Ali Bahrami <Ali.Bahrami at Sun.COM> writes: > jw137282 wrote: >> Mark J. Nelson wrote: >>> You can pull a copy of the repo from >>> ssh://anon at hg.opensolaris.org//hg/scm-migration/scmtest, then cd into >>> src/legacy/scm/tooltest and read the README there. >>> >>> I don't think you need to worry about the TET-harnessed stuff, but >>> others may certainly complain otherwise. (Is that even populated?) >> >> You can ignore the tet tests. >> >>> The abstraction of the comment checking should be tested effectively by >>> the cddl check tests, you're welcome to add mapfilechk tests if you're >>> so inclined. I don't think that should block your integration, but >>> again, others may speak up... >>> >> >> I think you should add a test. It's easy. >> >> Jim > > > I pulled over scmtest and ran it, and it did in fact pick up > something that I had changed. The cddlchk test relies on > Cddl.py having a variable Cddl.CmntChrs, and in my changes, I > had refactored that into CmtBlk.py (and so, CmtBlk.CmntChrs). > > So I added Cddl.CmntChrs back (taking its value from CmtBlk.py), > and now all the tests pass without modification. I'm very glad I > ran these before integration...
I'm confused by that. If you mean the *test* relies on that, surely fixing the test is the right thing to do? If Cddl relies on that, I'd assume it was using your CmtBlk code now. > I looked into adding a mapfilechk test, and it is truly very > easy. However, the result is a cut down copy of the cddlchk > test that does not actually test anything new. The reason for > this is that Cddl.py and Mapfile.py are now one liners that > reference CmdBlk.py (which is where all the guts of Cddl.py > hve moved). Hence, testing cddlchk really does fully test > mapfilechk as well. So in that case, rename the cddl tests to indicate they test the cmntblk stuff, and comment that what it covers is the implementation of both cddl and mapfile? -- Rich
