[Freeciv-Dev] Re: (PR#8521) sha.c crashes

["Raimar Falke" <i-freeciv-lists@xxxxxxxxxxxxx>]

<URL: http://rt.freeciv.org/Ticket/Display.html?id=8521 >

On Mon, Apr 19, 2004 at 07:21:36PM -0700, Jason Short wrote:
> 
> <URL: http://rt.freeciv.org/Ticket/Display.html?id=8521 >
> 
> Raimar Falke wrote:
> 
> >>Which leads me to conclude that destroy_unit should be removed and 
> >>unit_clone should be clone_unit_virtual.
> > 
> >>A "virtual" unit is the part of the unit allocation that is common 
> >>between client and server.  We don't need two concepts for this.
> > 
> > Uhhh. Care to cite something? AFAIK virtual means that the unit
> > doesn't exists (a virtual unit in the client is unknown to the server,
> > a virtual unit in the server-ai is unknown to the server,...). The
> > original reason for it was that some function's wanted "struct unit *"
> > so so you created a virtual unit to pass into them. It was first
> > introduced by Gregory in the air goto patches:
> > 
> > +/**********************************************************
> > + * Create a virtual unit to use in build want estimation
> > + *********************************************************/
> > +struct unit *create_unit_virtual(struct player *pplayer, int x, int y,
> > +                                Unit_Type_id type, bool make_veteran)
> 
> Yes.  And the units used by the historian agent are "virtual" units in 
> this sense.  They aren't a part of the normal hierarchy (unit hash, 
> pplayer->units, ptile->units) but are tracked internally by sha.c.

I agree that neither virtual units nor "sha units" are part of the
general reference graph. However I think that the construction and use
differ between these two. Virtual units are constructed from a few
bits of data (see parameters above) and are used to pass them to
functions. The sha however wants a data container (a clone) which
holds all data. The sha also doesn't pass the clone around.

> > So we should drop the _virtual and just name it
> > "unit_create". "unit_destroy" can then be used for both "unit_create"
> > and "unit_clone". unittools.c:create_unit should be renamed to
> > unittools.c:server_create_unit or similar to denote that more work is
> > done here.
> 
> Possibly.  This is just a matter of semantics, however.
> 
> My point is that there are only two concepts: server/client unit and 
> virtual unit.  A server/client unit is a subset/subclass of a virtual 
> unit.  This may not be how the concept of a virtual unit started but it 
> has undeniably evolved into this.

Since we want solutions, what do you suggest?

        Raimar

-- 
 email: rf13@xxxxxxxxxxxxxxxxx
  "Some development methodoligies suck. In fack, I think most of them
   do. The Linux kernel development is an unorganized mess. The
   OpenBSD kernel development is a closed mess. XFree86 development
   is a mysterious mess. And that's before we even get to proprietary
   software, which is often ten times worse"
    -- John Goerzen in freeciv-dev