Bugzilla Bug 236683

Severity trivial since it's just prettying an error message...

After having an emerge fail, I modified the ebuild to see what was up, and
tried to remanifest it, only to get a traceback.

ebuild /p/sys-apps/hal/hal-0.5.11-r2.ebuild manifest
>>> Creating Manifest for /str/portage/sys-apps/hal
!!! write_atomic('/str/portage/sys-apps/hal/Manifest')

Tried digest, same error.

Checked the manpage for a debug to try, added --force, got a traceback which
let me see the problem:

ebuild /p/sys-apps/hal/hal-0.5.11-r2.ebuild digest --force
Traceback (most recent call last):
  File "/usr/bin/ebuild", line 217, in <module>
    discard_digests(ebuild, tmpsettings, portage.portdb)
  File "/usr/bin/ebuild", line 163, in discard_digests
    mf.write()
  File "/usr/lib64/portage/pym/portage/manifest.py", line 226, in write
    "".join("%s\n" % str(myentry) for myentry in myentries))
  File "/usr/lib64/portage/pym/portage/util.py", line 945, in write_atomic
    raise PermissionDenied(func_call)
portage.exception.PermissionDenied:
write_atomic('/str/portage/sys-apps/hal/Manifest')

Note that exception, permission denied.  DOH!  I was trying to manifest/digest
as a regular user!  Ran it using sudo without issue.

I would have expected instead of that cryptic and a bit scary "write_atomic" a
nice error telling me to try it as root, or to check permissions.

Duncan

In svn r11922 its fixed to say 'Permission Denied' so there's no confusion.

As always, thanks!

(FWIW, I really do appreciate how user friendly you tend to be handling portage
bugs.  It must get depressing sometimes, but having someone that actually takes
time to explain what was going on and what the fix actually fixed, as well as
how to apply the patches until a new version comes out, really DOES make a
difference.  It's certainly noticed and appreciated here!)

This is fixed in 2.2_rc15.

(In reply to comment #2)
> As always, thanks!

You're welcome. :)