Re: [IPCop-devel] 1.4 or 1.5 (was: GUI)

[Gilles E. <g....@fr...>]

Selon ja...@gu...:

> > I have suggested some days ago (probably lost mail) :
> > nobody wants to waits 1.5 too much long time. Ok.
> > So let's make ipcop 1.5 =3D 1.4 on k2.6 and nothing more.
> > (the exact opposite as what I want !)
> >
I didn't say that we need v1.5 release in a short time.
I say that we have to take decisions on what could be done in v1.5 and wh=
at
could be addressed later.
This will have great influence on when v1.5 could be available.
Have a target in one year time look reasonable to me.
This does not mind at all there is a deadline to release v1.5.
As usual v1.5 should be release when it's ready. But we have to take now =
the
right decisions.

> > The reason: drivers present in k2.6 and not in k2.4
> > We can allow some variations: openswan2, new installer,
> > hardware detection.
> >
> > But we have to LEAVE THE GUI in 1.4 state=3D>erase and copy
> > again all 1.4 files.
> > And we no more do major changes on this IPCop 1.4/1.5.
> >
No there is already difference between v1.4 and v1.5 GUI.
I don't see a gain to loose those differences.
This is why we all ask you (Frank) to commit changes on v1.5 in the same =
time
(or before) as in v1.4. vpnmain.cgi was not updated (maybe others too).

> > In paralell, we start a NEW independant GUI (python / perl)
> > to handle the real changes (multiple interface, brain=3Dshorewall,
> > ...)
> >
A new GUI framework adoption should be reserved for a later version.
We already have enought work for 1.5.
This does not mean that preliminary work to define/study what could be th=
at new
structure could not start now. But this need to be done by different peop=
le
that the one actually involved or it will slow down v1.5 developpement.

I don't think that supporting multiples interfaces fully need to restart =
from
the ground to rebuild an entire new GUI structure.
> >
> > 1) we stabilize 1.5 build process
> > 2) reimplant ALL 1.4 GUI (thus removing all changes made by
> > block-out which is perfect as an addon).
> > 3) while designing future GUI, blockout stuff will naturally
> >  integrate, with many other features.
>
There is no urgency that will need to remove block-out to release v1.5 th=
at
fast.

> I think this is wrong direction.  1.5 was to be next big leap in
> IPCop, with modulation and more OO design in the interfaces.
>
> 1.4 should be in the bag.  No more large changes that take a year to
> roll out. Installl the new Kernal.  Add new card types.   Fix bugs.
> But no major changes.
>
> We can live on 1.4 for up to year, as the 1.5 works though the build
> process, but keep 1.4 only in the fix mode.
>
> Jack Beglinger
>

I agree on Jack remarks and previous Alan do.
Our actual key points are :
- how now set interfaces from new installer,
- keep green / red / blue /orange asset,
- allow multiples interfaces.

Gilles