[Openicc] Re: [Lcms-user] Beginner's questions

[Kai-Uwe Behrmann]

The offsets can easily been exchanged through Oyranos, as well as the
othe parameters influencing the transform. As an option the chache can
been removed after switching the computer off.

Ahhem .. better would be of course an interface, which provides pointers,
or what appropriate, after been asked about the specific transform
options including the belonging profiles. The application should
not care about memory stuff at all. Creation , invalidation and so one can
be better handled by an single library.

regards
Kai-Uwe Behrmann
                                + imaging development / panoramas
                                + color management
                                + email: ku.b at gmx.de
                                + CMS proposal <www.behrmann.name>


Am 13.12.04, 10:57 -0600 schrieb Bob Friesenhahn:

> On Mon, 13 Dec 2004, Kai-Uwe Behrmann wrote:
> >
> > Would an transform cache fit in an global memoy space? This using the
> > same memory across applications and save space. Is taking the profile id's
> > of both involved profiles plus creation options sufficent?
>
> Using a memory-mapped file obtains the benefit of shared "global"
> memory while obtaining the ability to re-use the cache between program
> invocations.  In order for this to work, the transform layout would
> need to observe certain rules.  For example, it would be best to store
> offsets from the base of the transform rather than storing pointers.
> That would allow the transform to be loaded at any address.  Also, it
> is best that the transform be able to live in a linear memory region
> rather than having some parts of it allocated separately from the
> heap.
>
> Bob
> ======================================
> Bob Friesenhahn
> bfriesen at simple.dallas.tx.us
> http://www.simplesystems.org/users/bfriesen
>