On Thursday, March 9, 2017 2:28:47 PM EST Michael Orlitzky wrote:
> On 03/09/2017 02:00 PM, William L. Thomson Jr. wrote:
> > Under what circumstances?
> > 
> > ...
> > 
> > Seems like it is not possible to generate the above permission issue.
> 
> I can make them up all day...

I cannot find the exact comment, but I recall being told before || die used 
along with rm -f was incorrect or something along those lines.

Case in point dev-db/firebird use to have a line like

rm -rf "${S}"/extern/{btyacc,editline,icu} || die

But if you look at current ebuild it is now

rm -r "${S}"/extern/{btyacc,editline,icu} || die

The force option/argument was dropped. Why? Seems it could have remained. 
Which is why I commented, as I am pretty sure that has been said to me before.

Essentially do not use -f with || die or something to that effect. Maybe for 
different reasoning. Still one of those things I never liked. Do it this way, 
for this reason, that is not documented. Others may digress and leads to 
confusion.

>   *  VENDOR_PATH=VENDORPN="" and we try to "rm -rf /"

That would likely be an incorrect ebuild and should not have been committed to 
tree.

>   *  A hard drive error occurs.

Likely have much greater issues than rm failing, not sure || die or the rest 
would work in that case.

>   *  Bad memory crashes "rm".

Same as above, if rm is crashing due to memory issues, you likely have serious 
issues. I doubt portage would continue on, or the system, etc.

>   *  Somebody is running a recursive chmod on /var/tmp during emerge.

That sounds like user error if doing updates etc on a system others are 
administrating at the same time.  I would put that in the stupid category.
Plus doing such would likely cause problems for other things that use that 
directory if not portage entirely.

>   *  The tarball contains something you don't expect and can't delete.

Why would that package be added to tree, or a developer be working with 
something blindly?

-- 
William L. Thomson Jr.