Weird Tcl/Tk configuration

From: Peter Eisentraut <peter_e(at)gmx(dot)net>
To: PostgreSQL Development <pgsql-hackers(at)postgresql(dot)org>
Subject: Weird Tcl/Tk configuration
Date: 2000-09-21 12:57:40
Message-ID: Pine.LNX.4.21.0009211441510.362-100000@peter
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

As things stand, if you use the --with-tcl configure option and a
sufficient Tcl installation could not be found, it will print a message
and continue without it. I have already on several occasions explained
why I consider that behaviour is undesirable, and it also seems quite
absurd, considering that the user explicitly asked for Tcl by specifying
the option in the first place.

That seems easy to fix, but what should we do about the Tk part?
Currently, --with-tcl implies Tk, except that it will be disabled if
tkConfig.sh or X Windows could not be found. I propose that we instead
make that a failure and advise the user to use the --without-x option to
disable Tk. (That option already exists, but it is not evaluated.) That
would imply that Tk = Tcl + X, and consequently X = Tk - Tcl, which would
fail if we ever add another X program that is unrelated to Tcl/Tk. If you
are concerned about that, maybe a --without-tk option would be better.
The general assumption here is that the majority of users that want to use
Tcl is also equipped with Tk and X, so that only a few users would have to
specifically disable Tk.

Comments?

--
Peter Eisentraut peter_e(at)gmx(dot)net http://yi.org/peter-e/

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Peter Eisentraut 2000-09-21 12:59:13 Re: Timezone discrepancies
Previous Message Tatsuo Ishii 2000-09-21 01:23:58 Re: -S is missing in postgresql.conf?