Re: Smoke [5.9.0] 22310 FAIL(FM) cygwin_nt-5.0 1.5.7(0.109/3/2) (x86/1 cpu)

On Tue, Feb 24, 2004 at 07:46:24AM -0800, Yitzchak Scott-Thoennes <[email protected]> wrote:
> On Tue, Feb 24, 2004 at 03:54:31PM +0100, "Gerrit P. Haase" <[email protected]> wrote:
> > Hi Merijn,
> > 
> > > On Tue 24 Feb 2004 12:52, "H.Merijn Brand" <[email protected]> wrote:
> > >> On Tue 24 Feb 2004 11:09, "Gerrit P. Haase" <[email protected]> wrote:
> > >> > > Failed Test                 Stat Wstat Total Fail  Failed  List of Failed
> > >> > >
> > >> -------------------------------------------------------------------------------
> > >> > > ../ext/IPC/SysV/t/ipcsysv.t    1   256    16   32 200.00%  1-16
> > >> > > ../ext/IPC/SysV/t/msg.t        0    12    ??   ??       %  ??
> > >> > > ../ext/IPC/SysV/t/sem.t        0    12    ??   ??       %  ??
> > >> > > ../lib/ExtUtils/t/Install.t    0   139    ??   ??       %  ??
> > >> > > ../lib/IPC/SysV.t              1   256    16   32 200.00%  1-16
> > >> > > op/sysio.t                                39    1   2.56%  39
> > >> > > op/taint.t                     0    12   223  148  66.37%  150-223
> > >> > > 50 tests and 514 subtests skipped.
> > >> > > Failed 7/872 test scripts, 99.20% okay. 107/80975 subtests failed, 99.87% okay.
> > >> > 
> > >> > 
> > >> > For IPC function without cygipc, the cygserver needs to run during the
> > >> > test runs.
> > >> 
> > >> IIRC Yitzchak was about to change the test to skip the test if the server was
> > >> not running.
> 
> It might be a while before I have time.
>  
> > > Sorry for the speedy follow up
> > 
> > > On second though, how feasible is it to
> > 
> > > if (cygserver_running ()) {
> > >     do_tests ();
> > >     }
> > > elif (cygserver_installed ()) {
> > >     start_cygserver ();
> > >     do_tests ();
> > >     stop_cygserver ();
> > >     }
> > > else {
> > >     skip_tests ();
> > >     }
> > 
> > > or is this too much of a security risk?
> > 
> > 
> > Security?  I think that is not a problem on a developer box.
> > Would be nice to have this auto-detect feature IMO.
> 
> I don't know that it's worth the trouble of differentiating between having
> the server not installed and the server installed but not running (or the
> server installed and running but CYGWIN=server not set), so I'm not inclined
> to worry about it;  skipping the tests should be good enough.

In addition, would putting something to start and stop the server
into Test::Smoke would be helpful?