While I'm firing off questions, any hints as to why there are no
ebuilds to satisfy bash? (I'm starting through the depgraph code, but
it is hairy and ugly....)
On 12/15/05, m h <[EMAIL PROTECTED]> wrote:
> Ok, I'm running into the consistancy message that Haubi mentioned.
> I've prefix set to
> /data1/portage/dec15/prefix
> but the toolsbox prefix (is that the correct nomenclature?) (I'll call
> it PREFIX2) is:
> /data1/portage/dec15/prefix/toolsbox-4-patchespre.20051215/i686-pc-linux-gnu
>
> I'll need to trace through to see whats happening. My make.conf in
> PREFIX2/etc is not being read even though I modify it....
>
> I'm sure I'm forgetting something somewhere... I'll step through to
> see what I find.
>
> On 12/15/05, m h <[EMAIL PROTECTED]> wrote:
> > Got around (hacked around) python issue (I have no portage user
> > currently, only the group). So in portage_data.py, I override
> > portage_uid to my userid
> > portage_uid=os.getuid()
> > and I'm rolling again.
> > Now I get the following:
> >
> > Invalid package name in package.provided: virtual/libc
> > Invalid package name in package.provided: virtual/linux-sources
> > Invalid package name in package.provided: virtual/os-headers
> > !!! Problem with sandbox binary. Disabling...
> >
> > PORTDIR /usr/portage
> > --- 'profiles/arch.list' is empty or not available. Empty portage tree?
> >
> > These are the packages that I would merge, in order:
> >
> > Calculating system dependencies
> > emerge: there are no ebuilds to satisfy ">=app-shells/bash-3.0-r9".
> >
> >
> >
> > !!! Problem resolving dependencies for >=app-shells/bash-3.0-r9
> > !!! Depgraph creation failed.
> > =========END OUTPUT============
> > Note the debug statement for PORTDIR, it is not prefixed!
> > I have the following set in make.conf:
> > PORTDIR=${PREFIX}/usr/portage
> > Am I missing something?
> >
> > I'll hack it for now (that should get rid of the arch.list error).
> >
> > Hopefully people don't mind the spam. Again, I'm not a gentoo dev, so
> > I'm learning about the guts of portage, hopefully this is useful (at
> > least as psuedo-dev notes) for someone else....
> >
>
--
[email protected] mailing list
