[PATCH] test-phase-exchange: stabilize for Windows

[FUJIWARA Katsunori]

At Mon, 17 Jul 2017 23:18:53 -0400,
Matt Harbison wrote:
> 
> On Mon, 17 Jul 2017 11:05:27 -0400, FUJIWARA Katsunori  
> <foozy at lares.dti.ne.jp> wrote:
> 
> > At Sat, 15 Jul 2017 23:49:54 -0400,
> > Matt Harbison wrote:
> >>
> >> # HG changeset patch
> >> # User Matt Harbison <matt_harbison at yahoo.com>
> >> # Date 1500176793 14400
> >> #      Sat Jul 15 23:46:33 2017 -0400
> >> # Node ID 21904723f1ce9ea0fc6541ee48611ba3ef7128fb
> >> # Parent  389536aff376d32d38f13305021c127245d4126a
> >> test-phase-exchange: stabilize for Windows
> >>
> >> diff --git a/tests/test-phases-exchange.t b/tests/test-phases-exchange.t
> >> --- a/tests/test-phases-exchange.t
> >> +++ b/tests/test-phases-exchange.t
> >> @@ -1298,11 +1298,17 @@
> >>
> >>  #endif
> >>
> >> -Test that clone behaves like pull and doesn't
> >> -publish changesets as plain push does
> >> +Test that clone behaves like pull and doesn't publish changesets as  
> >> plain push
> >> +does.  The conditional output accounts for changes in the conditional  
> >> block
> >> +above.
> >>
> >>    $ hg -R Upsilon phase -q --force --draft 2
> >>    test-debug-phase: move rev 2: 0 -> 1
> >> +  test-debug-phase: move rev 3: 0 -> 1 (unix-permissions no-root !)
> >> +  test-debug-phase: move rev 7: 0 -> 1 (unix-permissions no-root !)
> >> +  test-debug-phase: move rev 8: 0 -> 1 (unix-permissions no-root !)
> >> +  test-debug-phase: move rev 9: 0 -> 1 (unix-permissions no-root !)
> >> +
> >
> > According to commit log of 4eec2f04a672, "(unix-permissions no-root !)"
> > marking means:
> >
> >   - this line is required, if "unix-permissions no-root" is satisfied
> >   - this line is optional, otherwise
> >
> > AFAIK, additional lines above appear:
> >
> >   - not on POSIX, on which "unix-permissions no-root" is satisfied
> >   - but on windows, on which it isn't
> >
> > Therefore, I expect this change to cause:
> >
> >   - failure on POSIX, because all required lines disappear
> >   - success on Windows, because all optional lines appear
> >
> > But in practice, this change doesn't cause failure on POSIX.
> >
> > For more simple example, I expect that test script below fails,
> > because "true" command generate no output, even though "foobar" output
> > is always expected ("true" feature is always available). But it
> > doesn't fail in practice.
> >
> >   ====================
> >     $ true
> >     foobar (true !)
> >   ====================
> >
> > What do I misunderstand about "(feature !)" ?
> 
> Nice catch.  I think I see what's going on here, and another bug around  
> (?).  I should have a fix tonight or tomorrow.
> 
> >
> > BTW, it seems bug of run-tests.py that "(unknown-feature !)" doesn't
> > cause failure.
> 
> Agreed.  I thought I remembered there being a test for a missing feature,  
> and it complaining in that case, so I thought this was covered.  How  
> should it fail?  Dropping the line seems too subtle.

I think that simple failure of run-tests.py itself as same as "#if
unknown" is enough, because it is obviously test script error.

If all "(feature !)" in test script are evaluated by TTest._hghave(),
it causes such kind of failure automatically. And your patch series
(V2) seems to do so, AFAIK.

> >
> >>    $ hg clone -q Upsilon Pi -r 7
> >>    test-debug-phase: new rev 0:  x -> 0
> >>    test-debug-phase: new rev 1:  x -> 0
> >> _______________________________________________
> >> Mercurial-devel mailing list
> >> Mercurial-devel at mercurial-scm.org
> >> https://www.mercurial-scm.org/mailman/listinfo/mercurial-devel
> 

-- 
----------------------------------------------------------------------
[FUJIWARA Katsunori]                             foozy at lares.dti.ne.jp