[lsb-discuss] statfs specification addition: review

Ian Murdock imurdock at imurdock.com
Thu Dec 21 09:54:59 PST 2006

On 12/21/06, Robert Schweikert <robert.schweikert at abaqus.com> wrote:
> I think we can push the "deprecation" even a step farther by stipulating
> in the spec that as soon as (f)statvfs(64) has the file system type
> information (f)statfs(64) will be removed from the LSB and that this
> constitutes a special exception to the deprecation policy. Considering
> that (f)statvfs(64) will probably not have the required information
> until after 4.0  and that the interface will not really disappear from
> glibc I think this should be a reasonable compromise between those
> forces that need the call and those forces seeing this struct as
> obsolete.

I'd rather not do special exceptions to the deprecation policy,
particularly since glibc will no doubt continue to provide the interfaces
in question. Sends the wrong message that we're willing to cut corners.
We'll have enough cases where our hand is forced because the upstream
removed something more quickly than we'd like, and I'd rather not consider
the special exception option until we absolutely have to (and, even then,
I'm hoping we can either work with the upstream/distros to provide
some sort of compatibility layer, i.e., special exceptions should be
a last resort). In this case, marking the interface as deprecated,
with loud howlings by the tools when they're used, should be enough.

Ian Murdock

"Don't look back--something might be gaining on you." --Satchel Paige

More information about the lsb-discuss mailing list