11 Dec
2009
11 Dec
'09
9:40 p.m.
Mariano,
Would you be in favor of the abstract class, if the abstract class itself had to be booted
into the image?
It's just a practical issue for me ... if we can count on the abstract class being
preloaded in all dialects, then an abstract class makes sense.
If we can count on Metacello being preloaded in some of the dialects, then the
ensureMetacello method need only be present for portability to platforms in which
Metacello is not preloaded.
I don't disagree with the oddity of copy and paste.
Earlier versions of Metacello configurations did rely on a superclass, but booting the
superclass into the image was problematic. Granted bootstrapping would only be hard in
Squeak, but it isn't supposed to be hard!
The ensureMetacello method is only there for ease of bootstrapping software that is
supposed to make it easy to boostrap software, but can't do the bootstrapping until it
is bootstrapped itself:)
I just hate the idea of making it harder to share configurations with Squeak...
Dale
----- "Mariano Martinez Peck" <marianopeck(a)gmail.com> wrote:
| On Fri, Dec 11, 2009 at 8:26 PM, Dale Henrichs
| <dale.henrichs(a)gemstone.com>wrote;wrote:
|
| > I wouldn't be entirely opposed to having an abstract configuration
| class,
| > but I hate to switch the configuration structure in midstream
| again... it is
| > very nice that a config can be loaded into any old smalltalk image
| without
| > any prerequisites....
| >
| > The #ensureMetacello method should simply load
| ConfigurationOfMetacello and
| > go from there ...
| >
| > To require an abstract configuration class, I would want that class
| to be
| > in the core for Pharo, GemStone and Squeak ... Metacello is already
| bundled
| > with the core in GemStone so that leaves Squeak ... I really don't
| want to
| > have to preface all of the install instructions with "Before loading
| the
| > configuration, make sure that you've done X, Y and Z"
| >
| > If bootstrapping is the only consideration, I don't think it is
| worth
| > making the change ... there isn't a lot protocol that I would want
| to put
| > into the Abstract configuration class ... project instance variable
| and
| > method ...
| >
| > If there are enough other arguments for an Abstract configuration
| class,
| > then it would be good to make the switch sooner rather than later.
| >
| > Being able to view the configurations using a hierarchy browser is
| an
| > advantage, but that's more of a tools issue.
| >
| > As I think about it, most of the advantages that come to mind are
| tools
| > related. Even bootstrapping is a tools issue of sorts...
| >
| > What do others think?
| >
|
| The first time I have to create my configuration, when I did the
| tutorial,
| and I have to copy paste the template I wondered myself why there
| wasn't an
| abstract configuration class :)
| In my opinion, it is easier for the configurations. You put the
| ensure
| metacello and all those methods that are in the template there and
| that's
| all.
|
| So... +1 to have an abstract class for me.
|
|
| >
| > Dale
| >
| > ----- "Tudor Girba" <tudor.girba(a)gmail.com> wrote:
| >
| > | Oh! I am slightly embarrassed :)
| > |
| > | I updated ConfigurationOfMooseJEE to use the latest
| > | ConfigurationOfMetacello and indeed it works.
| > |
| > | But I think this basically means that it starts to make sense to
| have
| > |
| > | an abstract configuration in Pharo-Core, because copy paste just
| does
| > |
| > | not scale anymore.
| > |
| > | Cheers,
| > | Doru
| > |
| > | p.s. Btw, I love the progress bar :)
| > |
| > |
| > | On 11 Dec 2009, at 19:24, Dale Henrichs wrote:
| > |
| > | > Doru,
| > | >
| > | > I looked at the #ensureMetacello method ConfigurationOfMooseJEE,
| but
| > |
| > | > didn't notice (the first, second and third time that I looked:)
| that
| > |
| > | > it is using MetacelloConfiguration so latest version isn't
| getting a
| > |
| > | > recent enough version of Metacello ...
| > | >
| > | > So update the ConfigurationOfMooseJEE class>>ensureMetacello to
| boot
| > |
| > | > with ConfigurationOfMetacello instead of
| MetacelloConfiguration...
| > | >
| > | > Dale
| > | > ----- "Dale Henrichs" <dale.henrichs(a)gemstone.com> wrote:
| > | >
| > | > | Doru,
| > | > |
| > | > | Things worked fine (well almost fine) with my first try at
| > | loading
| > | > | MooseJEE ... initialization failed in FSPlatformResolver,
| because
| > | > | there is no FSUnixResolver, however the
| Morphic-MorphTreeWidget
| > | is
| > | > | being loaded. I used this expression:
| > | > |
| > | > | Gofer new squeaksource: 'MooseJEE'; addPackage:
| > | > | 'ConfigurationOfMooseJEE'; load.
| > | > | (Smalltalk at: #ConfigurationOfMooseJEE) perform:
| #loadDefault
| > | > |
| > | > | ... This test was run with 1.0-beta.18 preloaded, so I will
| try
| > | > again
| > | > | with a virgin image...
| > | > |
| > | > | Dale
| > | > | ----- "Tudor Girba" <tudor.girba(a)gmail.com> wrote:
| > | > |
| > | > | | Hi,
| > | > | |
| > | > | | A strange problem, that appeared before but was solved at
| the
| > | > time,
| > | > |
| > | > | | seems to resurface again in a slightly different context.
| > | > | |
| > | > | | It is related to the Morphic-MorphTreeWidget from the Momo
| > | > | | squeaksource repository. This is needed in Glamour.
| > | > | |
| > | > | | Glamour is needed in Moose.
| > | > | |
| > | > | | Now, if you load ConfigurationOfMoose, it all works fine:
| > | > | |
| > | > | | Gofer new squeaksource: 'Moose'; addPackage:
| > | > 'ConfigurationOfMoose';
| > | > |
| > | > | |
| > | > | | load.
| > | > | | (Smalltalk at: #ConfigurationOfMoose) perform: #loadDefault
| > | > | |
| > | > | | However, Moose is needed in MooseJEE. And if you load
| > | > | | ConfigurationOfMooseJEE it does not work anymore:
| > | > | |
| > | > | | Gofer new squeaksource: 'MooseJEE'; addPackage:
| > | > | | 'ConfigurationOfMooseJEE'; load.
| > | > | | (Smalltalk at: #ConfigurationOfMooseJEE) perform:
| #loadDefault
| > | > | |
| > | > | | If you look in the packages to load,
| Morphic-MorphTreeWidget
| > | does
| > | > | not
| > | > | |
| > | > | | seem to appear among them in the latter case.
| > | > | |
| > | > | | I would need help for this one, if someone has time to look
| > | into
| > | > | that
| > | > | | :)
| > | > | |
| > | > | | Cheers,
| > | > | | Doru
| > | > | |
| > | > | | --
| > | > | | www.tudorgirba.com
| > | > | |
| > | > | | "The coherence of a trip is given by the clearness of the
| > | goal."
| > |
| > | --
| > | www.tudorgirba.com
| > |
| > | "Presenting is storytelling."
| >