Re: [lmi] wxmsw-2.9.0 (svn trunk) anomaly

[Greg Chicares]

On 2009-03-14 19:01Z, Vadim Zeitlin wrote:
> On Sat, 14 Mar 2009 18:41:39 +0000 Greg Chicares <address@hidden> wrote:
> 
> GC> On 2009-03-14 17:42Z, Vadim Zeitlin wrote:
> GC> > On Thu, 12 Mar 2009 12:54:47 +0000 Greg Chicares <address@hidden> wrote:
> GC> > 
> GC> > GC> [...] a couple hundred lines like this:
> GC> > GC>   stl_uninitialized.h:82: warning: '__cur' might be used 
> uninitialized in this function
> GC> > GC> that I don't believe we can do anything about.
> GC> > 
> GC> >  Wouldn't using -isystem for this directory help?
> GC> 
> GC> I suppose it should help. But I also would suppose that the right
> GC> place to specify '-isystem' would be either in the autotools files,
> GC> or in the compiler's specs file, and I don't want to touch either.
> GC> Just to be clear, I see this stuff only when building wx itself with
> GC> some variant of './configure && make', but never when building lmi.
> 
>  You surely know this already but this is just because LMI uses
> -Wno-uninitialized.

I suspect the cause might be different.

I see these messages only when building wx itself, in particular
by using lmi's 'install_wx.make' to build it, but that file doesn't
specify '-Wno-uninitialized', and no other lmi file is involved in
building wx itself. I save '>some_log_file 2>&1' output from every
build, and '-Wno-uninitialized' doesn't occur in my wx build logs,
where I see these diagnostics. Conversely, '-Wno-uninitialized'
occurs throughout my lmi build logs, but they don't contain any of
these diagnostics.