Hi, On 2022-03-30 12:34:34 -0400, Tom Lane wrote: > Andres Freund <and...@anarazel.de> writes: > > Unless somebody speaks in favor of doing this across branches, I'd just go > > for > > HEAD. > > +1 for HEAD only, especially if we think we might change it some more > later. It seems possible this might break somebody's tooling if we > drop it into minor releases.
Yea. I certainly have written scripts that parse check-world output - they didn't break, but... > One refinement that comes to mind as I look at the patch is to distinguish > between "check" and "installcheck". Not sure that's worthwhile, but not > sure it isn't, either. As it's just about "free" to do so, I see no reason not to go for showing that difference. How about: echo "+++ (tap|regress|isolation) [install-]check in $(subdir) +++" && \ I see no reason to distinguish the PGXS / non-PGXs tap installcheck cases? Random aside: Am I the only one bothered by a bunch of places in Makefile.global.in quoting like $(MAKE) -C '$(top_builddir)' DESTDIR='$(abs_top_builddir)'/tmp_install install >'$(abs_top_builddir)'/tmp_install/log/install.log 2>&1 and rm -rf '$(CURDIR)'/tmp_check && etc yielding commands like: make -C '.' DESTDIR='/home/andres/build/postgres/dev-assert/vpath'/tmp_install install >'/home/andres/build/postgres/dev-assert/vpath'/tmp_install/log/install.log 2>&1 and rm -rf '/home/andres/build/postgres/dev-assert/vpath/contrib/test_decoding'/tmp_check & Greetings, Andres Freund
