[PATCHES] remove "legacy" option

[Simo Sorce]

The "legacy" option was introduced to speed up development.
It is actually confusing and not necessary.
> From now on we store all using the member/memberof schema internally.

(Tested only with the native ldap driver).

0001 mainly converts the core sysdb functions
0002 fixes many errors in fill_grent and removes legacy support form
there
0003 adds an upgrade function to convert existing databases on the first
run.

Simo.

[Sumit Bose]

On Thu, Aug 27, 2009 at 12:58:41AM -0400, Simo Sorce wrote:
> The "legacy" option was introduced to speed up development.
> It is actually confusing and not necessary.
> > From now on we store all using the member/memberof schema internally.
>
> (Tested only with the native ldap driver).
>
> 0001 mainly converts the core sysdb functions
> 0002 fixes many errors in fill_grent and removes legacy support form
> there
> 0003 adds an upgrade function to convert existing databases on the first
> run.
>
> Simo.

ACK

bye,
Sumit

[Simo Sorce]

On Thu, 2009-08-27 at 14:33 +0200, Sumit Bose wrote:
> On Thu, Aug 27, 2009 at 12:58:41AM -0400, Simo Sorce wrote:
> > The "legacy" option was introduced to speed up development.
> > It is actually confusing and not necessary.
> > > From now on we store all using the member/memberof schema
>
> internally.
> > (Tested only with the native ldap driver).
> >
> > 0001 mainly converts the core sysdb functions
> > 0002 fixes many errors in fill_grent and removes legacy support form
> > there
> > 0003 adds an upgrade function to convert existing databases on the
>
> first
> > run.
> >
> > Simo.
>
> ACK

Pushed,
Simo.

[Jenny Galipeau]

Simo Sorce wrote:
> On Thu, 2009-08-27 at 14:33 +0200, Sumit Bose wrote:
> > On Thu, Aug 27, 2009 at 12:58:41AM -0400, Simo Sorce wrote:
> > > The "legacy" option was introduced to speed up development.
> > > It is actually confusing and not necessary.
> > > > From now on we store all using the member/memberof schema
> >
> > internally.
> > > (Tested only with the native ldap driver).
> > >
> > > 0001 mainly converts the core sysdb functions
> > > 0002 fixes many errors in fill_grent and removes legacy support form
> > > there
> > > 0003 adds an upgrade function to convert existing databases on the
> >
> > first
> > > run.
> > >
> > > Simo.
> >
> > ACK
>
> Pushed,
> Simo.

Does this mean that I can remove legacy = TRUE from any test 
configurations? What happens if I set legacy = FALSE? Please advice.
~Jenny

[Simo Sorce]

On Thu, 2009-08-27 at 09:55 -0400, Jenny Galipeau wrote:
> Simo Sorce wrote:
> > On Thu, 2009-08-27 at 14:33 +0200, Sumit Bose wrote:
> > > On Thu, Aug 27, 2009 at 12:58:41AM -0400, Simo Sorce wrote:
> > > > The "legacy" option was introduced to speed up development.
> > > > It is actually confusing and not necessary.
> > > > > From now on we store all using the member/memberof schema
> > >
> > > internally.
> > > > (Tested only with the native ldap driver).
> > > >
> > > > 0001 mainly converts the core sysdb functions
> > > > 0002 fixes many errors in fill_grent and removes legacy support form
> > > > there
> > > > 0003 adds an upgrade function to convert existing databases on the
> > >
> > > first
> > > > run.
> > > >
> > > > Simo.
> > >
> > > ACK
> >
> > Pushed,
> > Simo.
>
> Does this mean that I can remove legacy = TRUE from any test 
> configurations? What happens if I set legacy = FALSE? Please advice.
> ~Jenny

legacy should simply be ignored.
Simo.