> I don't see this as an acceptable resolution; programs or scripts
> written to depend on coreutils may not run unless we supply these
> utilities.

Isn't this true, though, of the omission of any component from
coreutils?  It's likely there are fewer programs that depend on
coreutils' "su" or or even "shred" but we face the same issue with
leaving those programs out as well.

I don't know whether the LSB "standard" covers any of these components
but in the absence of a GNU standards compliance requirement for
OpenSolaris, it seems omitting certain items from coreutils is
reasonable if not unavoidable (eg, for the security issues mentioned
earlier.)

In any case, I assume the unaltered GNU *sum utilities could still be
shipped in /usr/gnu/bin even if the "native" ones proposed in
PSARC/2005/530 were delivered in /usr/bin so a script requiring the
coreutils' version would need to augment their PATH accordingly.

(That said, I'm still uncomfortable with a ARC case that has reserved
the name-space but not actually integrated anything preventing other
projects from integrating an alternate implementation.  Perhaps it just
means that PSARC/2005/530 needs a follow-on case to include whatever
changes necessary to be able to replace the coreutils versions.)

dsc

Reply via email to