > an elog(ERROR) in CreateRole and checks that the created role has a<br
/>> wanted prefix, or have a plugin that uses the utility hook to do this<br /> > filtering.<p dir="ltr">If we
makea hidden regression_test_safety GUC then we could have pg_regress enable it and have these specific tests disable
itexplicitly with comments on why it's safe.<p dir="ltr">It might even be handy for other people writing application
regressiontests depending on what other things it blocked.<p dir="ltr">A hook might even be possible to use the same
way.pg_regress would have to build and install a .so which might be tricky.