Re: SI packages
At 2002/5/30 14:01-0700 Sean 'Shaleh' Perry writes:
>
> The issue is there is no difference (apparently) between something
> that is allowed but not required and items that are required in the
> test suite. A failure is a failure. We can not state "sorry we do
> not provide functionality X".
Yes, whilst some of the commands and directories are only required if
"the corresponding subsystem is installed" (which the fhs test suites
handle by the upfront questions) there are many which are required by
the FHS. I believe the intent is to radically reduce the number of
these in the next version of the FHS, leaving a few examples, but
mainly relying on rules to classify where a given binary should live.
This will however make it harder to test (where a binary should live
can depend on the exact functionality offered by that binary) but if
we have been precise enough in the specification of the commands in
the gLSB we can have a LSB specific FHS test suite. And existence of
directories will remain important.
Regards,
Chris
--
cyeoh@au.ibm.com
IBM OzLabs Linux Development Group
Canberra, Australia
--
To UNSUBSCRIBE, email to lsb-impl-request@lists.linuxbase.org
with subject of "unsubscribe". Trouble? Email listmaster@lists.linuxbase.org
Reply to: