David Laight wrote:
> 
> > And sure enough, last night's failed again due to flist inconsistencies.
> 
> flist discrepencies are a slight problem, but are not, IMHO, major failures
> are they are trivial to fix.
> 
Except they cause the process to fail....


> Ok, since they are extra files it is (probably) save to just delete them
> from DESTDIR.  Indeed the installboot pages have been removed from the
> build and distribution, so only exist in your DESTDIR from a previous
> build.
> 
Funny, I'd already checked the web site, and searched the mailing lists 
for any change to installboot at all in the past week, and NOTHING shows up!  I wonder how this could have slipped under the radar?


> So again the problem you are seeing is a side effect of an update build
> and you must delete these files from DESTDIR by hand.
> 
> So you 'last night' build actually worked.
> 
Except for that pesky problem with errors stopping the process.

Back on 26 Jun 2003 23:33:17 +0700, KRE mentioned that he just turns
checkflist off, as it complains about the otherwise normal stuff that 
he puts in the destdir to test the system before install.  (I'd never 
thought of that -- KRE is a very clever fellow.)

It might be an improvement to move the checkflist to 'build.sh release'
so that 'build.sh distribution' merely ensures everything is prepared 
for 'build.sh install', without failing due to EXTRA files.  That is, 
most of the time we're concerned with MISSING files, not extra files.

Then again, maybe I'm just fuzzy on the historic differences between 
build, distribution, and release targets.  The documentation isn't 
particularly clear.
-- 
William Allen Simpson
    Key fingerprint =  17 40 5E 67 15 6F 31 26  DD 0D B9 9B 6A 15 2C 32