On Tue, Sep 19, 2006 at 04:00:14PM -0700, Fred Cox wrote:
> --- Kris Kennaway <[EMAIL PROTECTED]> wrote:
> 
> > On Tue, Sep 19, 2006 at 03:25:45PM -0700, Fred Cox
> > wrote:
> > > 
> > > 
> > > --- Kris Kennaway <[EMAIL PROTECTED]> wrote:
> > > 
> > > > "Will fail to package" is pretty far from
> > perfection
> > > > in my book :)
> > > > 
> > > 
> > > I don't believe I ever said that.  It builds fine
> > and
> > > even runs, it just has lots of bugs.
> > > 
> > > Can you let me know what I said that gave you that
> > > impression?
> > 
> > Based on what we've discussed, the proposed changes
> > to the port will
> > make it fail to build with default settings (since
> > mysql 4 will be
> > installed), for example on the package build
> > cluster.
> > 
> 
> I missed that requirement.
> 
> So basically, until it can run with PHP 5 and MySQL 5,
> it's not good enough to be a port.  Restricting it to
> run with PHP 4 will mean that it will be rejected?

No, I guess you've still misunderstood.  I don't know how many times I
can say this, but let me try to explain once more: your port should be
buildable with the default settings of all ports involved.

This means that you can't place special requirements like "you have to
first install mysql 3.x, then install the php4-mysql port, then
install this port", because that is too non-generic and will not be
true on systems that already have php4-mysql installed with the
default mysql client.

The solution, which I explained several messages ago, is to make an
alternative php4-mysql3 port, which always depends on mysql 3.x, and
use that instead of php4-mysql (it may need to conflict with
php4-mysql, I don't know).  This really isn't very hard and you
perhaps could have done it already by now :)

Kris

Attachment: pgp0eXfQcpW2J.pgp
Description: PGP signature

Reply via email to