Re: JDBC build patches

Поиск
Список
Период
Сортировка
От Peter Mount
Тема Re: JDBC build patches
Дата
Msg-id 5.0.2.1.0.20010312201548.00a20510@mail.retep.org.uk
обсуждение исходный текст
Ответ на Re: JDBC build patches  (Peter Eisentraut <peter_e@gmx.net>)
Ответы Re: JDBC build patches  (Peter Eisentraut <peter_e@gmx.net>)
Список pgsql-jdbc
At 17:47 12/03/01 +0100, Peter Eisentraut wrote:
>Peter Mount writes:
>
> > What's the difference between uninstall and clean?
>
>'clean' undoes 'all', 'uninstall' undoes 'install'.

Ah ok.


> > >* honour the default port as specified to configure
> >
> > Hmm, not a good idea. The driver should be as globally useable (part of
> > write once run anywhere), so this would be pretty disasterous.
>
>ISTM that if the user actually goes through the trouble of building the
>driver himself (as opposed to downloading the jar file) as well as
>configuring with a non-standard port (as opposed to configuring the port
>at runtime) he probably wants the JDBC driver in on that deal.  It also
>seems unlikely that he would take this driver and send it to some random
>person to make a point about "run anywhere".

True. I just didn't want there to be a rogue driver with a different port.
Over time I've kept the driver pretty clean on that front (been one of the
goals). As long as they understand the driver is unique to them, it
shouldn't catch anyone out.


> > So far our standard has been to use 5432 if no port is supplied.
>
>It's still this way.  As I said, only people that use configure
>--with-pgport see any change, and those people know what they're getting
>into.

Perhaps a message saying this (or even in the version strings within JDBC).


> > >* allow building outside the source tree
> > >(This doesn't actually work, because Ant always puts the build dir
> > >relative to the source dir, but at least the whole thing will proceed
> > >smoothly when the rest of the tree builds this way.)
> >
> > You can by passing the base dir in the ant command line using -D
> > ant -Ddest=/tmp/dest -Djars=/tmp/jars
> > would build unti /tmp/dest and put the jars under /tmp/jars
>
>That's what I thought, but for some reason it always appends the path to
>the current directory, even if it's absolute.  Might be an old version of
>Ant.

Might be, I'll check.

There is a nice little gotcha with Cygwin, but then that's the problem of
Win32 IMHO :-)

Peter


В списке pgsql-jdbc по дате отправления:

Предыдущее
От: Peter Eisentraut
Дата:
Сообщение: Re: JDBC build patches
Следующее
От: Peter Mount
Дата:
Сообщение: Re: [INTERFACES] object id