Don Armstrong <[EMAIL PROTECTED]> writes:
> On Thu, 23 Feb 2006, Frank Küster wrote:
>> Don Armstrong <[EMAIL PROTECTED]> wrote:
>> > On Wed, 22 Feb 2006, Frank Küster wrote:
>> >> Adeodato Simó <[EMAIL PROTECTED]> wrote:
>> >> > Correct, so one would put in foo.postrm:
>> >> >
>> >> > rmdir --ignore-fail-on-non-empty /usr/local/lib/foo
>> >>
>> >> That's not sufficient, because /usr/local may be mounted ro, and
>> >> therefore the command may fail even if the directory is empty.
>> >>
>> >> rmdir --ignore-fail-on-non-empty /usr/local/lib/foo || true
>> >
>> > So you're suggesting that it's better to fail silently instead of
>> > failing loudly?
>>
>> Yes, please read Policy 9.1.2.
>
> Hrm, right... it just seems totally wrong to me for the package to
> create the directories and then not fail if removing them fails, since
> presumably the removal could fail for reasons unrelated to /usr/local
> being r/o. [Perhaps the ideal solution to resolve both of these issues
> is to have the script complain bitterly if it can't remove the
> directories and they exist, but not fail.]
>
>
> Don Armstrong
I think the same applies to creating the dir. It should try but not
fail if creating it fails.
As for complaining bitterly: That is either anyoing because it is
interactive or gets lost in all the output because it isn't. Damned if
you do, damned if you don't.
MfG
Goswin
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
