cvs commit: src UPDATING

Nate Lawson nate at root.org
Tue Feb 27 23:19:23 UTC 2007


Ruslan Ermilov wrote:
> On Mon, Feb 19, 2007 at 10:49:43PM +0000, Nate Lawson wrote:
>> njl         2007-02-19 22:49:43 UTC
>>
>>   FreeBSD src repository
>>
>>   Modified files:
>>     .                    UPDATING 
>>   Log:
>>   Note problems I had with bsnmpd while updating from an older -current.
>>   There may be better ways to fix/work around these issues but this worked
>>   for me.
>>   
>>   Revision  Changes    Path
>>   1.474     +13 -0     src/UPDATING
>>
> It could be a pilot error on your part as well.  The gensnmptree
> tool is already part of the bootstrap-tools set, and is correctly
> bootstrapped (built, installed into a temporary place, and used)
> when/if needed
> 
> : .if ${BOOTSTRAPPING} < 700015
> : _gensnmptree=   usr.sbin/bsnmpd/gensnmptree
> : .endif

I didn't know that.  I did use buildworld.  The system was an untouched
7-current from March 2006 (I think).  I recently cvsupped to HEAD and
did the buildworld/installworld.

> ... as determined by a value of __FreeBSD_version from the
> installed copy of /usr/include/osreldate.h.  It's possible to
> fool the build system into thinking that you have a different
> (e.g., more recent) version of FreeBSD installed by doing
> partial installs (when osreldate.h doesn't match the installed
> version) which is not uncommon on developers' machines.  :-)
> In that case, it's possible to work around this by telling
> the build system you have a "very old" environment, like this:
> 
> make buildworld OSRELDATE=0
> 
> I'd like to see this note removed, or more details to be
> provided that we could identify a real problem (e.g., if
> more incompatible changes were made to gensnmptree after
> 700015).

I hadn't done anything with the system since about June 2006.  I don't
rule out the possibility that while trying to fix the buildworld problem
that I copied in /usr/src/include to /usr/include, but I definitely
didn't do it before the first buildworld failed there.

Is there any harm in leaving the message in UPDATING?  It's just an FYI
in case someone else sees the same problem.  You can modify it to say
this is likely pilot error if you want, but I think it's helpful to have
it there since I saw a few similar reports on list archives with no
response.

-- 
Nate


More information about the cvs-src mailing list