On 11/13/06, Edgar J. De Cleene <edgardec2001 at yahoo.com.ar> wrote:
> Pavel Krivanek puso en su mail :
>
> > Edgar, I think that it will be better if you will focus on Monticello
> > "memory leak" now.
> Ok.
>
> And what is that ?
> I guess why using Monticello results in a bigger image .

Yes, you may test it on some smallter package than MinimalMorphic
(that's everyone, I guess ;-)

> By the way, you see the proof of concept
> /Users/elpelotero/Public/MinimalMorphic/package-cache/PackageInfo-edc.12.mcz
>
> for having preamble and postcript (copied from ChangeSet) on Monticello ?.
> I don't resume on the load of this , but think you also need it.

I had no time for that but I will. The bigger problem of Monticello we
should solve are overrrides. We have got some primitive packages and
if we load MinimalMorphic, this package overrides some existing
methods. In current MC it means that the new package takes ownership
of the method. But If you want to publish the lower level package, the
package is incomplete.

This is the warning for you and others too. You may publish changes in
MinimalMorphic package but if you want to publish changes for example
in Monticello kernel, you have to do it from the image without
MinimalMorphic.

I solved the initialization problems of MC package by renaming the
initialize class methods. But it's very ugly solution and we should
create something more general.

> No more complaint about Traits, as I decide give my best to this , but I
> could think different and do what you say :=)

Good combination ;-)

-- Pavel