Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 24 Jan 2011 10:33:21 +0000 (UTC)
From:      "Bjoern A. Zeeb" <bz@FreeBSD.org>
To:        Warner Losh <imp@bsdimp.com>
Cc:        svn-src-head@FreeBSD.org, svn-src-all@FreeBSD.org, src-committers@FreeBSD.org, Garrett Cooper <gcooper@FreeBSD.org>
Subject:   Re: svn commit: r217733 - in head: . share/man/man7
Message-ID:  <20110124095030.Q3489@maildrop.int.zabbadoz.net>
In-Reply-To: <4D3D1AC6.7030003@bsdimp.com>
References:  <201101222257.p0MMvSZY007310@svn.freebsd.org> <AANLkTimmiLVVQ3y3-tbWG41UDaqQpE51BapKuDdxZP0O@mail.gmail.com> <20110123104812.A3489@maildrop.int.zabbadoz.net> <4D3D1AC6.7030003@bsdimp.com>

next in thread | previous in thread | raw e-mail | index | archive | help
On Sun, 23 Jan 2011, Warner Losh wrote:

> On 01/23/2011 03:51, Bjoern A. Zeeb wrote:
>> On Sat, 22 Jan 2011, Garrett Cooper wrote:
>> 
>>> On Sat, Jan 22, 2011 at 2:57 PM, Bjoern A. Zeeb <bz@freebsd.org> wrote:
>>>> Author: bz
>>>> Date: Sat Jan 22 22:57:28 2011
>>>> New Revision: 217733
>>>> URL: http://svn.freebsd.org/changeset/base/217733
>>>> 
>>>> Log:
>>>>  Properly document what the top-level `make tinderbox` does.
>>>>
>>>>  Reviewed by:  jmallett
>>>>  MFC After:    3 days
>>>>  X-MFC:                build.7 only
>>>
>>>    As much as I like make universe and tinderbox, for the purpose of
>>> development testing would it make sense if there was a tinderbox
>>> target that actually died on its first failure? tinderbox and universe
>> 
>> I think not really; if you want to catch errors, you want all of them
>> and not just the first broken kernel if there are 12.
>> 
>>> seems to be useful when dealing with known breakage or when executing
>>> periodic builds, but I don't know think it's as effective with
>>> development testing, as for some of the stuff I just want to build on
>>> something other than my host arch and another arch that isn't x86 (say
>>> arm, mips, powerpc) just to make sure that I didn't break anything
>>> endian wise (if the code is that nasty).
>>
>>     make universe TARGETS=arm
>> or
>>     make universe TARGETS="i386 amd64 sparc64 ia64"
>> 
>> works perfectly well (again).  Even
>>
>>     make universe TARGETS=mips TARGET_ARCH=mips64eb KERNCONFS=XLR64
>> 
>> should be perfectly fine.
>
> Except, it isn't fine.  That's a horrible interface.

Right.  But it works perfectly fine for most cases, can be scripted and
is a tool for developers and automated (full) testing, which I think has
been underutilized the last years.


> It doesn't work if TARGETS="i386 mips".

.. and you set TARGET_ARCH=mips64eb. Well yeah.  That never worked and is
kind of obvious given boolean algebra.  But that's not what you are
saying as the "doesn't work" not means it will instantly fail but
more means that it ignores the TARGET_ARCH in that case, right?
Not the result I would have expected btw. but probably better than what
I expected:)

make -s -j8 universe __MAKE_CONF=/dev/null TARGETS="i386 mips" TARGET_ARCH=mips64eb
--------------------------------------------------------------
>>> make universe started on Mon Jan 24 10:23:00 UTC 2011
--------------------------------------------------------------
>> mips started on Mon Jan 24 10:23:00 UTC 2011
>> i386 started on Mon Jan 24 10:23:00 UTC 2011
>> mips.mipseb buildworld started on Mon Jan 24 10:23:00 UTC 2011
>> mips.mips64eb buildworld started on Mon Jan 24 10:23:00 UTC 2011
>> mips.mips64el buildworld started on Mon Jan 24 10:23:00 UTC 2011
>> i386.i386 buildworld started on Mon Jan 24 10:23:00 UTC 2011
>> mips.mipsel buildworld started on Mon Jan 24 10:23:00 UTC 2011
^C

-- 
Bjoern A. Zeeb                                 You have to have visions!
         <ks> Going to jail sucks -- <bz> All my daemons like it!
   http://www.freebsd.org/doc/en_US.ISO8859-1/books/handbook/jails.html



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20110124095030.Q3489>