Re: [PATCH] coverage: test semantics of "dummy" per-object flags

[Stefano Lattarini]

On Sunday 23 January 2011, Ralf Wildenhues wrote:
> Hi Stefano,
> 
> * Stefano Lattarini wrote on Sun, Jan 23, 2011 at 11:28:25AM CET:
> > On Saturday 22 January 2011, Ralf Wildenhues wrote:
> > > * Stefano Lattarini wrote on Sat, Jan 22, 2011 at 06:14:36PM CET:
> > > > On Saturday 22 January 2011, Ralf Wildenhues wrote:
> > > > * tests/specflg-dummy.test: Check that we can "fool" automake
> > > > into thinking that per-object CFLAGS are used by simply doing
> > > > `foo_CFLAGS = $(AM_CFLAGS)', even if AM_CFLAGS is undefined.
> > > 
> > > I don't think it is so much "fooling", as the semantics are quite
> > > clearly defined in the manual, and per-target flags are documented
> > > in several places.  See 'Renamed Objects' and 'Objects created both
> > > with libtool and without' for quite explicit mention of these
> > > semantics.
> > >
> > Ah, but there I only see examples of "real" per-target flags, while
> > my test is meant to check that even "dummy" ones triggers the use of
> > renamed objects.
> 
> What is a "dummy" one then?
> 
> If foo_CFLAGS is set, then it is a per-target flag.  It doesn't matter
> whether it is set to $(AM_CFLAGS) or -foo or anything else.  Well, it
> shouldn't matter at least.
>
Exactly, and the main point of the testcase is to check that it truly
doesn't matter :-)  Sorry if I failed to explain myself properly before.

> > What about this squash-in?
> 
> I'm fine with all your proposed changes.
> 
> Thanks,
> Ralf
> 

I've pushed the patch to maint now.

Thanks,
  Stefano