Re: eselect profile list fails

On Wed, Jun 3, 2026 at 9:45=E2=80=AFAM Javier Martinez <[email protected]=
> wrote:
>
> El 3/6/26 a las 15:40, John Blinka escribi=C3=B3:
> > Hello, everyone,
> >
> > Could use some help.
> >
> > eselect profile list gives this error on all my boxes:
> > !!! Error: Failed to get a list of valid profiles
> >
> > All boxes have identical setup, all are amd64.
> >
> > ls -l make.profile gives
> > make.profile ->
> > ../../var/db/repos/gentoo/profiles/default/linux/amd64/23.0/desktop,
> > whose contents are:
> > eapi  gnome/  parent  plasma/  systemd/
> >
> > I run openrc and xfce4 with a smattering of rarely used gnome/kde/
> > apps. For the most part I run a stable system, with occasional use of
> > ~amd64. I keep up-to-date with portage. My systems have been very
> > stable, although I am wrestling with the recent python upgrade, with
> > sound support on one box, and the recent spate of kernel upgrades. The
> > problems with eselect profile list have existed for months and predate
> > my current issues. This problem has just now arisen to the top of my
> > TODO list.
> >
> > The error message doesn't give me any guidance. The usual Googling
> > doesn't uncover much except recommendations to update portage, which I
> > do very regularly.
> >
> > Thanks for any ideas on where to start looking!
> >
> > John Blinka
> >
>
> With a strace -f -o /tmp/output eselect profile list, in the output file
> you can get more info about why it's failing

Thanks for the quick response, Javier!

I'm rather unskilled at looking at such output, but, scrolling up from
the bottom and looking for system calls returning suspicious values, I
find

20071 wait4(-1, 0x7ffdd4a1f4d0, WNOHANG, NULL) =3D -1 ECHILD (No child proc=
esses)

This condition seems to trigger subsequent
20071 write(1, "\33[31m\33[1m!!! Error: \33(B\33[mFailed"..., 65) =3D 65
and then
20060 rt_sigreturn({mask=3D[CHLD]})       =3D -1 EINTR (Interrupted system =
call)
which I'm guessing are just part of error recovery.

I'm not enlightened by this. Probably my deficiency! ;)

Is there another approach you can suggest?  For example, emerging a
clean copy of something that might correct whatever is wrong.

Thanks,

John