Re: FreeBSD 10.0-ALPHA4 now available

Hi, Reference:
> From:		Benjamin Kaduk <kaduk_at_mit.edu> 
> Date:		Tue, 8 Oct 2013 20:07:59 -0400 (EDT) 

Benjamin Kaduk wrote:
> [-re]
> 
> On Mon, 7 Oct 2013, Julian H. Stacey wrote:
> 
> >> The fourth ALPHA build of the 10.0-RELEASE release cycle is now available
> >> on the FTP servers for the amd64, i386, ia64, powerpc, powerpc64 and
> >> sparc64 architectures.
> >>
> >> Note:  Due to build issues within the head/ branch, ALPHA3 ISO builds
> >> were skipped.
> >>
> >> The 10.0-ALPHA4 builds correlate to svn revision r255933 of the head/
> >> branch.
> >
> > (After zapping boot on an adjacent parition & repairing,)
> > I avoided installer, & booted another partition, mounted
> > FreeBSD-10.0-ALPHA4-amd64-memstick.img & tar'ed to empty partition
> >
> > 10.0-ALPHA4 runs, built & ran new generic kernel & ports X11 etc. :-)
> >
> > I had supposed an Alpha /usr/src would compile even if it broke at run time.
> > Not so, 10.0-ALPHA4 src/ is broken (with & without obj & depend).
> > 	 cd /usr/src ; make -i install
> > killed the system. To avoid killing, I hacked Makefiles, list below.
> > Reinstalled & survived.
> 
> I thought buildworld+installworld was the documented way to build 
> everything.


Yes, for an evolving current that's true 
(I'd perhaps forgotten, as I've been away from current & was tired) 
but in the case of a release (& Alpha4 is intended for a release
after fixes) I expect binaries on media to exactly match the identical
src they are presumably made from, so as a simple sanity & machine
hardware & file corruption etc check I always do the same simple
make with any release.

If re_at_ are not shipping a matched set of bins built from src theyre
shipped with, I'd be very suprised, & would want to know.


> Do you have other documentation I should look at?

Sorry, not sure what you want ?

> > I'll try with a newer current later, but Alpha4 src/ needs fixing.
> >
> > cd /usr/src ; cp /dev/null ~/tmp/`uname -r`.diffs
> > foreach i ( `find . -name Makefile\*.pre_jhs | sort`)
> > echo >> ~/tmp/`uname -r`.diffs
> > diff -c $i `dirname $i`/`basename $i .pre_jhs` >> ~/tmp/`uname -r`.diffs
> 
> BTW, unified diffs are preferred.

I find context diffs easier to read, & these were just to read, to
show where ALPHA4 is broken. I do not want those patches stoed to
apply, that would be wrong, they are just temporary to show where
ALPHA4 will commit suicide, untill re_at_ fixes underlying breakages.

Cheers,
Julian
-- 
Julian Stacey, BSD Unix Linux C Sys Eng Consultant, Munich http://berklix.com
 Reply below not above, like a play script.  Indent old text with "> ".
 Send plain text.  No quoted-printable, HTML, base64, multipart/alternative.