On 6/19/2007 9:22 AM, Christopher Browne wrote:
> "Andrew Hammond" <[EMAIL PROTECTED]> writes:
>> On 6/18/07, Jan Wieck <[EMAIL PROTECTED]> wrote:
>>> On 6/16/2007 5:32 PM, Stéphane Schildknecht wrote:
>>> > David Fetter a écrit :
>>> >>
>>> >>> # ----
>>> >>> -if test $# -ne 2 ; then
>>> >>> - echo "usage: $0 subscriber-dbname clustername" >&2
>>> >>> - exit 1
>>> >>> +if test $# -lt 2 ; then
>>> >>> + echo "usage: $0 subscriber-dbname clustername [-H host] [-P
>>> >>> port] [-U user]" >&2
>>> >>>
>>> >>
>>> >> Why -H and -P instead of -h and -p, respectively?
>>> >>
>>> >>
>>> >
>>> > For no good reason, except I already used -h for --help and -P was just
>>> > a silly way to have the same (upper) case for all options.
>>>
>>> -? is a good idea for --help.
>>
>> I agree. Following the convention in pg_dump, psql and friends is, I
>> think, quite important. That means that -h is host, and -p is port.
>
> The thing I keep wondering about all this is "But why bother?"
>
> If you set PGHOST, PGPORT, PGUSER, and such, then, via the "magic" of
> libpq, you have control over all of these parameters, and actually
> more, because PGSERVICE is also handled automatically by libpq.
So that turns
Slony1_dump.sh -h foo -u me mydb
into
PGHOST=foo PGUSER=me Slony1_dump.sh mydb
Hmmm, don't know if that makes it that much easier.
Jan
>
> That's why I haven't been quick to respond or to grab every proposed
> patch as it comes along; I'm not convinced that there is any
> particular value to adding in all that code.
>
> If someone is keen on arguing that command line options always need to
> be around because they find environment values "evil," well, that's an
> argument that can be made.
>
> We found at Afilias that some careful selections of PG environment
> variables turned out to usually be pretty useful, with little
> downside. So I'm inclined to disagree with the "anti-environment"
> positions. (And FYI, anyone willing to even *touch* Perl pretty much
> needs to be in that camp, as Perl's overall behaviour goes wildly
> above and beyond that, as "man perlvar" would attest ;-).)
>
> Anyways, my question still is, "Why bother?"
--
#======================================================================#
# It's easier to get forgiveness for being wrong than for being right. #
# Let's break this rule - forgive me. #
#================================================== [EMAIL PROTECTED] #
_______________________________________________
Slony1-general mailing list
[email protected]
http://gborg.postgresql.org/mailman/listinfo/slony1-general
