Loading problem in Seaside

Hi all,

I have a problem using the new Seaside3 configuration in the ConfigurationOfJQueryWidgetbox because of the following: when I try to load the stable version, it actually tries to load version: '3.1.0-gemstone'. It says this is the current version for Seaside3, because this version is fully loaded (it is empty, so mathematically speaking, that is correct).

The result is an error (This version does not know the group "Core"). But this is not what I want …. how should I solve this? Should I remove the gemstone specific things and distribute these over the functional groups? Or is this a (known) bug in Metacello?

Diego

--
You received this message because you are subscribed to the Google Groups "Metacello" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/groups/opt_out.

Diege,

From reading the Seaside-dev list it seems to me that you solved this problem by removing the "empty" version?

If not, could you give me a pointer to where the Seaside3 and ConfigurationOfJQueryWidgetbox configurations can be found ... a load script that fails would be the most useful....

But, I think your analysis is correct ... empty versions have no meaning and are best removed completely...

Dale

----- Original Message -----
| From: "Diego Lont" <[hidden email]>
| To: [hidden email]
| Sent: Monday, July 22, 2013 2:15:30 AM
| Subject: [Metacello] Loading problem in Seaside
|
| Hi all,
|
| I have a problem using the new Seaside3 configuration in the
| ConfigurationOfJQueryWidgetbox because of the following: when I try
| to load the stable version, it actually tries to load version:
| '3.1.0-gemstone'. It says this is the current version for Seaside3,
| because this version is fully loaded (it is empty, so mathematically
| speaking, that is correct).
|
| The result is an error (This version does not know the group "Core").
| But this is not what I want …. how should I solve this? Should I
| remove the gemstone specific things and distribute these over the
| functional groups? Or is this a (known) bug in Metacello?
|
| Diego
|
| --
| You received this message because you are subscribed to the Google
| Groups "Metacello" group.
| To unsubscribe from this group and stop receiving emails from it,
| send an email to [hidden email].
| For more options, visit https://groups.google.com/groups/opt_out.
|
|
|

--
You received this message because you are subscribed to the Google Groups "Metacello" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/groups/opt_out.

Hi Dale,

To get the ConfigurationOfSeaside3 config:

Gofer new
        url: 'http://www.smalltalkhub.com/mc/Seaside/MetacelloConfigurations/main';
        package: 'ConfigurationOfSeaside3';
        load
       
You can probably already see that something is wrong from the result of:

ConfigurationOfSeaside3 project latestVersion

which yields: 3.1.0-gemstone

The problem lies with the way the configuration for 3.1 is being defined: tt seems that this configuration's baseline was split up into methods, each containing a logical part. By tagging them each as a version, they can be recomposed using Metacello's inclusion mechanism. But I guess this is a good place to 'hack' inclusion by invoking the separate methods from the baseline configuration method. (I know you prefer not too, but what would be the good approach here?)

Johan


On 22 Jul 2013, at 16:30, "Dale K. Henrichs" <[hidden email]> wrote:

--
You received this message because you are subscribed to the Google Groups "Metacello" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/groups/opt_out.

When I run

  ConfigurationOfSeaside3 project latestVersion

in GemStone/S, I get:

  3.0.8 [ConfigurationOfSeaside3]

which is different from what you are getting ... what else do I need to do?

Dale

----- Original Message -----
| From: "Johan Brichau" <[hidden email]>
| To: [hidden email]
| Sent: Monday, July 22, 2013 8:16:58 AM
| Subject: Re: [Metacello] Loading problem in Seaside
|
| Hi Dale,
|
| To get the ConfigurationOfSeaside3 config:
|
| Gofer new
| url:
| 'http://www.smalltalkhub.com/mc/Seaside/MetacelloConfigurations/main';
| package: 'ConfigurationOfSeaside3';
| load
|
| You can probably already see that something is wrong from the result
| of:
|
| ConfigurationOfSeaside3 project latestVersion
|
| which yields: 3.1.0-gemstone
|
| The problem lies with the way the configuration for 3.1 is being
| defined: tt seems that this configuration's baseline was split up
| into methods, each containing a logical part. By tagging them each
| as a version, they can be recomposed using Metacello's inclusion
| mechanism. But I guess this is a good place to 'hack' inclusion by
| invoking the separate methods from the baseline configuration
| method. (I know you prefer not too, but what would be the good
| approach here?)
|
| Johan
|
|
| On 22 Jul 2013, at 16:30, "Dale K. Henrichs"
| <[hidden email]> wrote:
|
| > Diege,
| >
| > From reading the Seaside-dev list it seems to me that you solved
| > this problem by removing the "empty" version?
| >
| > If not, could you give me a pointer to where the Seaside3 and
| > ConfigurationOfJQueryWidgetbox configurations can be found ... a
| > load script that fails would be the most useful....
| >
| > But, I think your analysis is correct ... empty versions have no
| > meaning and are best removed completely...
| >
| > Dale
| >
| > ----- Original Message -----
| > | From: "Diego Lont" <[hidden email]>
| > | To: [hidden email]
| > | Sent: Monday, July 22, 2013 2:15:30 AM
| > | Subject: [Metacello] Loading problem in Seaside
| > |
| > | Hi all,
| > |
| > | I have a problem using the new Seaside3 configuration in the
| > | ConfigurationOfJQueryWidgetbox because of the following: when I
| > | try
| > | to load the stable version, it actually tries to load version:
| > | '3.1.0-gemstone'. It says this is the current version for
| > | Seaside3,
| > | because this version is fully loaded (it is empty, so
| > | mathematically
| > | speaking, that is correct).
| > |
| > | The result is an error (This version does not know the group
| > | "Core").
| > | But this is not what I want …. how should I solve this? Should I
| > | remove the gemstone specific things and distribute these over the
| > | functional groups? Or is this a (known) bug in Metacello?
| > |
| > | Diego
| > |
| > | --
| > | You received this message because you are subscribed to the
| > | Google
| > | Groups "Metacello" group.
| > | To unsubscribe from this group and stop receiving emails from it,
| > | send an email to [hidden email].
| > | For more options, visit https://groups.google.com/groups/opt_out.
| > |
| > |
| > |
| >
| > --
| > You received this message because you are subscribed to the Google
| > Groups "Metacello" group.
| > To unsubscribe from this group and stop receiving emails from it,
| > send an email to [hidden email].
| > For more options, visit https://groups.google.com/groups/opt_out.
| >
| >
|
| --
| You received this message because you are subscribed to the Google
| Groups "Metacello" group.
| To unsubscribe from this group and stop receiving emails from it,
| send an email to [hidden email].
| For more options, visit https://groups.google.com/groups/opt_out.
|
|
|

--
You received this message because you are subscribed to the Google Groups "Metacello" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/groups/opt_out.

Hi Dale,

Sorry, you need to take the version prior to Diego's changes:

Gofer new
  url:
  'http://www.smalltalkhub.com/mc/Seaside/MetacelloConfigurations/main';
  version: 'ConfigurationOfSeaside3-DiegoLont.9';
  load


On 22 Jul 2013, at 18:48, "Dale K. Henrichs" <[hidden email]> wrote:

--
You received this message because you are subscribed to the Google Groups "Metacello" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/groups/opt_out.

----- Original Message -----
| From: "Johan Brichau" <[hidden email]>
| To: [hidden email]
| Sent: Monday, July 22, 2013 8:16:58 AM
| Subject: Re: [Metacello] Loading problem in Seaside
|
| Hi Dale,
|
| To get the ConfigurationOfSeaside3 config:
|
| Gofer new
| url:
| 'http://www.smalltalkhub.com/mc/Seaside/MetacelloConfigurations/main';
| package: 'ConfigurationOfSeaside3';
| load
|
| You can probably already see that something is wrong from the result
| of:
|
| ConfigurationOfSeaside3 project latestVersion
|
| which yields: 3.1.0-gemstone
|
| The problem lies with the way the configuration for 3.1 is being
| defined: tt seems that this configuration's baseline was split up
| into methods, each containing a logical part. By tagging them each
| as a version, they can be recomposed using Metacello's inclusion
| mechanism. But I guess this is a good place to 'hack' inclusion by
| invoking the separate methods from the baseline configuration
| method. (I know you prefer not too, but what would be the good
| approach here?)
|

I've looked over the ConfigurationOfSeaside3 and the problem with the 'hack' of invoking separate methods is that any tool that you might use like Versioner will not "know" about the 'hack' and it will try to compose a monstrous method...

My long term answer is to deprecate #currentVersion. In the Metacello Scripting API, Metacello keeps track of which versions of projects are loaded and #currentVersion is no longer used.

This doesn't necessarily solve your problem in the short term, because I haven't released the Metacello Scripting API yet and I have more (minor) changes that I plan to make to the API, so I still am not prepared to "ship it".

I'm thinking that it would be very straightforward to push out a release of Metacello where we could define a new  blessing called #structural (Basically two methods need to be touched). #structural versions would be excluded from the currentVersion and latestVersion calculations ...

Since it's not likely that the literal limit problem will ever go away, I'd be inclined to create a #structural:imports: pragma so that we could introduce the notion of named structural elements without overloading the version name space ... then tools could also reason about strucural elements as well. This to me seems to be the proper solution but I would also be inclined to implement this as part of the Metacello Scripting API release ...

Let me know what you think,

Dale

--
You received this message because you are subscribed to the Google Groups "Metacello" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/groups/opt_out.

Hmmm,

From reading the configuration, I can imagine that there is a problem:)

Am I supposed to do the load in Pharo or does it fail for you in GemStone as well?

If it fails for you in GemStone, then I'll need details about GLASS version, GemStone version and Metacello version...

Dale
----- Original Message -----
| From: "Johan Brichau" <[hidden email]>
| To: [hidden email]
| Sent: Monday, July 22, 2013 10:12:18 AM
| Subject: Re: [Metacello] Loading problem in Seaside
|
| Hi Dale,
|
| Sorry, you need to take the version prior to Diego's changes:
|
| Gofer new
|   url:
|   'http://www.smalltalkhub.com/mc/Seaside/MetacelloConfigurations/main';
|   version: 'ConfigurationOfSeaside3-DiegoLont.9';
|   load
|
|
| On 22 Jul 2013, at 18:48, "Dale K. Henrichs"
| <[hidden email]> wrote:
|
| > When I run
| >
| >  ConfigurationOfSeaside3 project latestVersion
| >
| > in GemStone/S, I get:
| >
| >  3.0.8 [ConfigurationOfSeaside3]
| >
| > which is different from what you are getting ... what else do I
| > need to do?
| >
| > Dale
| >
| > ----- Original Message -----
| > | From: "Johan Brichau" <[hidden email]>
| > | To: [hidden email]
| > | Sent: Monday, July 22, 2013 8:16:58 AM
| > | Subject: Re: [Metacello] Loading problem in Seaside
| > |
| > | Hi Dale,
| > |
| > | To get the ConfigurationOfSeaside3 config:
| > |
| > | Gofer new
| > | url:
| > | 'http://www.smalltalkhub.com/mc/Seaside/MetacelloConfigurations/main';
| > | package: 'ConfigurationOfSeaside3';
| > | load
| > |
| > | You can probably already see that something is wrong from the
| > | result
| > | of:
| > |
| > | ConfigurationOfSeaside3 project latestVersion
| > |
| > | which yields: 3.1.0-gemstone
| > |
| > | The problem lies with the way the configuration for 3.1 is being
| > | defined: tt seems that this configuration's baseline was split up
| > | into methods, each containing a logical part. By tagging them
| > | each
| > | as a version, they can be recomposed using Metacello's inclusion
| > | mechanism. But I guess this is a good place to 'hack' inclusion
| > | by
| > | invoking the separate methods from the baseline configuration
| > | method. (I know you prefer not too, but what would be the good
| > | approach here?)
| > |
| > | Johan
| > |
| > |
| > | On 22 Jul 2013, at 16:30, "Dale K. Henrichs"
| > | <[hidden email]> wrote:
| > |
| > | > Diege,
| > | >
| > | > From reading the Seaside-dev list it seems to me that you
| > | > solved
| > | > this problem by removing the "empty" version?
| > | >
| > | > If not, could you give me a pointer to where the Seaside3 and
| > | > ConfigurationOfJQueryWidgetbox configurations can be found ...
| > | > a
| > | > load script that fails would be the most useful....
| > | >
| > | > But, I think your analysis is correct ... empty versions have
| > | > no
| > | > meaning and are best removed completely...
| > | >
| > | > Dale
| > | >
| > | > ----- Original Message -----
| > | > | From: "Diego Lont" <[hidden email]>
| > | > | To: [hidden email]
| > | > | Sent: Monday, July 22, 2013 2:15:30 AM
| > | > | Subject: [Metacello] Loading problem in Seaside
| > | > |
| > | > | Hi all,
| > | > |
| > | > | I have a problem using the new Seaside3 configuration in the
| > | > | ConfigurationOfJQueryWidgetbox because of the following: when
| > | > | I
| > | > | try
| > | > | to load the stable version, it actually tries to load
| > | > | version:
| > | > | '3.1.0-gemstone'. It says this is the current version for
| > | > | Seaside3,
| > | > | because this version is fully loaded (it is empty, so
| > | > | mathematically
| > | > | speaking, that is correct).
| > | > |
| > | > | The result is an error (This version does not know the group
| > | > | "Core").
| > | > | But this is not what I want …. how should I solve this?
| > | > | Should I
| > | > | remove the gemstone specific things and distribute these over
| > | > | the
| > | > | functional groups? Or is this a (known) bug in Metacello?
| > | > |
| > | > | Diego
| > | > |
| > | > | --
| > | > | You received this message because you are subscribed to the
| > | > | Google
| > | > | Groups "Metacello" group.
| > | > | To unsubscribe from this group and stop receiving emails from
| > | > | it,
| > | > | send an email to [hidden email].
| > | > | For more options, visit
| > | > | https://groups.google.com/groups/opt_out.
| > | > |
| > | > |
| > | > |
| > | >
| > | > --
| > | > You received this message because you are subscribed to the
| > | > Google
| > | > Groups "Metacello" group.
| > | > To unsubscribe from this group and stop receiving emails from
| > | > it,
| > | > send an email to [hidden email].
| > | > For more options, visit
| > | > https://groups.google.com/groups/opt_out.
| > | >
| > | >
| > |
| > | --
| > | You received this message because you are subscribed to the
| > | Google
| > | Groups "Metacello" group.
| > | To unsubscribe from this group and stop receiving emails from it,
| > | send an email to [hidden email].
| > | For more options, visit https://groups.google.com/groups/opt_out.
| > |
| > |
| > |
| >
| > --
| > You received this message because you are subscribed to the Google
| > Groups "Metacello" group.
| > To unsubscribe from this group and stop receiving emails from it,
| > send an email to [hidden email].
| > For more options, visit https://groups.google.com/groups/opt_out.
| >
| >
|
| --
| You received this message because you are subscribed to the Google
| Groups "Metacello" group.
| To unsubscribe from this group and stop receiving emails from it,
| send an email to [hidden email].
| For more options, visit https://groups.google.com/groups/opt_out.
|
|
|

--
You received this message because you are subscribed to the Google Groups "Metacello" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/groups/opt_out.

Hi Dale,

I have yet to try Seaside3.1 in Gemstone, so I don't know yet.

Diego noticed a failing load when trying to load Magritte, which references this config's #stable version. I have only been doing direct loads of '3.1.0' in Pharo, which works fine.
But I noticed the problem by sending #latestVersion which oddly told me that it was 3.1.0-gemstone.

Anyway, what you mention about #structural versions certainly sounds OK to me.

Johan

On 22 Jul 2013, at 20:30, "Dale K. Henrichs" <[hidden email]> wrote:

--
You received this message because you are subscribed to the Google Groups "Metacello" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/groups/opt_out.

Ah, so I should try this in Pharo2.0?

I'll see if I can bang out a new Metacello release with the #structural blessing, if I find that it fixes the problem, assuming I can reproduce the problem in the first place:)

Dale

----- Original Message -----
| From: "Johan Brichau" <[hidden email]>
| To: [hidden email]
| Sent: Monday, July 22, 2013 11:43:22 AM
| Subject: Re: [Metacello] Loading problem in Seaside
|
| Hi Dale,
|
| I have yet to try Seaside3.1 in Gemstone, so I don't know yet.
|
| Diego noticed a failing load when trying to load Magritte, which
| references this config's #stable version. I have only been doing
| direct loads of '3.1.0' in Pharo, which works fine.
| But I noticed the problem by sending #latestVersion which oddly told
| me that it was 3.1.0-gemstone.
|
| Anyway, what you mention about #structural versions certainly sounds
| OK to me.
|
| Johan
|
| On 22 Jul 2013, at 20:30, "Dale K. Henrichs"
| <[hidden email]> wrote:
|
| > Hmmm,
| >
| > From reading the configuration, I can imagine that there is a
| > problem:)
| >
| > Am I supposed to do the load in Pharo or does it fail for you in
| > GemStone as well?
| >
| > If it fails for you in GemStone, then I'll need details about GLASS
| > version, GemStone version and Metacello version...
| >
| > Dale
| > ----- Original Message -----
| > | From: "Johan Brichau" <[hidden email]>
| > | To: [hidden email]
| > | Sent: Monday, July 22, 2013 10:12:18 AM
| > | Subject: Re: [Metacello] Loading problem in Seaside
| > |
| > | Hi Dale,
| > |
| > | Sorry, you need to take the version prior to Diego's changes:
| > |
| > | Gofer new
| > |   url:
| > |   'http://www.smalltalkhub.com/mc/Seaside/MetacelloConfigurations/main';
| > |   version: 'ConfigurationOfSeaside3-DiegoLont.9';
| > |   load
| > |
| > |
| > | On 22 Jul 2013, at 18:48, "Dale K. Henrichs"
| > | <[hidden email]> wrote:
| > |
| > | > When I run
| > | >
| > | >  ConfigurationOfSeaside3 project latestVersion
| > | >
| > | > in GemStone/S, I get:
| > | >
| > | >  3.0.8 [ConfigurationOfSeaside3]
| > | >
| > | > which is different from what you are getting ... what else do I
| > | > need to do?
| > | >
| > | > Dale
| > | >
| > | > ----- Original Message -----
| > | > | From: "Johan Brichau" <[hidden email]>
| > | > | To: [hidden email]
| > | > | Sent: Monday, July 22, 2013 8:16:58 AM
| > | > | Subject: Re: [Metacello] Loading problem in Seaside
| > | > |
| > | > | Hi Dale,
| > | > |
| > | > | To get the ConfigurationOfSeaside3 config:
| > | > |
| > | > | Gofer new
| > | > | url:
| > | > | 'http://www.smalltalkhub.com/mc/Seaside/MetacelloConfigurations/main';
| > | > | package: 'ConfigurationOfSeaside3';
| > | > | load
| > | > |
| > | > | You can probably already see that something is wrong from the
| > | > | result
| > | > | of:
| > | > |
| > | > | ConfigurationOfSeaside3 project latestVersion
| > | > |
| > | > | which yields: 3.1.0-gemstone
| > | > |
| > | > | The problem lies with the way the configuration for 3.1 is
| > | > | being
| > | > | defined: tt seems that this configuration's baseline was
| > | > | split up
| > | > | into methods, each containing a logical part. By tagging them
| > | > | each
| > | > | as a version, they can be recomposed using Metacello's
| > | > | inclusion
| > | > | mechanism. But I guess this is a good place to 'hack'
| > | > | inclusion
| > | > | by
| > | > | invoking the separate methods from the baseline configuration
| > | > | method. (I know you prefer not too, but what would be the
| > | > | good
| > | > | approach here?)
| > | > |
| > | > | Johan
| > | > |
| > | > |
| > | > | On 22 Jul 2013, at 16:30, "Dale K. Henrichs"
| > | > | <[hidden email]> wrote:
| > | > |
| > | > | > Diege,
| > | > | >
| > | > | > From reading the Seaside-dev list it seems to me that you
| > | > | > solved
| > | > | > this problem by removing the "empty" version?
| > | > | >
| > | > | > If not, could you give me a pointer to where the Seaside3
| > | > | > and
| > | > | > ConfigurationOfJQueryWidgetbox configurations can be found
| > | > | > ...
| > | > | > a
| > | > | > load script that fails would be the most useful....
| > | > | >
| > | > | > But, I think your analysis is correct ... empty versions
| > | > | > have
| > | > | > no
| > | > | > meaning and are best removed completely...
| > | > | >
| > | > | > Dale
| > | > | >
| > | > | > ----- Original Message -----
| > | > | > | From: "Diego Lont" <[hidden email]>
| > | > | > | To: [hidden email]
| > | > | > | Sent: Monday, July 22, 2013 2:15:30 AM
| > | > | > | Subject: [Metacello] Loading problem in Seaside
| > | > | > |
| > | > | > | Hi all,
| > | > | > |
| > | > | > | I have a problem using the new Seaside3 configuration in
| > | > | > | the
| > | > | > | ConfigurationOfJQueryWidgetbox because of the following:
| > | > | > | when
| > | > | > | I
| > | > | > | try
| > | > | > | to load the stable version, it actually tries to load
| > | > | > | version:
| > | > | > | '3.1.0-gemstone'. It says this is the current version for
| > | > | > | Seaside3,
| > | > | > | because this version is fully loaded (it is empty, so
| > | > | > | mathematically
| > | > | > | speaking, that is correct).
| > | > | > |
| > | > | > | The result is an error (This version does not know the
| > | > | > | group
| > | > | > | "Core").
| > | > | > | But this is not what I want …. how should I solve this?
| > | > | > | Should I
| > | > | > | remove the gemstone specific things and distribute these
| > | > | > | over
| > | > | > | the
| > | > | > | functional groups? Or is this a (known) bug in Metacello?
| > | > | > |
| > | > | > | Diego
| > | > | > |
| > | > | > | --
| > | > | > | You received this message because you are subscribed to
| > | > | > | the
| > | > | > | Google
| > | > | > | Groups "Metacello" group.
| > | > | > | To unsubscribe from this group and stop receiving emails
| > | > | > | from
| > | > | > | it,
| > | > | > | send an email to [hidden email].
| > | > | > | For more options, visit
| > | > | > | https://groups.google.com/groups/opt_out.
| > | > | > |
| > | > | > |
| > | > | > |
| > | > | >
| > | > | > --
| > | > | > You received this message because you are subscribed to the
| > | > | > Google
| > | > | > Groups "Metacello" group.
| > | > | > To unsubscribe from this group and stop receiving emails
| > | > | > from
| > | > | > it,
| > | > | > send an email to [hidden email].
| > | > | > For more options, visit
| > | > | > https://groups.google.com/groups/opt_out.
| > | > | >
| > | > | >
| > | > |
| > | > | --
| > | > | You received this message because you are subscribed to the
| > | > | Google
| > | > | Groups "Metacello" group.
| > | > | To unsubscribe from this group and stop receiving emails from
| > | > | it,
| > | > | send an email to [hidden email].
| > | > | For more options, visit
| > | > | https://groups.google.com/groups/opt_out.
| > | > |
| > | > |
| > | > |
| > | >
| > | > --
| > | > You received this message because you are subscribed to the
| > | > Google
| > | > Groups "Metacello" group.
| > | > To unsubscribe from this group and stop receiving emails from
| > | > it,
| > | > send an email to [hidden email].
| > | > For more options, visit
| > | > https://groups.google.com/groups/opt_out.
| > | >
| > | >
| > |
| > | --
| > | You received this message because you are subscribed to the
| > | Google
| > | Groups "Metacello" group.
| > | To unsubscribe from this group and stop receiving emails from it,
| > | send an email to [hidden email].
| > | For more options, visit https://groups.google.com/groups/opt_out.
| > |
| > |
| > |
| >
| > --
| > You received this message because you are subscribed to the Google
| > Groups "Metacello" group.
| > To unsubscribe from this group and stop receiving emails from it,
| > send an email to [hidden email].
| > For more options, visit https://groups.google.com/groups/opt_out.
| >
| >
|
| --
| You received this message because you are subscribed to the Google
| Groups "Metacello" group.
| To unsubscribe from this group and stop receiving emails from it,
| send an email to [hidden email].
| For more options, visit https://groups.google.com/groups/opt_out.
|
|
|

--
You received this message because you are subscribed to the Google Groups "Metacello" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/groups/opt_out.

On 22 Jul 2013, at 20:51, "Dale K. Henrichs" <[hidden email]> wrote:

> Ah, so I should try this in Pharo2.0?

I'm having this both in 1.4 and 2.0

mind again: use this version of ConfigurationOfSeaside3: ConfigurationOfSeaside3-DiegoLont.9

ConfigurationOfSeaside3 project latestVersion load
-> error because it tries to load 3.1.0-gemstone

(ConfigurationOfSeaside3 project version:'3.1.0') load
-> loads just fine


--
You received this message because you are subscribed to the Google Groups "Metacello" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/groups/opt_out.

Okay ... I'll jump on this guy right away...

Dale

----- Original Message -----
| From: "Johan Brichau" <[hidden email]>
| To: [hidden email]
| Sent: Monday, July 22, 2013 12:08:50 PM
| Subject: Re: [Metacello] Loading problem in Seaside
|
|
| On 22 Jul 2013, at 20:51, "Dale K. Henrichs"
| <[hidden email]> wrote:
|
| > Ah, so I should try this in Pharo2.0?
|
| I'm having this both in 1.4 and 2.0
|
| mind again: use this version of ConfigurationOfSeaside3:
| ConfigurationOfSeaside3-DiegoLont.9
|
| ConfigurationOfSeaside3 project latestVersion load
| -> error because it tries to load 3.1.0-gemstone
|
| (ConfigurationOfSeaside3 project version:'3.1.0') load
| -> loads just fine
|
|
| --
| You received this message because you are subscribed to the Google
| Groups "Metacello" group.
| To unsubscribe from this group and stop receiving emails from it,
| send an email to [hidden email].
| For more options, visit https://groups.google.com/groups/opt_out.
|
|
|

--
You received this message because you are subscribed to the Google Groups "Metacello" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/groups/opt_out.

Do you think it's a bug? Isn't it because all these methods are defined as versions and that 3.1.0-gemstone is the alphanumerically-sorted last one?
So, I expect this. For me, it's an issue of needing the #structural versions as you say.

So, we can temporarily patch using the 'hack' and once there is a better solution, we can fix this
no?

On 22 Jul 2013, at 21:26, "Dale K. Henrichs" <[hidden email]> wrote:

--
You received this message because you are subscribed to the Google Groups "Metacello" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/groups/opt_out.

Bug/missing feature. Short answer is yes. Long answer is...

The method literal limit is the real culprit here ...

A version in Metacello is intended to be specified in a single method (structure).

Using a version as a pure structural element starts as a workaround for the method literal limit bug and ends up being a fairly nice way to organize related packages within a class without using pure messages sends (which a tool can't see) ... so platform-specific (structural) versions are created.

Using a platform-specific (structural) version method exposes a bug in Metacello that a version with no specs will match the criterion for #currentVersion and/or #latestVersion. I've spent a fair amount of time fighting #currentVersion/#latestVersion and I came to the conclusion that the battle of refining the #currentVersion/#latestVersion to accomodate all of the different packageprojects combinations encountered would be a constant battle that would never be won, so I decided to deprecate the #currentVersion/#latestVersion methods.

The currentVersion of a project should be defined as the version that the devloper loaded into his/her image....and in the Metacelo Preview that is what is done.

Deprecating #currentVersion/#latestVersion is not an option for pre-Metacello Preview versions of Metacello, so I need to come up with a patch that addresses the issue.

Using a #structural blessing is just a special case of the #broken blessing, just a tad bit less restrictive as it will only apply to the #currentVersion/#latestVersion calculations, which is exactly what we're looking for here.

The #structural blessing is very cheap and safe to implement. Any other solution will have broader impacts...

Breaking a version specification up into potentially reusable components is a desirable feature in it's own right, but leaning on the blessing for this is not really appropriate as the version name space is still polluted with structural version names ...

In the Metacello Preview I am moving away from using blessings to denote baselines and have added a <baseline:> pragma. Baselines and Structure deserve explicit declarations and blessings are not the right model to use (what happens if I have a structural baseline? BTW, baselines are already excluded from #currentVersion and #latestVersion calculations)...So it will be natural to introduce a <structural:imports:> pragma...

Dale

----- Original Message -----
| From: "Johan Brichau" <[hidden email]>
| To: [hidden email]
| Sent: Monday, July 22, 2013 12:35:57 PM
| Subject: Re: [Metacello] Loading problem in Seaside
|
| Do you think it's a bug? Isn't it because all these methods are
| defined as versions and that 3.1.0-gemstone is the
| alphanumerically-sorted last one?
| So, I expect this. For me, it's an issue of needing the #structural
| versions as you say.
|
| So, we can temporarily patch using the 'hack' and once there is a
| better solution, we can fix this
| no?
|
| On 22 Jul 2013, at 21:26, "Dale K. Henrichs"
| <[hidden email]> wrote:
|
| > Okay ... I'll jump on this guy right away...
| >
| > Dale
| >
| > ----- Original Message -----
| > | From: "Johan Brichau" <[hidden email]>
| > | To: [hidden email]
| > | Sent: Monday, July 22, 2013 12:08:50 PM
| > | Subject: Re: [Metacello] Loading problem in Seaside
| > |
| > |
| > | On 22 Jul 2013, at 20:51, "Dale K. Henrichs"
| > | <[hidden email]> wrote:
| > |
| > | > Ah, so I should try this in Pharo2.0?
| > |
| > | I'm having this both in 1.4 and 2.0
| > |
| > | mind again: use this version of ConfigurationOfSeaside3:
| > | ConfigurationOfSeaside3-DiegoLont.9
| > |
| > | ConfigurationOfSeaside3 project latestVersion load
| > | -> error because it tries to load 3.1.0-gemstone
| > |
| > | (ConfigurationOfSeaside3 project version:'3.1.0') load
| > | -> loads just fine
| > |
| > |
| > | --
| > | You received this message because you are subscribed to the
| > | Google
| > | Groups "Metacello" group.
| > | To unsubscribe from this group and stop receiving emails from it,
| > | send an email to [hidden email].
| > | For more options, visit https://groups.google.com/groups/opt_out.
| > |
| > |
| > |
| >
| > --
| > You received this message because you are subscribed to the Google
| > Groups "Metacello" group.
| > To unsubscribe from this group and stop receiving emails from it,
| > send an email to [hidden email].
| > For more options, visit https://groups.google.com/groups/opt_out.
| >
| >
|
| --
| You received this message because you are subscribed to the Google
| Groups "Metacello" group.
| To unsubscribe from this group and stop receiving emails from it,
| send an email to [hidden email].
| For more options, visit https://groups.google.com/groups/opt_out.
|
|
|

--
You received this message because you are subscribed to the Google Groups "Metacello" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/groups/opt_out.

Johan, Diego,

Well, in theory the blessing approach could have worked:), but in practice it does not, I can get #currentVersion to be correct, but not #latestVersion and theat isn't really acceptable...

I'll have to sacrifice some more brain cells to the cause:)

Dale

----- Original Message -----
| From: "Dale K. Henrichs" <[hidden email]>
| To: [hidden email]
| Sent: Monday, July 22, 2013 1:18:40 PM
| Subject: Re: [Metacello] Loading problem in Seaside
|
| Bug/missing feature. Short answer is yes. Long answer is...
|
| The method literal limit is the real culprit here ...
|
| A version in Metacello is intended to be specified in a single method
| (structure).
|
| Using a version as a pure structural element starts as a workaround
| for the method literal limit bug and ends up being a fairly nice way
| to organize related packages within a class without using pure
| messages sends (which a tool can't see) ... so platform-specific
| (structural) versions are created.
|
| Using a platform-specific (structural) version method exposes a bug
| in Metacello that a version with no specs will match the criterion
| for #currentVersion and/or #latestVersion. I've spent a fair amount
| of time fighting #currentVersion/#latestVersion and I came to the
| conclusion that the battle of refining the
| #currentVersion/#latestVersion to accomodate all of the different
| packageprojects combinations encountered would be a constant battle
| that would never be won, so I decided to deprecate the
| #currentVersion/#latestVersion methods.
|
| The currentVersion of a project should be defined as the version that
| the devloper loaded into his/her image....and in the Metacelo
| Preview that is what is done.
|
| Deprecating #currentVersion/#latestVersion is not an option for
| pre-Metacello Preview versions of Metacello, so I need to come up
| with a patch that addresses the issue.
|
| Using a #structural blessing is just a special case of the #broken
| blessing, just a tad bit less restrictive as it will only apply to
| the #currentVersion/#latestVersion calculations, which is exactly
| what we're looking for here.
|
| The #structural blessing is very cheap and safe to implement. Any
| other solution will have broader impacts...
|
| Breaking a version specification up into potentially reusable
| components is a desirable feature in it's own right, but leaning on
| the blessing for this is not really appropriate as the version name
| space is still polluted with structural version names ...
|
| In the Metacello Preview I am moving away from using blessings to
| denote baselines and have added a <baseline:> pragma. Baselines and
| Structure deserve explicit declarations and blessings are not the
| right model to use (what happens if I have a structural baseline?
| BTW, baselines are already excluded from #currentVersion and
| #latestVersion calculations)...So it will be natural to introduce a
| <structural:imports:> pragma...
|
| Dale
|
| ----- Original Message -----
| | From: "Johan Brichau" <[hidden email]>
| | To: [hidden email]
| | Sent: Monday, July 22, 2013 12:35:57 PM
| | Subject: Re: [Metacello] Loading problem in Seaside
| |
| | Do you think it's a bug? Isn't it because all these methods are
| | defined as versions and that 3.1.0-gemstone is the
| | alphanumerically-sorted last one?
| | So, I expect this. For me, it's an issue of needing the #structural
| | versions as you say.
| |
| | So, we can temporarily patch using the 'hack' and once there is a
| | better solution, we can fix this
| | no?
| |
| | On 22 Jul 2013, at 21:26, "Dale K. Henrichs"
| | <[hidden email]> wrote:
| |
| | > Okay ... I'll jump on this guy right away...
| | >
| | > Dale
| | >
| | > ----- Original Message -----
| | > | From: "Johan Brichau" <[hidden email]>
| | > | To: [hidden email]
| | > | Sent: Monday, July 22, 2013 12:08:50 PM
| | > | Subject: Re: [Metacello] Loading problem in Seaside
| | > |
| | > |
| | > | On 22 Jul 2013, at 20:51, "Dale K. Henrichs"
| | > | <[hidden email]> wrote:
| | > |
| | > | > Ah, so I should try this in Pharo2.0?
| | > |
| | > | I'm having this both in 1.4 and 2.0
| | > |
| | > | mind again: use this version of ConfigurationOfSeaside3:
| | > | ConfigurationOfSeaside3-DiegoLont.9
| | > |
| | > | ConfigurationOfSeaside3 project latestVersion load
| | > | -> error because it tries to load 3.1.0-gemstone
| | > |
| | > | (ConfigurationOfSeaside3 project version:'3.1.0') load
| | > | -> loads just fine
| | > |
| | > |
| | > | --
| | > | You received this message because you are subscribed to the
| | > | Google
| | > | Groups "Metacello" group.
| | > | To unsubscribe from this group and stop receiving emails from
| | > | it,
| | > | send an email to [hidden email].
| | > | For more options, visit
| | > | https://groups.google.com/groups/opt_out.
| | > |
| | > |
| | > |
| | >
| | > --
| | > You received this message because you are subscribed to the
| | > Google
| | > Groups "Metacello" group.
| | > To unsubscribe from this group and stop receiving emails from it,
| | > send an email to [hidden email].
| | > For more options, visit https://groups.google.com/groups/opt_out.
| | >
| | >
| |
| | --
| | You received this message because you are subscribed to the Google
| | Groups "Metacello" group.
| | To unsubscribe from this group and stop receiving emails from it,
| | send an email to [hidden email].
| | For more options, visit https://groups.google.com/groups/opt_out.
| |
| |
| |
|

--
You received this message because you are subscribed to the Google Groups "Metacello" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/groups/opt_out.

Johan, Diego,

I was premature in declaring the blessing approach dead. I have been able to get both #currentVersion and #latestVersion to work using the #structural blessing (modifying only 3 methods ... 4 if you count #printOn:).

Patching Pharo2.0 will be a bit complicated, since Pharo2.0 out of the box has a custom set of Metacello packages that are owned by Pharo and not me and no ConfigurationOfMetacello for doing an easy upgrade.

At the moment I only support the Metacello Preview on Pharo2.0 (thanks to Christophe Demarey!), and it is asking a bit much to load the Metacello Preview into Pharo2.0 just to load Seaside ....

Sooo, I plan to publish a new version of Metacello (1.0-beta.31.1.8) with the fix for this bug/feature[1]. 1.0-beta.31.1.8 can be loaded into Pharo1.4 and Squeak4.x.

I will also publish a new version of Metacello Preview (1.0.0-beta.32.9) with the fix for this bug/feature[1]. 1.0.0-beta.32.9 can be loaded into Pharo1.4, Squeak4.x Pharo2.0 and GemStone (the Metacello Preview is the standard version of Metacello on GemStone) ...

So that leaves the version of Metacello installed in the standard version of Pharo2.0 unhandled. The patch is simple enough that it could be slip-streamed into the Pharo2.0 release process....I've attached the patch for Issue #188 and the changes that I made to Diego's version of the ConfigurationOfSeaside3 (change `spec blessing: #development` to `spec blessing: #structural` in all of the 31 version methods and add a #common section to #version310:). To tide you guys over until an official Pharo2.0 patch can be made available...

I've run out of time today but the patch (loadable into any recent version of Metacello or Metacello Preview or Pharo2.0) should get you through tonight and I should be able to publish the new versions of Metacello by end of day tomorrow. Can't answer for timing of getting patch into Pharo2.0 (short of loading the Metacello Preview.

Hope this helps,

Dale

[1] https://github.com/dalehenrich/metacello-work/issues/188
----- Original Message -----
| From: "Dale K. Henrichs" <[hidden email]>
| To: [hidden email]
| Sent: Monday, July 22, 2013 3:24:46 PM
| Subject: Re: [Metacello] Loading problem in Seaside
|
| Johan, Diego,
|
| Well, in theory the blessing approach could have worked:), but in
| practice it does not, I can get #currentVersion to be correct, but
| not #latestVersion and theat isn't really acceptable...
|
| I'll have to sacrifice some more brain cells to the cause:)
|
| Dale
|
| ----- Original Message -----
| | From: "Dale K. Henrichs" <[hidden email]>
| | To: [hidden email]
| | Sent: Monday, July 22, 2013 1:18:40 PM
| | Subject: Re: [Metacello] Loading problem in Seaside
| |
| | Bug/missing feature. Short answer is yes. Long answer is...
| |
| | The method literal limit is the real culprit here ...
| |
| | A version in Metacello is intended to be specified in a single
| | method
| | (structure).
| |
| | Using a version as a pure structural element starts as a workaround
| | for the method literal limit bug and ends up being a fairly nice
| | way
| | to organize related packages within a class without using pure
| | messages sends (which a tool can't see) ... so platform-specific
| | (structural) versions are created.
| |
| | Using a platform-specific (structural) version method exposes a bug
| | in Metacello that a version with no specs will match the criterion
| | for #currentVersion and/or #latestVersion. I've spent a fair amount
| | of time fighting #currentVersion/#latestVersion and I came to the
| | conclusion that the battle of refining the
| | #currentVersion/#latestVersion to accomodate all of the different
| | packageprojects combinations encountered would be a constant battle
| | that would never be won, so I decided to deprecate the
| | #currentVersion/#latestVersion methods.
| |
| | The currentVersion of a project should be defined as the version
| | that
| | the devloper loaded into his/her image....and in the Metacelo
| | Preview that is what is done.
| |
| | Deprecating #currentVersion/#latestVersion is not an option for
| | pre-Metacello Preview versions of Metacello, so I need to come up
| | with a patch that addresses the issue.
| |
| | Using a #structural blessing is just a special case of the #broken
| | blessing, just a tad bit less restrictive as it will only apply to
| | the #currentVersion/#latestVersion calculations, which is exactly
| | what we're looking for here.
| |
| | The #structural blessing is very cheap and safe to implement. Any
| | other solution will have broader impacts...
| |
| | Breaking a version specification up into potentially reusable
| | components is a desirable feature in it's own right, but leaning on
| | the blessing for this is not really appropriate as the version name
| | space is still polluted with structural version names ...
| |
| | In the Metacello Preview I am moving away from using blessings to
| | denote baselines and have added a <baseline:> pragma. Baselines and
| | Structure deserve explicit declarations and blessings are not the
| | right model to use (what happens if I have a structural baseline?
| | BTW, baselines are already excluded from #currentVersion and
| | #latestVersion calculations)...So it will be natural to introduce a
| | <structural:imports:> pragma...
| |
| | Dale
| |
| | ----- Original Message -----
| | | From: "Johan Brichau" <[hidden email]>
| | | To: [hidden email]
| | | Sent: Monday, July 22, 2013 12:35:57 PM
| | | Subject: Re: [Metacello] Loading problem in Seaside
| | |
| | | Do you think it's a bug? Isn't it because all these methods are
| | | defined as versions and that 3.1.0-gemstone is the
| | | alphanumerically-sorted last one?
| | | So, I expect this. For me, it's an issue of needing the
| | | #structural
| | | versions as you say.
| | |
| | | So, we can temporarily patch using the 'hack' and once there is a
| | | better solution, we can fix this
| | | no?
| | |
| | | On 22 Jul 2013, at 21:26, "Dale K. Henrichs"
| | | <[hidden email]> wrote:
| | |
| | | > Okay ... I'll jump on this guy right away...
| | | >
| | | > Dale
| | | >
| | | > ----- Original Message -----
| | | > | From: "Johan Brichau" <[hidden email]>
| | | > | To: [hidden email]
| | | > | Sent: Monday, July 22, 2013 12:08:50 PM
| | | > | Subject: Re: [Metacello] Loading problem in Seaside
| | | > |
| | | > |
| | | > | On 22 Jul 2013, at 20:51, "Dale K. Henrichs"
| | | > | <[hidden email]> wrote:
| | | > |
| | | > | > Ah, so I should try this in Pharo2.0?
| | | > |
| | | > | I'm having this both in 1.4 and 2.0
| | | > |
| | | > | mind again: use this version of ConfigurationOfSeaside3:
| | | > | ConfigurationOfSeaside3-DiegoLont.9
| | | > |
| | | > | ConfigurationOfSeaside3 project latestVersion load
| | | > | -> error because it tries to load 3.1.0-gemstone
| | | > |
| | | > | (ConfigurationOfSeaside3 project version:'3.1.0') load
| | | > | -> loads just fine
| | | > |
| | | > |
| | | > | --
| | | > | You received this message because you are subscribed to the
| | | > | Google
| | | > | Groups "Metacello" group.
| | | > | To unsubscribe from this group and stop receiving emails from
| | | > | it,
| | | > | send an email to [hidden email].
| | | > | For more options, visit
| | | > | https://groups.google.com/groups/opt_out.
| | | > |
| | | > |
| | | > |
| | | >
| | | > --
| | | > You received this message because you are subscribed to the
| | | > Google
| | | > Groups "Metacello" group.
| | | > To unsubscribe from this group and stop receiving emails from
| | | > it,
| | | > send an email to [hidden email].
| | | > For more options, visit
| | | > https://groups.google.com/groups/opt_out.
| | | >
| | | >
| | |
| | | --
| | | You received this message because you are subscribed to the
| | | Google
| | | Groups "Metacello" group.
| | | To unsubscribe from this group and stop receiving emails from it,
| | | send an email to [hidden email].
| | | For more options, visit https://groups.google.com/groups/opt_out.
| | |
| | |
| | |
| |
|

--
You received this message because you are subscribed to the Google Groups "Metacello" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/groups/opt_out.

Diego,

If you have some time, could you give Dale's ultrafast fix a try?

fixed Issue #188: Problems using <version:imports:> for structural reasons
 https://github.com/dalehenrich/metacello-work/issues/188

On 23 Jul 2013, at 02:01, Dale K. Henrichs <[hidden email]> wrote:

--
You received this message because you are subscribed to the Google Groups "Metacello" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/groups/opt_out.

Diego,

Keep in mind that the new version of Metacello is for Pharo1.4 only ... getting Pharo2.0 updated is out of my control...

Dale

----- Original Message -----
| From: "Johan Brichau" <[hidden email]>
| To: [hidden email]
| Sent: Wednesday, July 24, 2013 12:17:31 PM
| Subject: Re: [Metacello] Loading problem in Seaside
|
| Diego,
|
| If you have some time, could you give Dale's ultrafast fix a try?
|
| fixed Issue #188: Problems using <version:imports:> for structural
| reasons
|  https://github.com/dalehenrich/metacello-work/issues/188
|
| On 23 Jul 2013, at 02:01, Dale K. Henrichs
| <[hidden email]> wrote:
|
| > Johan, Diego,
| >
| > I was premature in declaring the blessing approach dead. I have
| > been able to get both #currentVersion and #latestVersion to work
| > using the #structural blessing (modifying only 3 methods ... 4 if
| > you count #printOn:).
| >
| > Patching Pharo2.0 will be a bit complicated, since Pharo2.0 out of
| > the box has a custom set of Metacello packages that are owned by
| > Pharo and not me and no ConfigurationOfMetacello for doing an easy
| > upgrade.
| >
| > At the moment I only support the Metacello Preview on Pharo2.0
| > (thanks to Christophe Demarey!), and it is asking a bit much to
| > load the Metacello Preview into Pharo2.0 just to load Seaside ....
| >
| > Sooo, I plan to publish a new version of Metacello
| > (1.0-beta.31.1.8) with the fix for this bug/feature[1].
| > 1.0-beta.31.1.8 can be loaded into Pharo1.4 and Squeak4.x.
| >
| > I will also publish a new version of Metacello Preview
| > (1.0.0-beta.32.9) with the fix for this bug/feature[1].
| > 1.0.0-beta.32.9 can be loaded into Pharo1.4, Squeak4.x Pharo2.0
| > and GemStone (the Metacello Preview is the standard version of
| > Metacello on GemStone) ...
| >
| > So that leaves the version of Metacello installed in the standard
| > version of Pharo2.0 unhandled. The patch is simple enough that it
| > could be slip-streamed into the Pharo2.0 release process....I've
| > attached the patch for Issue #188 and the changes that I made to
| > Diego's version of the ConfigurationOfSeaside3 (change `spec
| > blessing: #development` to `spec blessing: #structural` in all of
| > the 31 version methods and add a #common section to #version310:).
| > To tide you guys over until an official Pharo2.0 patch can be made
| > available...
| >
| > I've run out of time today but the patch (loadable into any recent
| > version of Metacello or Metacello Preview or Pharo2.0) should get
| > you through tonight and I should be able to publish the new
| > versions of Metacello by end of day tomorrow. Can't answer for
| > timing of getting patch into Pharo2.0 (short of loading the
| > Metacello Preview.
| >
| > Hope this helps,
| >
| > Dale
| >
| > [1] https://github.com/dalehenrich/metacello-work/issues/188
| > ----- Original Message -----
| > | From: "Dale K. Henrichs" <[hidden email]>
| > | To: [hidden email]
| > | Sent: Monday, July 22, 2013 3:24:46 PM
| > | Subject: Re: [Metacello] Loading problem in Seaside
| > |
| > | Johan, Diego,
| > |
| > | Well, in theory the blessing approach could have worked:), but in
| > | practice it does not, I can get #currentVersion to be correct,
| > | but
| > | not #latestVersion and theat isn't really acceptable...
| > |
| > | I'll have to sacrifice some more brain cells to the cause:)
| > |
| > | Dale
| > |
| > | ----- Original Message -----
| > | | From: "Dale K. Henrichs" <[hidden email]>
| > | | To: [hidden email]
| > | | Sent: Monday, July 22, 2013 1:18:40 PM
| > | | Subject: Re: [Metacello] Loading problem in Seaside
| > | |
| > | | Bug/missing feature. Short answer is yes. Long answer is...
| > | |
| > | | The method literal limit is the real culprit here ...
| > | |
| > | | A version in Metacello is intended to be specified in a single
| > | | method
| > | | (structure).
| > | |
| > | | Using a version as a pure structural element starts as a
| > | | workaround
| > | | for the method literal limit bug and ends up being a fairly
| > | | nice
| > | | way
| > | | to organize related packages within a class without using pure
| > | | messages sends (which a tool can't see) ... so
| > | | platform-specific
| > | | (structural) versions are created.
| > | |
| > | | Using a platform-specific (structural) version method exposes a
| > | | bug
| > | | in Metacello that a version with no specs will match the
| > | | criterion
| > | | for #currentVersion and/or #latestVersion. I've spent a fair
| > | | amount
| > | | of time fighting #currentVersion/#latestVersion and I came to
| > | | the
| > | | conclusion that the battle of refining the
| > | | #currentVersion/#latestVersion to accomodate all of the
| > | | different
| > | | packageprojects combinations encountered would be a constant
| > | | battle
| > | | that would never be won, so I decided to deprecate the
| > | | #currentVersion/#latestVersion methods.
| > | |
| > | | The currentVersion of a project should be defined as the
| > | | version
| > | | that
| > | | the devloper loaded into his/her image....and in the Metacelo
| > | | Preview that is what is done.
| > | |
| > | | Deprecating #currentVersion/#latestVersion is not an option for
| > | | pre-Metacello Preview versions of Metacello, so I need to come
| > | | up
| > | | with a patch that addresses the issue.
| > | |
| > | | Using a #structural blessing is just a special case of the
| > | | #broken
| > | | blessing, just a tad bit less restrictive as it will only apply
| > | | to
| > | | the #currentVersion/#latestVersion calculations, which is
| > | | exactly
| > | | what we're looking for here.
| > | |
| > | | The #structural blessing is very cheap and safe to implement.
| > | | Any
| > | | other solution will have broader impacts...
| > | |
| > | | Breaking a version specification up into potentially reusable
| > | | components is a desirable feature in it's own right, but
| > | | leaning on
| > | | the blessing for this is not really appropriate as the version
| > | | name
| > | | space is still polluted with structural version names ...
| > | |
| > | | In the Metacello Preview I am moving away from using blessings
| > | | to
| > | | denote baselines and have added a <baseline:> pragma. Baselines
| > | | and
| > | | Structure deserve explicit declarations and blessings are not
| > | | the
| > | | right model to use (what happens if I have a structural
| > | | baseline?
| > | | BTW, baselines are already excluded from #currentVersion and
| > | | #latestVersion calculations)...So it will be natural to
| > | | introduce a
| > | | <structural:imports:> pragma...
| > | |
| > | | Dale
| > | |
| > | | ----- Original Message -----
| > | | | From: "Johan Brichau" <[hidden email]>
| > | | | To: [hidden email]
| > | | | Sent: Monday, July 22, 2013 12:35:57 PM
| > | | | Subject: Re: [Metacello] Loading problem in Seaside
| > | | |
| > | | | Do you think it's a bug? Isn't it because all these methods
| > | | | are
| > | | | defined as versions and that 3.1.0-gemstone is the
| > | | | alphanumerically-sorted last one?
| > | | | So, I expect this. For me, it's an issue of needing the
| > | | | #structural
| > | | | versions as you say.
| > | | |
| > | | | So, we can temporarily patch using the 'hack' and once there
| > | | | is a
| > | | | better solution, we can fix this
| > | | | no?
| > | | |
| > | | | On 22 Jul 2013, at 21:26, "Dale K. Henrichs"
| > | | | <[hidden email]> wrote:
| > | | |
| > | | | > Okay ... I'll jump on this guy right away...
| > | | | >
| > | | | > Dale
| > | | | >
| > | | | > ----- Original Message -----
| > | | | > | From: "Johan Brichau" <[hidden email]>
| > | | | > | To: [hidden email]
| > | | | > | Sent: Monday, July 22, 2013 12:08:50 PM
| > | | | > | Subject: Re: [Metacello] Loading problem in Seaside
| > | | | > |
| > | | | > |
| > | | | > | On 22 Jul 2013, at 20:51, "Dale K. Henrichs"
| > | | | > | <[hidden email]> wrote:
| > | | | > |
| > | | | > | > Ah, so I should try this in Pharo2.0?
| > | | | > |
| > | | | > | I'm having this both in 1.4 and 2.0
| > | | | > |
| > | | | > | mind again: use this version of ConfigurationOfSeaside3:
| > | | | > | ConfigurationOfSeaside3-DiegoLont.9
| > | | | > |
| > | | | > | ConfigurationOfSeaside3 project latestVersion load
| > | | | > | -> error because it tries to load 3.1.0-gemstone
| > | | | > |
| > | | | > | (ConfigurationOfSeaside3 project version:'3.1.0') load
| > | | | > | -> loads just fine
| > | | | > |
| > | | | > |
| > | | | > | --
| > | | | > | You received this message because you are subscribed to
| > | | | > | the
| > | | | > | Google
| > | | | > | Groups "Metacello" group.
| > | | | > | To unsubscribe from this group and stop receiving emails
| > | | | > | from
| > | | | > | it,
| > | | | > | send an email to [hidden email].
| > | | | > | For more options, visit
| > | | | > | https://groups.google.com/groups/opt_out.
| > | | | > |
| > | | | > |
| > | | | > |
| > | | | >
| > | | | > --
| > | | | > You received this message because you are subscribed to the
| > | | | > Google
| > | | | > Groups "Metacello" group.
| > | | | > To unsubscribe from this group and stop receiving emails
| > | | | > from
| > | | | > it,
| > | | | > send an email to [hidden email].
| > | | | > For more options, visit
| > | | | > https://groups.google.com/groups/opt_out.
| > | | | >
| > | | | >
| > | | |
| > | | | --
| > | | | You received this message because you are subscribed to the
| > | | | Google
| > | | | Groups "Metacello" group.
| > | | | To unsubscribe from this group and stop receiving emails from
| > | | | it,
| > | | | send an email to [hidden email].
| > | | | For more options, visit
| > | | | https://groups.google.com/groups/opt_out.
| > | | |
| > | | |
| > | | |
| > | |
| > |
| >
| > --
| > You received this message because you are subscribed to the Google
| > Groups "Metacello" group.
| > To unsubscribe from this group and stop receiving emails from it,
| > send an email to [hidden email].
| > For more options, visit https://groups.google.com/groups/opt_out.
| >
| >
| > <Issue188_fix.cs><ConfigurationOfSeaside3.cs>
|
| --
| You received this message because you are subscribed to the Google
| Groups "Metacello" group.
| To unsubscribe from this group and stop receiving emails from it,
| send an email to [hidden email].
| For more options, visit https://groups.google.com/groups/opt_out.
|
|
|

--
You received this message because you are subscribed to the Google Groups "Metacello" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/groups/opt_out.