Re: [Stp-devel] [PATCH] client/stp_client.pl - minor fixes to resolve errors

[Mark W. <ma...@os...>]

On Tue, 7 Jun 2005 12:18:59 -0700 (PDT)
Judith Lebzelter <ju...@os...> wrote:

> 
> 
> On Tue, 7 Jun 2005, Mark Wong wrote:
> 
> > On Tue, 07 Jun 2005 08:44:52 -0700
> > Leann Ogasawara <oga...@os...> wrote:
> > 
> > > On Tue, 2005-06-07 at 08:05, Mark Wong wrote:
> > > > On Tue, 07 Jun 2005 07:39:43 -0700
> > > > Leann Ogasawara <oga...@os...> wrote:
> > > > 
> > > > > On Mon, 2005-06-06 at 15:54, Mark Wong wrote:
> > > > > > On Mon, 06 Jun 2005 14:34:37 -0700
> > > > > > Leann Ogasawara <oga...@os...> wrote:
> > > > > > 
> > > > > > > Minor fixes to stp_client.pl to resolve some errors we saw when
> > > > > > > testing.  Check return on the system_upgrades() call and only call
> > > > > > > clean_exit() for STP (not BRT).  I ran it through perltidy -pt=0 -sbt=0
> > > > > > > -bt=0 -bbt=1 -nsfs -lp -ce.  Also used 'monotone diff' to generate the
> > > > > > > patch.  Thanks,
> > > > > > 
> > > > > > I'm thinking it would be more appropriate to have a "remote" flag or
> > > > > > something of that nature rather than "BRT".  Can you do that instead?
> > > > > 
> > > > > We actually need the "BRT" flag because we use that to override some of
> > > > > the STP TestRequest.pm sub routines.  
> > > > 
> > > > Right, I understand that.  What I'm saying is that a more general
> > > > identifier than "BRT" should be used because it's more about how it's
> > > > run, not what project it's a part of, no?
> > > 
> > > I tend to disagree.  If we switch to a more general identifier like
> > > "remote" to describe how it is run, I don't see that being completely
> > > correct.  This framework is not going to strictly be running remotely,
> > > we are going to use this in house as well.
> > 
> > I prefer not completely correct over not descriptive at all. ;)  But I
> > thought this mode of running was outside the automated framework?  I
> > would argue that your description is more like "offsite".  But let's
> > confer with Judith about the extent of the changes.
> 
> The function of running remotely is identified by the configuration 
> parameter 'lite'.  The 'BRT' option is different in that it uses the BRT 
> Build sets and binary application sets, which STP does not use.

Ok, that makes sense.

Mark