geda-developers team mailing list archive

[Vladimir Zhbanov <vzhbanov@xxxxxxxxx>]

On Thu, Sep 17, 2015 at 01:10:20PM +0200, Roland Lutz wrote:
> On Thu, 17 Sep 2015, Vladimir Zhbanov wrote:
> >So you consider adding various rewrites of the programs we have using
> >various languages to be a progress?
> 
> Xorn isn't a rewrite of existing programs.  It consists of:
> 
>   libxornstorage - a new C library which helps managing objects in memory
guile has a garbage collector

>   xorn.storage - Python bindings for libxornstorage
we already have guile bindings in our programs

>   xorn.geda - a subset of libgeda, ported to Python and refactored
OK

>   xorn.geda.netlist - the gnetlist code, ported to Python and refactored
OK

> 
> There are good reasons to go that way, as I outlined on the geda-user list.
Probably I could not read all the list as quickly as the messages come
and probably I've missed something. Could you please shortly list the
advantages of xorn over the current geda infrastructure. (Sorry for the
question, however, some developers might be not aware of them.)

> 
> >Aren't you afraid this will turn into one-man show?
> 
> I hope it won't; that's part of the reason why I agreed to merge Xorn into
> mainline gEDA/gaf at an early point.
OK.

> 
> >>Branch from just before the xorn integration, then.  I think that
> >>until we have a solid development schedule, we're going to have to
> >>risk things happening "not when convenient".
> >
> >Yes, yes. And then branch from that branch, and then again. Why is the
> >master branch needed then?
> 
> Branching for a release is actually the way it's usually done, and there are
> good reasons for it.
Common words, sorry.

> 
> >Now, when geda-gaf builds also without problems on *BSD, it would be the
> >most appropriate moment.  But you move in an experimental stuff, making
> >things to be in flux again, and drive me into a corner.
> 
> This is one of the good reasons.
> 
> On Thu, 17 Sep 2015, Vladimir Zhbanov wrote:
> >the stuff they're working on will be thrown out
> 
> I'm checking in regular intervals for changes on gEDA/gaf master that touch
> the relevant parts and make sure thay are reflected in Xorn, so no, your
> work isn't thrown out.  This means I constantly have to keep the redundant
> code parts in sync, though, so I have an incentive to keep the code
> duplication low.
> 
> >and replaced with something 'more perfect' that nobody uses yet.
> 
> That's the natural state in which any code starts out.  I hope some more
> people will be using it now it's in mainline gEDA.
OK, I see your point.
Please tell me why it needs to be so eagerly merged into the master
branch? Does this help to promote it further? Or is it just the way to
move it on without taking other opinions into account?

Cheers,
  Vladimir