pcp
[Top] [All Lists]

Re: [pcp] pcp updates - fix build failure, tart up pmhostname

To: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Subject: Re: [pcp] pcp updates - fix build failure, tart up pmhostname
From: Nathan Scott <nathans@xxxxxxxxxx>
Date: Sun, 8 Sep 2013 18:39:11 -0400 (EDT)
Cc: pcp@xxxxxxxxxxx
Delivered-to: pcp@xxxxxxxxxxx
In-reply-to: <522CD94E.20806@xxxxxxxxxxxxxxxx>
References: <522CD94E.20806@xxxxxxxxxxxxxxxx>
Reply-to: Nathan Scott <nathans@xxxxxxxxxx>
Thread-index: 0CyqLy87pI2eIkeK90pQaKh22lTawg==
Thread-topic: pcp updates - fix build failure, tart up pmhostname

----- Original Message -----
> Changes committed to git://oss.sgi.com/kenj/pcp.git dev
> 
>  src/libpcp_pmcd/src/GNUmakefile |    2 -
>  src/pmhostname/pmhostname.c     |   79
>  +++++++++++++++++++++++++++++++++++-----
>  2 files changed, 72 insertions(+), 9 deletions(-)
> 
> commit f7efeb3216e3ba68c8cc3c1f0afd75c8e3cdc83a
> Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
> Date:   Mon Sep 9 06:06:02 2013 +1000
> 
>     libpcp_pmcd - build failure
>     
>     Wrong guard round install lines ... pmcd.stp is not created
>     unless $(ENABLE_PROBES) is true.

Yes - we need both ENABLE_PROBES (guarding pmcd.stp creation)
and the TARGET_OS check (systemtap is Linux-specific).

> commit 7f944c2f80e48130fe5f41e5405f1ad497feaf1a
> Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
> Date:   Mon Sep 9 06:04:23 2013 +1000
> 
>     pmhostname - add -D on the command line
>     
>     So I can expose diagnositics when it is not working.

These DESPERATE-level diagnostics should move into the libpcp
routines.  Particularly problematic is the exposure of internal
"typedef struct addrinfo __pmAddrInfo", which is conditional
on the presence of NSPR/NSS (so this change wont work for the
secure builds, I expect).

cheers.

--
Nathan

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