Mercurial > hgsubversion
view Makefile @ 1541:adba152a5177
test_util: make developer warnings cause our tests to fail
Previously hgsubversion didn't get any help from develwarn calls,
because we don't use core's testrunner. The upstream configitem
refactor was easier to hunt down using the develwarn being a failure,
which prompted this code. It also exposed several locking correctness
bugs, so it feels like it's worth keeping around long term.
We don't bother with develwarn on anything so old that we don't try to
use smartset. The smartset deprecation was not handled gracefully, so
hg complains for many versions without actually breaking anything, and
it'd pointlessly complicate our code to handle smartset living in
multiple places between when the warning was introduced and the
present.
author | Augie Fackler <raf@durin42.com> |
---|---|
date | Sun, 29 Oct 2017 09:07:12 -0400 |
parents | 055f9254d790 |
children |
line wrap: on
line source
# Makefile for testing hgsubversion PYTHON=python .PHONY: all check check-demandimport check-subvertpy check-swig all: @echo "Use the following commands to build and install hgsubversion:" @echo @echo "$$ cd $(PWD)" @echo "$$ $(PYTHON) ./setup.py install" @echo @exit 1 check: check-demandimport check-subvertpy check-swig check-demandimport: # verify that hgsubversion loads properly without bindings, but fails # when actually used ! LC_ALL=C HGSUBVERSION_BINDINGS=none HGRCPATH=/dev/null \ hg --config extensions.hgsubversion=./hgsubversion \ version 2>&1 \ | egrep '(^abort:|failed to import extension)' LC_ALL=C HGSUBVERSION_BINDINGS=none HGRCPATH=/dev/null \ hg --config extensions.hgsubversion=./hgsubversion \ version --svn 2>&1 \ | egrep '(^abort:|failed to import extension)' check-subvertpy: $(PYTHON) tests/run.py --all --bindings=subvertpy check-swig: $(PYTHON) tests/run.py --all --bindings=swig