svn commit: r215070 - head/tools/tools/nanobsd

[Garrett Cooper]

On Tue, Nov 9, 2010 at 10:26 PM, Warner Losh <imp at bsdimp.com> wrote:
>  On 11/09/2010 23:10, Garrett Cooper wrote:
>>
>> On Tue, Nov 9, 2010 at 8:41 PM, Warner Losh<imp at bsdimp.com>  wrote:
>>>
>>>  On 11/09/2010 18:25, Garrett Cooper wrote:
>>>>
>>>> On Tue, Nov 9, 2010 at 4:56 PM, Warner Losh<imp at freebsd.org>    wrote:
>>>>>
>>>>> Author: imp
>>>>> Date: Wed Nov 10 00:56:44 2010
>>>>> New Revision: 215070
>>>>> URL: http://svn.freebsd.org/changeset/base/215070
>>>>>
>>>>> Log:
>>>>>  Build make.conf when the world is not selected to build, but the
>>>>>  kernel is.
>>>>>
>>>>>  PR:           151696
>>>>>  Submitted by: lev@
>>>>>
>>>>> Modified:
>>>>>  head/tools/tools/nanobsd/nanobsd.sh
>>>>>
>>>>> Modified: head/tools/tools/nanobsd/nanobsd.sh
>>>>>
>>>>>
>>>>> ==============================================================================
>>>>> --- head/tools/tools/nanobsd/nanobsd.sh Wed Nov 10 00:39:42 2010
>>>>>  (r215069)
>>>>> +++ head/tools/tools/nanobsd/nanobsd.sh Wed Nov 10 00:56:44 2010
>>>>>  (r215070)
>>>>> @@ -917,6 +917,9 @@ else
>>>>>  fi
>>>>>
>>>>>  if $do_kernel ; then
>>>>> +       if ! $do_world ; then
>>>>> +               make_conf_build
>>>>> +       fi
>>>>>        build_kernel
>>>>>  else
>>>>>        pprint 2 "Skipping buildkernel (as instructed)"
>>>>
>>>>     Funny that it overwrites __MAKE_CONF, but not SRCCONF (seems like
>>>> the latter would infect a system more than the former *shrugs*); I
>>>> think I understand why it's not set in this script though, because
>>>> IIRC src.conf predates nanobsd.
>>>>     I set my core variables and features in src.conf so I think it'd
>>>> probably be better to block that out too.
>>>>     Should I provide a patch for this item?
>>>
>>> __MAKE_CONF is the root of all evil.  SRCCONF shouldn't affect anything
>>> if
>>> __MAKE_CONF is overriden.  If it does, then point me at the code in src
>>> that
>>> pulls it in?
>>
>> $ cat test-srcconf.mk
>> .include<bsd.own.mk>
>> $ tail -n 1 /etc/src.conf
>> .error "Hallo old chap!"
>> $ make __MAKE_CONF=/dev/null -f test-srcconf.mk
>> "/etc/src.conf", line 64: "Hallo old chap!"
>>
>>     They're completely different methods of supplying different bits of
>> info :).
>
> At least the patch looks trivial..

    I realize that nanobsd aims to block that stuff off, but since
it's executed as a script, does it make more sense to supply values
via the CLI or something, but default to __MAKECONF=/dev/null and
setting _WITHOUT_SRCCONF?
    The commit you made looks good though for taking care of src.conf though.
Thanks!
-Garrett