BTW, is there any pre-processed file with extension .source which will be converted to sql file for testing? Should we also consider these auto-generated sql files? If so, maybe we should move these intermediate files to a temp dir, so that we can distinguish them from normal sql files.
On Tue, Jul 12, 2016 at 1:18 PM, Paul Guo <[email protected]> wrote: > I would leave .diff file at least unmasked so that people could quickly > find what cases failed. > > 2016-07-12 11:38 GMT+08:00 陶征霖 <[email protected]>: > > > Besides .out, there also exists .diff, binary which should be ignored. > > > > 2016-07-12 11:32 GMT+08:00 Paul Guo <[email protected]>: > > > > > I'd mask the output files after running feature tests in .gitignore and > > > clean it up after running "make clean" or "make distclean". Anyone has > > any > > > suggestions? Thanks. > > > > > > diff --git a/src/test/feature/.gitignore b/src/test/feature/.gitignore > > > index a2e6bd4..c7332b2 100644 > > > --- a/src/test/feature/.gitignore > > > +++ b/src/test/feature/.gitignore > > > @@ -1 +1,2 @@ > > > doc/ > > > +**/*.out > > > > > > diff --git a/src/test/feature/Makefile b/src/test/feature/Makefile > > > index adc6acc..e0985d1 100644 > > > --- a/src/test/feature/Makefile > > > +++ b/src/test/feature/Makefile > > > @@ -35,6 +35,7 @@ doc: > > > doxygen doxygen_template > > > > > > clean distclean: sharelibclean > > > + find . -type f -name "*.out" |xargs rm -f > > > $(RM) feature-test > > > $(RM) feature-test.dSYM > > > > > >
