Undefined symbol "__mb_sb_limit" during 7.0-BETA1 installworld

[韓家標 Bill Hacker]

Ken Smith wrote:
> On Wed, 2007-10-24 at 23:31 -0400, 韓家標 Bill Hacker wrote:
>> Ken Smith wrote:
>> *snip*
>>
>>> If you updated an older machine using cvsup this issue won't impact you
>>> at all.
>> Not 100% certain of that - Last night's pull of amd64 HEAD off the jp mirror 
>> produced a dysfunctional 8-CURRENT /boot/loader executable that hung 
>> 'till-Hong-Kong-harbour-freezes' on final reboot & subsequent.
>>
>> Core-2 Quad platform, 4GB, physically reduced to 2GB, w/o joy (unrecognized 
>> Allantic NIC onboard Asus P5K, RealTek NIC on PCI in-use).
>>
>> Building a new world / kernel from a fresh pull now, with the expectation the 
>> issue was transitory.
>>
>> Will chase further if otherwise.
>>
>> Bill Hacker
> 
> Sorry, my comment about updating an older system applied to RELENG_7,
> not HEAD.  Code freeze on HEAD is over so all bets are off for
> that... :-)
> 

Suspect it is time to 'branch' 7-anything mailing list from freebsd-current 
mailing list anyway, as confusion may [soon | already?]  outweigh the advantage 
of a common discussion area between 7 & 8.

Understand the 'bets are off..' part, but thought you meant the problem came 
about in that a true HEAD tag would draw 8-CURRENT, not latest 7_<whatever>.

..At least with the supfile used here of:

*default tag=.
*default-host=cvsup.jp.freebd.org
*default release=cvs
*default delete use-rel-suffic

It so seems....

RELENG_7 tag OTOH, seems to have been working 'adaptively' for the 7-<whatever> 
thoughout the recent naming changes..

Side issue - but why not rename (or link):

(HEAD ==> <version_number>-WHIPS (Work Hopefully In Progress Soon)

- to clarify what one might expect - and/or satisfy the masochists among us...

(my buildworld having just barfed a few minutes ago in a way I should have 
anticipated....)

;-)


Bill