On Tue, Jan 13, 2009 at 10:20:19AM -0600, Paul Hartman wrote:
> >> I normally do "emerge -uDvN @world" (or in other words "emerge
> >> --update --deep --verbose --newuse @world"). Right now, it tells me
> >> this:

<snip>

> > read the man page.
> >
> > Especially the bit about bdeps - these are usually not included
> > in 'emerge -uND world' but will be included when you use -e

<snip>

> When I do "emerge -Dtp openoffice" it shows this:
> 
<snip>

> I'm still not entirely clear on why those appear with --deep
> openoffice but not --deep world. If they are build-time deps, wouldn't
> they be included when I emerge openofifce without --deep? If they are
> related to dev-perl/Archive-Zip-1.26 (as it appears), which is
> installed, why don't they get touched by @world? Maybe I just can't
> wrap my brain around how it works, please have patience with me.
> thanks :)

The problem with this last one is not --deep. It is --update. 

When you issued emerge --update --deep world, it found that according
to the installed ebuild portage kept, that 

  a) OpenOffice does not need rebuilding-- since there is no update
according to version number.
  b) none of the dependencies installed by openoffice needs
rebuilding, since there is no update.

But when you issued emerge --deep openoffice, you asked for portage to
consider rebuilding openoffice, which now looks at the ebuild in the
tree, which, as you noted, is different from the ebuild that you
emerged months ago. As such, any new dependencies that differs between
the two ebuilds and any build-time dependencies will need to be
satisfied. 

And my suggestion would be to just ignore this little thing with
openoffice. If open office runs, and that the ebuild change is minor
enough to not get a new release number, then you can probably wait
until the next release of openoffice to worry about updates. 

W
-- 
Willie W. Wong                                      ww...@math.princeton.edu
408 Fine Hall,  Department of Mathematics,  Princeton University,  Princeton
A mathematician's reputation rests on the number of bad proofs he has given.

Reply via email to