From mboxrd@z Thu Jan  1 00:00:00 1970
From: =?ISO-8859-1?Q?J=F6rg?= Krause <joerg.krause@embedded.rocks>
Date: Mon, 20 Jul 2015 22:48:14 +0200
Subject: [Buildroot] [autobuild.buildroot.net] Build results for
 2015-07-19
In-Reply-To: <20150720223851.68e8f706@free-electrons.com>
References: <20150720063017.2400B101C74@stock.ovh.net>
 <1437422115.1945.4.camel@embedded.rocks>
 <20150720223851.68e8f706@free-electrons.com>
Message-ID: <1437425294.1945.8.camel@embedded.rocks>
List-Id: <buildroot.busybox.net>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
To: buildroot@busybox.net

Dear Thomas Petazzoni,

On Mo, 2015-07-20 at 22:38 +0200, Thomas Petazzoni wrote:
> Dear J?rg Krause,
> 
> On Mon, 20 Jul 2015 21:55:15 +0200, J?rg Krause wrote:
> 
> > >          arm |                 libtirpc-0.3.1 | NOK | 
> > > http://autobuild.buildroot.net/results/80321de8d4c6c0a00439b07289
> > > a581
> > > 8962128b43/
> > 
> > This is odd:
> > # BR2_TOOLCHAIN_EXTERNAL_HAS_THREADS is not set
> > # libtirpc needs a toolchain w/ threads
> > 
> > -> why is libtirpc build anyhow?
> 
> Take the configuration, run menuconfig and exit, you will see:
> 
> warning: (BR2_PACKAGE_NFS_UTILS) selects BR2_PACKAGE_RPCBIND which 
> has unmet direct dependencies (BR2_TOOLCHAIN_HAS_THREADS && 
> !BR2_TOOLCHAIN_EXTERNAL_BLACKFIN_UCLINUX_2012R2 && BR2_USE_MMU)

I have done so, but I must missed the warning somehow. Thanks for the
hint!

> 
> The problem is that nfs-utils selects libtirpc without properly 
> caring
> for the thread dependency. This is caused by a patch from Maxime
> Hadjinlian that added a "select BR2_PACKAGE_RPCBIND" in
> nfs-utils/Config.in.
> 
> Since we probably don't care much about using nfs-utils in a
> non-threaded scenario, maybe we could simply do:
> 
> comment "nfs-utils needs a toolchain w/ threads"
>         depends on BR2_USE_MMU
>         depends on !BR2_TOOLCHAIN_HAS_THREADS
> 
> config BR2_PACKAGE_NFS_UTILS
>         bool "nfs-utils"
>         depends on BR2_TOOLCHAIN_HAS_THREADS # libtirpc, rpcbind
>         select BR2_PACKAGE_LIBTIRPC if !BR2_TOOLCHAIN_HAS_NATIVE_RPC
>         select BR2_PACKAGE_RPCBIND # runtime
> 
> We could try to be smarter and use portmap when no threads are
> available, but really this is too much complexity for such a corner
> case.
> 
> Care to send a patch?

Thanks for clarification! I'll prepare a patch.

Best regards
J?rg Krause