cvs commit: ports/www/yabb Makefile pkg-plist ports/www/yabb/files pkg-message.in

[Andrew Pantyukhin]

On 1/28/06, Kris Kennaway <kris at obsecurity.org> wrote:
> On Sat, Jan 28, 2006 at 11:13:14PM +0300, Andrew Pantyukhin wrote:
> > On 1/28/06, Kris Kennaway <kris at obsecurity.org> wrote:
> > > On Sat, Jan 28, 2006 at 10:44:29PM +0300, Andrew Pantyukhin wrote:
> > > > On 1/28/06, Kris Kennaway <kris at obsecurity.org> wrote:
> > > > > On Sat, Jan 28, 2006 at 04:59:32PM +0000, Pav Lucistnik wrote:
> > > > > > pav         2006-01-28 16:59:32 UTC
> > > > > >
> > > > > >   FreeBSD ports repository
> > > > > >
> > > > > >   Modified files:
> > > > > >     www/yabb             Makefile pkg-plist
> > > > > >     www/yabb/files       pkg-message.in
> > > > > >   Log:
> > > > > >   - Now installs into /usr/local/www instead of /usr/local/www/{data,cgi-bin}
> > > > > >   - Data flatfiles are now cautiously left behind whether they have changed or not
> > > > >
> > > > > It's not OK for packages to leave files behind after deinstallation.
> > > > > Can you please fix this?
> > > > >
> > > > > Kris
> > > > >
> > > > >
> > > > >
> > > >
> > > > NO_PACKAGE was set in the same commit, if that helps
> > >
> > > It's still not OK, but now you just hid the problem from the package
> > > cluster :)
> > >
> > > Kris
> > >
> > >
> > >
> >
> > No, it's not about the package cluster. The plist is incomplete,
> > that's the simplest/stupidest way to deal with 3 dozen important
> > data files in different dirs.
> >
> > I wish there was some way to leverage the file checksums in
> > the package database and tell pkg_delete (and pkg_deinstall,
> > too) not to remove any modified files unless the user has very
> > specifically told otherwise. AFAIK there isn't. I guess I'll have
> > to do it manually (in pkg-<scripts>). I haven't done it yet, but
> > I felt it was not _that_ important. Warnings are in place,
> > package building is denied.
> >
> > If you can't look at it without remorse, shoot me, mark it
> > broken, delete it, whatever. I still can't see what's itching so
> > bad.
>
> It's about standards that ports need to live up to - otherwise pretty
> soon we wind up back in the situation where 2000 packages have packing
> lists that bear little resemblence to what their ports install, and
> related quality problems.  Anyway, thanks for working on it.

Thanks to you for keeping an eye wide open!

If anything is done to this port other than a total removal,
will one of you please add a BROKEN or NO_PACAKGE
to it (the latter being closer to my heart). Thanks!