On Thu, 16 Dec 1999, Peter Eisentraut wrote:
> On Wed, 15 Dec 1999, Karel Zak - Zakkr wrote:
>
> > $ pg_dump --help
> > /usr/lib/postgresql/bin/pg_dump: invalid option -- -
> >
> > hmm ?
> >
> > Prepare anyone long options for pg_dump, pg_passwd, pg_version ?
> > If not, I make it, current state is disgraceful.
Hi,
> Only GNU (Linux) systems support long options. What you consider
> disgraceful is normal behavior on the majority of platforms.
Yes, I know this (it is not first program, which I hacking :-),
IMHO is good resolution (?) add getopt_long() to str/utils and use it
for non-GNU (non getopt_long) platforms, or not?
> I put in long options into psql and into the wrapper scripts (under some
> protests). But I agree that we should if at all only provide them, not
> advertise them, since it's going to be a support nightmare.
I saw/use your current psql (great!). Plan you add long option to
usage() output? If yes, will good make identical output format for
all pg_ routines (nice is example 'mc --help' styl).
> If you plan on doing this, please look there and use the same options
> whereever possible. I have been trying to establish some consistent
> options naming across client applications. Also check out how psql copes
> with systems where there are no long options available.
Yes, I agree and I use your options.
> Meanwhile I have been getting closer to resolving to scratch pg_dump and
> write a new one for 7.1, so whatever you do now might not live long. :(
Hmm.. :-(, but I not worry, it is only small work for me and you can use
this options (and code) for your 7.1 version.
Karel
