X-MimeOLE: Produced By Microsoft Exchange V6.5
Received: by onstor-exch02.onstor.net 
	id <01C820D3.E7CC3BEA@onstor-exch02.onstor.net>; Tue, 6 Nov 2007 16:19:47 -0800
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
Content-class: urn:content-classes:message
Subject: RE: seep changes have broken BSD builds
Date: Tue, 6 Nov 2007 16:19:47 -0800
Message-ID: <BB375AF679D4A34E9CA8DFA650E2B04E066D1DD9@onstor-exch02.onstor.net>
In-Reply-To: <BB375AF679D4A34E9CA8DFA650E2B04E02EB314D@onstor-exch02.onstor.net>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: seep changes have broken BSD builds
Thread-Index: AcggyvWIqRJuHPmhRXafnywCV6SovgAAgfqhAABEO3AAAUwcnwAAB+2w
References: <BB375AF679D4A34E9CA8DFA650E2B04E02EB314D@onstor-exch02.onstor.net>
From: "Maxim Kozlovsky" <maxim.kozlovsky@onstor.com>
To: "Ken Renshaw" <ken.renshaw@onstor.com>,
	"Andy Sharp" <andy.sharp@onstor.com>
Cc: "Larry Scheer" <larry.scheer@onstor.com>,
	"dl-Cougar" <dl-Cougar@onstor.com>,
	"Paul Hammer" <paul.hammer@onstor.com>

It seems to be your job to get the nightly builds right, so it does not
really matter how you get it done, either by using the make features, or
"adding a check to tally up the total errors for the bsd builds".

Nightly builds, if they are working right of course, provide sufficient
coverage so normal developers do not have to build the BSD with every
change they make. It is not laxness, but right trade off between the
effort spent to catch the errors and how soon the error is caught.

>-----Original Message-----
>From: Ken Renshaw
>Sent: Tuesday, November 06, 2007 4:15 PM
>To: Maxim Kozlovsky; Andy Sharp
>Cc: Larry Scheer; dl-Cougar; Paul Hammer
>Subject: Re: seep changes have broken BSD builds
>
>I'm not inclined to rewrite the bsd makefiles Max, nor do I think
nightly
>builds are to be used to build laxness into the development process.
>
>These constant build breaks are old, whatever the claimed reasons.
>
>
>
>
>-----Original Message-----
>From: Maxim Kozlovsky
>To: Ken Renshaw; Andy Sharp
>CC: Larry Scheer; dl-Cougar; Paul Hammer
>Sent: Tue Nov 06 15:42:00 2007
>Subject: RE: seep changes have broken BSD builds
>
>It would be very inefficient for everybody to start building the BSD as
the
>probability of breaking the BSD build by regular code modification is
too
>low.
>
>The make will stop on error just fine if you don't run it parallel or
>specifically requesting it to ignore errors. I am sure you can find the
>right make invocation option if you put some effort into it.
>
>
>
>>-----Original Message-----
>>From: Ken Renshaw
>>Sent: Tuesday, November 06, 2007 3:30 PM
>>To: Andy Sharp
>>Cc: Larry Scheer; dl-Cougar; Paul Hammer
>>Subject: Re: seep changes have broken BSD builds
>>
>>Hi Andy.
>>
>>No real mystery, just a chance to remind us a couple reasons that the
bsd
>>builds are...imperfect.
>>
>>Basically the two reaons are that many of the bsd makefiles do not
stop on
>>error, so having a script that checks the return code and goes by that
>will
>>miss things. Secondly, even with a good build there are a couple dozen
>>'allowable' "Error " strings, mostly NLS related.
>>
>>Given those two things its pretty easy to miss something unless you
read
>>through all the logs manually, which I just happened to do in the case
of
>>the first time install branch build. From there I rebuilt it out of
the
>dev
>>branch, gleaned the logs, and lo and behold found this error.
>>
>>I will add a check to tally up the total errors for the bsd builds,
>>hopefully keeping track of that.....and having cougar members build
>openbsd
>>as systemw is required to do for linux, and this shouldn't happen.
>>
>>Thanks,
>>
>>-Ken
>>
>>
>>
>>-----Original Message-----
>>From: Andy Sharp
>>To: Ken Renshaw
>>CC: Larry Scheer; dl-Cougar; Paul Hammer
>>Sent: Tue Nov 06 15:15:45 2007
>>Subject: Re: seep changes have broken BSD builds
>>
>>I'm afraid I don't understand.  The changes in question were made
weeks
>>ago, but I've been getting emails every day saying the nightly build
>>passed.
>>
>>On Tue, 6 Nov 2007 15:12:50 -0800 "Ken Renshaw"
>><ken.renshaw@onstor.com> wrote:
>>
>>> Thanks for the follow-up Larry.
>>>
>>>
>>>
>>> We have multiple builds waiting behind this fix, so I'd very much
>>> appreciate a resolution ETA for this one
>>>
>>>
>>>
>>> Thanks guys,
>>>
>>>
>>>
>>> -Ken
>>>
>>>
>>>
>>> ________________________________
>>>
>>> From: Larry Scheer
>>> Sent: Tuesday, November 06, 2007 1:29 PM
>>> To: Ken Renshaw; dl-Cougar
>>> Subject: RE: seep changes have broken BSD builds
>>>
>>>
>>>
>>> I did a quick check  of the source in dev. The definition for
>>> SSC_I2C_ADDR appears inside of  an #ifdef PROM statement in the
>>> prom/{bc,ch,cg}/code/sm-seep/seep.h files.
>>>
>>>
>>>
>>> ________________________________
>>>
>>> From: Ken Renshaw
>>> Sent: Tuesday, November 06, 2007 12:48 PM
>>> To: dl-Cougar
>>> Cc: Ken Renshaw
>>> Subject: seep changes have broken BSD builds
>>>
>>>
>>>
>>> Hi guys, I have a problem whereby the recent seep rewrites have
broken
>>> the BSD builds, specifically this file's compilation:
>>>
>>>
>>>
>>> //depot/dev/openbsd/src/sys/arch/pmonmips/machdep.c
>>>
>>>
>>>
>>> and possibly others as well.
>>>
>>>
>>>
>>> The problem is that one or both of these changes:
>>>
>>>
>>>
>>> Change 25915 on 2007/10/18 by jeffm@jeffm-dev '       These are
>>> changes to mak'
>>>
>>> Change 25799 on 2007/10/10 by timg@timg-work 'seep file unification.
>>> seep rel'
>>>
>>>
>>>
>>> break the compilation of both bobcat and cheetah kernels, with the
>>> following error in bsd-build-all:
>>>
>>>
>>>
>>> /usr/openbsd.new/usr/cross/pmonmips/usr/bin/cc -g -O2 -Werror -Wall
>>> -Wmissing-prototypes -Wstrict-prototypes  -Wno-uninitialized
>>> -Wno-format -Wno-main -mno-abicalls -mips2 -mcpu=3Dr4000    -I.
>>> -I../../../../arch -I../../../..
>>> -I../../../../../../../nfx-tree/Includes/ch
>>> -I../../../../../../../nfx-tree/Includes
>>> -I../../../../../../../nfx-tree/code/sm-eee
>>> -I../../../../../../../nfx-tree/code -DTIMEZONE=3D"0" -DDST=3D"0" =
-DNTP
>>> -DSWAPPAGER -DDEVPAGER -DKTRACE -DKMEMSTATS -DCRYPTO -DSYSVMSG
>>> -DSYSVSEM -DSYSVSHM -DUVM_SWAP_ENCRYPT -DCOMPAT_11 -DCOMPAT_43 -DFFS
>>> -DQUOTA -DMFS -DTCP_SACK -DNFSCLIENT -DNFSSERVER -DFDESC -DFIFO
>>> -DKERNFS -DNULLFS -DUNION -DINET -DIPFILTER -DIPFILTER_LOG
>>> -DBOOT_CONFIG -DNATIVE_ELF -DPCIBR_GAL -DPCIBR_V3 -DGTUARTCONSOLE
>>> -DCONSPEED=3D"B57600" -DGALILEO_REAL_SSC -DCHEETAH -DENABLE_WATCHDOG
>>> -Dtssccomm -DNFSBOOT_NETMASK=3D"0xffff0000"
>>> -DNFSBOOT_BROADCAST=3D"0xa01ffff" -DNETEEE -DNETEEE_ALIAS
>>> -DNETEEE_FRAGMENT -D_KERNEL -Dmips_cachealias -Dpmonmips
>>> -D__pmonmips__  -c ../.!
>>>
>>>  ./../../arch/pmonmips/pmonmips/machdep.c
>>>
>>> ../../../../arch/pmonmips/pmonmips/machdep.c: In function `boot':
>>>
>>> ../../../../arch/pmonmips/pmonmips/machdep.c:1017: `SSC_I2C_ADDR'
>>> undeclared (first use in this function)
>>>
>>> ../../../../arch/pmonmips/pmonmips/machdep.c:1017: (Each undeclared
>>> identifier is reported only once
>>>
>>> ../../../../arch/pmonmips/pmonmips/machdep.c:1017: for each function
>>> it appears in.)
>>>
>>> *** Error code 1
>>>
>>>
>>>
>>> Stop in /usr/openbsd.new/src/sys/arch/pmonmips/compile/SSC_I (line
>>> 1028 of Makefile).
>>>
>>> make: *** [bsd-kernel] Error 1
>>>
>>>
>>>
>>> In the machdep.c file there is this block of defines at the
beginning
>>> that I believe no longer works:
>>>
>>>
>>>
>>> #include <nfx-types.h>
>>>
>>> #include <sm-gt64240/i2c-api.h>
>>>
>>> #include <sm-hw/hw-api.h>
>>>
>>> #include <sm-seep/seep-api.h>
>>>
>>> #include <sm-seep/seep.h>
>>>
>>> #include <sm-seep/env-api.h>
>>>
>>> #include <nfx-types.h>
>>>
>>> #include "sm-prom-upgrade/from-api.h"
>>>
>>> #include <sys/amd29f040.h>
>>>
>>> #include "sm-prom-upgrade/prom-part-api.h"
>>>
>>> #include "ssc-crashsave/ssc-crashdump-api.h"
>>>
>>>
>>>
>>> Could someone please fix this ASAP? Thanks much.
>>>
>>>
>>>
>>> -Ken
>>>
