2013/6/29 J. Liles <[email protected]>:
>
>
> On Sat, Jun 29, 2013 at 10:59 AM, Karl Lindén <[email protected]> wrote:
>>
>> 2013/6/29 J. Liles <[email protected]>:
>> >
>> >
>> >
>> > On Sat, Jun 29, 2013 at 3:16 AM, Karl Lindén <[email protected]>
>> > wrote:
>> >>
>> >> Hi,
>> >>
>> >> I have made a patch for NTK's build system that makes waf take a
>> >> --no-ldconfig option which prevents the build system from calling
>> >> /sbin/ldconfig. The previous behaviour is preserved when not giving
>> >> --no-ldconfig option to waf.
>> >>
>> >> This patch is needed in Gentoo because our package manager
>> >> automatically calls ldconfig and running ldconfig from the build
>> >> system generates an error (a so called sandbox violation).
>> >>
>> >> I have attached the patch to this mail. Can you please take a look at
>> >> it and also see if you can apply it upstream?
>> >>
>> >> Thanks in advance,
>> >> Karl
>> >
>> >
>> > Are you not using DESTDIR? The running of ldconfig should already be
>> > disabled when DESTDIR is used (which is what packagers typically do).
>> >
>> The eclass appends --destdir="${D}", where D is the destination
>> directory, in the install phase. Everything is installed into the
>> correct place, but DESTDIR does not seem to be set (ldconfig is run).
>> If --destdir is given, why is not DESTDIR automatically set?
>>
>>
> I suppose that's a bug in waf. I didn't even notice that the --destdir
> option existed. Do you get the intended result if you just set DESTDIR in
> the environment instead of using that option?
>
Yes, it works with both DESTDIR in conjunction with --destdir and only
DESTDIR. The bug might be solved in some later version of waf, but I
can't find the bug it in their issue tracker. Anyway, the shipped one
is version 1.7.5 and the most recent one is 1.7.11. Would it be worth
the work to update the waf binary?
