pcp
[Top] [All Lists]

Re: [pcp] pcp updates

To: Nathan Scott <nathans@xxxxxxxxxx>
Subject: Re: [pcp] pcp updates
From: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Thu, 25 Aug 2011 17:09:53 +1000
Cc: pcp@xxxxxxxxxxx
In-reply-to: <1021648368.198009.1314254233859.JavaMail.root@xxxxxxxxxxxxxxxxxxxxxx>
References: <1021648368.198009.1314254233859.JavaMail.root@xxxxxxxxxxxxxxxxxxxxxx>
On Thu, 2011-08-25 at 16:37 +1000, Nathan Scott wrote:
> 
> ----- Original Message -----
> > On Thu, 2011-08-25 at 13:10 +1000, Nathan Scott wrote:
> > > ...
> > >
> > > > src/libpcp/src/logutil.c | 8
> > >
> > > Instead of the gcc-specific hack, is it not possible to cast
> > > the returned value to (void)? I've seen this done in some
> > > other places, so here: "(void)strtol(q+1, &end, 10);" and if
> > > that works toss that big comment and local "tmpl" variable?
> > 
> > The gcc pixies refuse to let (void) cast over-ride the warning if the
> > __attribute(()) is set on the function ... there are long mail wars
> > between the gcc mob, the glibc mob and everyone else.
> > 
> > I played with this a lot, and this is the only way I've found to make
> > the warnings go away on all platforms.
> 
> OK.  We should check with Max, I thought he was still building Solaris
> bits with the native (non-gcc) compiler?  Mark mentioned he's away on
> holiday atm though.

If there is a non-gcc build that is needed, I'll rework my change to use
the (void) cast approach for non-gcc builds (sensible compilers should
do the right thing there) ... Max (or anyone else), please let me know
what non-gcc compilers are being used to build pcp.

> > ps please make sure you pull in my last round of changes ... there are
> > a couple of additional tweaks needed for the configure.in changes.
> 
> Saw that mail - git pull is finding no new changes there though (in dev
> branch)...?

Overlapping git transactions and email here it seems ... my last commit
is in the official tree.


<Prev in Thread] Current Thread [Next in Thread>