Re: Metacello

Janko,

This isn't a waste of time:)

I agree that there are problems with the Monticello naming convention. It _isn't_ enforced by anything but the tools and the sort order has changed over time even within the same tool set...

Part of the rationale behind Metacello is to break the link between the mcz file names and versions. In Metacello, there is an explicit version number class. I've taken the version numbers from your packages, converted them to metacello version numbers and sorted the list:

(#('5.4.1' '5.4' '5.6' '5.4-beta1' '6.0' '6.0-beta.1' '6.0-beta.2')
  collect: [:str | str asMetacelloVersionNumber ]) sort: [:a :b | a <= b ]

  5.4-beta1
  5.4
  5.4.1
  5.6
  6.0-beta.1
  6.0-beta.2
  6.0

Which I believe is the correct sort order for the versions. As an aside numbers are sorted in numeric order so 9.0 sorts before 10.0. Strings are sorted using string order so -alpha sorts before -beta. Finally 6.0-anything sorts before 6.0.

Gofer is used to load packages in Metacello and the Gofer sort order based on the package names is:

  '5.4.1' '5.4' '5.6' '5.4-beta1' '6.0' '6.0-beta.1' '6.0-beta.2'

to top it off the name 'Aida6.0' looks like a package name and Gofer endeavors to load the latest version of the package which is 6.0-beta.2:(

I'm not sure that I can force Gofer to do things differently without breaking a lot of other behavior. I will look a little bit more tomorrow, but I think that we are caught between a rock and a hard place.

If you could create packages with mcz file names that conform to the packagename-author.pkgversionnumber, then the Metacello configuration could be used to represent each of the unique product versions and the package name would be less important a development artifact rather than a version artifact. I see from the package history the you use the "standard" naming convention for development so Metacello could point into your development repository...

As I said, I will look a little bit more tomorrow, but I'm not optimistic ... the computer has to have some algorithm for determining sort order and I think that the name Aida6.0 is too ambiguous (is it a base package name or an mcz file name)....

Hopefully we'll be able to figure something out.

Dale

----- "Janko Mivšek" <[hidden email]> wrote:

| Hi Dale,
|
| On 27. 01. 2010 19:12, Dale Henrichs wrote:
|
| > I will check into to see what might be going on ... I think that the
| naming convention that you use for Aida, Swazoo and Sport doesn't
| quite fit the "standard" naming conventions which is confusing
| SqueakSource and Gofer (which is used as the loader for Metacello).
|
| Thanks a lot and sorry for wasting your time due to naming convention.
|
| Which is not enforced and I always regarded as unwritten rule, but not
| very handy specially for official released versions, where author
| certainly is not so important to be in the name. The MC package
| contains
| the author among its properties anway.
|
| Also, from usability and newbies standpoint the released package name
| should be short and a new long-name author convention is direct
| oposite
| of that rule.  While for development ones it is good to have all three
| attributes in the name.
|
| Just few thoughts
| Janko
|
|
| >
| > Dale
| > ----- "Janko Mivšek" <[hidden email]> wrote:
| >
| > | Hi Dale,
| > |
| > | Thanks a lot for that configuration! I just tried it and it loads
| > | nicelly in Pharo RC2, just that it loads Aida6.0-beta.2 instead of
| > | final
| > | Aida6.0. From looking at class ConfigurationOfAida, method
| > | #version60:,
| > | I didn's see any problem. Is that Metacello wrongly determine the
| > | version to load?
| > |
| > | Ok, I just "released" and "blessed" Aida6.0.mcz on SqueakSource,
| > | maybe
| > | this will help... no, it still loads 6.0-beta.2 instead.
| > | http://www.squeaksource.com/Aida.html
| > |
| > | Funny thing is also that on SqueakSource 6.0 is now among Released
| > | and
| > | Blessed, but not among Versions, neither among Latest. What is
| wrong?
| > | I'm not very proficient with MC as you see :)
| > |
| > | Best regards
| > | Janko
| > |
| > |
| > | On 27. 01. 2010 04:20, Dale Henrichs wrote:
| > | >
| > | > ----- "Geert Claes" <[hidden email]> wrote:
| > | >
| > | > | Is there a metacello (ConfigurationOfAIDAWeb type of thing)
| > | install
| > | > | for AIDA
| > | > | Web?
| > | > | --
| > | > | View this message in context:
| > | > | http://n4.nabble.com/Metacello-tp997116p997116.html
| > | > | Sent from the AIDA/Web mailing list archive at Nabble.com.
| > | > | _______________________________________________
| > | > | Aida mailing list
| > | > | [hidden email]
| > | > | http://lists.aidaweb.si/mailman/listinfo/aida
| > | >
| > | > Geert,
| > | >
| > | > I've just committed ConfigurationOfAida to the
| MetacelloRepository.
| > | >
| > | > "on Pharo"
| > | > Gofer new
| > | > squeaksource: 'MetacelloRepository';
| > | > package: 'ConfigurationOfAida';
| > | > load.
| > | >
| > | > ((Smalltalk at: #ConfigurationOfAida) project version: '6.0')
| load.
| > | >
| > | > "on Squeak trunk"
| > | > (Installer ss project: 'MetacelloRepository')
| > | >     install: 'ConfigurationOfAida'.
| > | > ((Smalltalk at: #ConfigurationOfAida) project version: '6.0')
| load.
| > | >
| > | > Adding GemStone to the configuration is on my todo list.
| > | >
| > | > Dale
| > | > _______________________________________________
| > | > Aida mailing list
| > | > [hidden email]
| > | > http://lists.aidaweb.si/mailman/listinfo/aida
| > | >
| > |
| > | --
| > | Janko Mivšek
| > | AIDA/Web
| > | Smalltalk Web Application Server
| > | http://www.aidaweb.si
| >
|
| --
| Janko Mivšek
| AIDA/Web
| Smalltalk Web Application Server
| http://www.aidaweb.si
| _______________________________________________
| Aida mailing list
| [hidden email]
| http://lists.aidaweb.si/mailman/listinfo/aida
_______________________________________________
Aida mailing list
[hidden email]
http://lists.aidaweb.si/mailman/listinfo/aida

Dale,

so the quick and dirty solution would be to repost Aida as
Aida6.0-jm.mcz? Let me try, I need to repost anyway, because the
connection to the parent version was somehow lost...

Best regards
Janko


On 28. 01. 2010 07:02, Dale Henrichs wrote:
--
Janko Mivšek
AIDA/Web
Smalltalk Web Application Server
http://www.aidaweb.si
_______________________________________________
Aida mailing list
[hidden email]
http://lists.aidaweb.si/mailman/listinfo/aida

Janko,

Yes that should do the trick. As soon as you've posted the new mcz file, I'll update the configuration.

Dale
----- "Janko Mivšek" <[hidden email]> wrote:

| Dale,
|
| so the quick and dirty solution would be to repost Aida as
| Aida6.0-jm.mcz? Let me try, I need to repost anyway, because the
| connection to the parent version was somehow lost...
|
| Best regards
| Janko
|
|
| On 28. 01. 2010 07:02, Dale Henrichs wrote:
| > Janko,
| >
| > This isn't a waste of time:)
| >
| > I agree that there are problems with the Monticello naming
| convention. It _isn't_ enforced by anything but the tools and the sort
| order has changed over time even within the same tool set...
| >
| > Part of the rationale behind Metacello is to break the link between
| the mcz file names and versions. In Metacello, there is an explicit
| version number class. I've taken the version numbers from your
| packages, converted them to metacello version numbers and sorted the
| list:
| >
| > (#('5.4.1' '5.4' '5.6' '5.4-beta1' '6.0' '6.0-beta.1' '6.0-beta.2')
|
| >   collect: [:str | str asMetacelloVersionNumber ]) sort: [:a :b | a
| <= b ]
| >
| >   5.4-beta1
| >   5.4
| >   5.4.1
| >   5.6
| >   6.0-beta.1
| >   6.0-beta.2
| >   6.0
| >
| > Which I believe is the correct sort order for the versions. As an
| aside numbers are sorted in numeric order so 9.0 sorts before 10.0.
| Strings are sorted using string order so -alpha sorts before -beta.
| Finally 6.0-anything sorts before 6.0.
| >
| > Gofer is used to load packages in Metacello and the Gofer sort order
| based on the package names is:
| >
| >   '5.4.1' '5.4' '5.6' '5.4-beta1' '6.0' '6.0-beta.1' '6.0-beta.2'
| >
| > to top it off the name 'Aida6.0' looks like a package name and Gofer
| endeavors to load the latest version of the package which is
| 6.0-beta.2:(
| >
| > I'm not sure that I can force Gofer to do things differently without
| breaking a lot of other behavior. I will look a little bit more
| tomorrow, but I think that we are caught between a rock and a hard
| place.
| >
| > If you could create packages with mcz file names that conform to the
| packagename-author.pkgversionnumber, then the Metacello configuration
| could be used to represent each of the unique product versions and the
| package name would be less important a development artifact rather
| than a version artifact. I see from the package history the you use
| the "standard" naming convention for development so Metacello could
| point into your development repository...
| >
| > As I said, I will look a little bit more tomorrow, but I'm not
| optimistic ... the computer has to have some algorithm for determining
| sort order and I think that the name Aida6.0 is too ambiguous (is it a
| base package name or an mcz file name)....
| >
| > Hopefully we'll be able to figure something out.
| >
| > Dale
| >
| > ----- "Janko Mivšek" <[hidden email]> wrote:
| >
| > | Hi Dale,
| > |
| > | On 27. 01. 2010 19:12, Dale Henrichs wrote:
| > |
| > | > I will check into to see what might be going on ... I think that
| the
| > | naming convention that you use for Aida, Swazoo and Sport doesn't
| > | quite fit the "standard" naming conventions which is confusing
| > | SqueakSource and Gofer (which is used as the loader for
| Metacello).
| > |
| > | Thanks a lot and sorry for wasting your time due to naming
| convention.
| > |
| > | Which is not enforced and I always regarded as unwritten rule, but
| not
| > | very handy specially for official released versions, where author
| > | certainly is not so important to be in the name. The MC package
| > | contains
| > | the author among its properties anway.
| > |
| > | Also, from usability and newbies standpoint the released package
| name
| > | should be short and a new long-name author convention is direct
| > | oposite
| > | of that rule.  While for development ones it is good to have all
| three
| > | attributes in the name.
| > |
| > | Just few thoughts
| > | Janko
| > |
| > |
| > | >
| > | > Dale
| > | > ----- "Janko Mivšek" <[hidden email]> wrote:
| > | >
| > | > | Hi Dale,
| > | > |
| > | > | Thanks a lot for that configuration! I just tried it and it
| loads
| > | > | nicelly in Pharo RC2, just that it loads Aida6.0-beta.2
| instead of
| > | > | final
| > | > | Aida6.0. From looking at class ConfigurationOfAida, method
| > | > | #version60:,
| > | > | I didn's see any problem. Is that Metacello wrongly determine
| the
| > | > | version to load?
| > | > |
| > | > | Ok, I just "released" and "blessed" Aida6.0.mcz on
| SqueakSource,
| > | > | maybe
| > | > | this will help... no, it still loads 6.0-beta.2 instead.
| > | > | http://www.squeaksource.com/Aida.html
| > | > |
| > | > | Funny thing is also that on SqueakSource 6.0 is now among
| Released
| > | > | and
| > | > | Blessed, but not among Versions, neither among Latest. What
| is
| > | wrong?
| > | > | I'm not very proficient with MC as you see :)
| > | > |
| > | > | Best regards
| > | > | Janko
| > | > |
| > | > |
| > | > | On 27. 01. 2010 04:20, Dale Henrichs wrote:
| > | > | >
| > | > | > ----- "Geert Claes" <[hidden email]> wrote:
| > | > | >
| > | > | > | Is there a metacello (ConfigurationOfAIDAWeb type of
| thing)
| > | > | install
| > | > | > | for AIDA
| > | > | > | Web?
| > | > | > | --
| > | > | > | View this message in context:
| > | > | > | http://n4.nabble.com/Metacello-tp997116p997116.html
| > | > | > | Sent from the AIDA/Web mailing list archive at
| Nabble.com.
| > | > | > | _______________________________________________
| > | > | > | Aida mailing list
| > | > | > | [hidden email]
| > | > | > | http://lists.aidaweb.si/mailman/listinfo/aida
| > | > | >
| > | > | > Geert,
| > | > | >
| > | > | > I've just committed ConfigurationOfAida to the
| > | MetacelloRepository.
| > | > | >
| > | > | > "on Pharo"
| > | > | > Gofer new
| > | > | > squeaksource: 'MetacelloRepository';
| > | > | > package: 'ConfigurationOfAida';
| > | > | > load.
| > | > | >
| > | > | > ((Smalltalk at: #ConfigurationOfAida) project version:
| '6.0')
| > | load.
| > | > | >
| > | > | > "on Squeak trunk"
| > | > | > (Installer ss project: 'MetacelloRepository')
| > | > | >     install: 'ConfigurationOfAida'.
| > | > | > ((Smalltalk at: #ConfigurationOfAida) project version:
| '6.0')
| > | load.
| > | > | >
| > | > | > Adding GemStone to the configuration is on my todo list.
| > | > | >
| > | > | > Dale
| > | > | > _______________________________________________
| > | > | > Aida mailing list
| > | > | > [hidden email]
| > | > | > http://lists.aidaweb.si/mailman/listinfo/aida
| > | > | >
| > | > |
| > | > | --
| > | > | Janko Mivšek
| > | > | AIDA/Web
| > | > | Smalltalk Web Application Server
| > | > | http://www.aidaweb.si
| > | >
| > |
| > | --
| > | Janko Mivšek
| > | AIDA/Web
| > | Smalltalk Web Application Server
| > | http://www.aidaweb.si
| > | _______________________________________________
| > | Aida mailing list
| > | [hidden email]
| > | http://lists.aidaweb.si/mailman/listinfo/aida
| > _______________________________________________
| > Aida mailing list
| > [hidden email]
| > http://lists.aidaweb.si/mailman/listinfo/aida
|
| --
| Janko Mivšek
| AIDA/Web
| Smalltalk Web Application Server
| http://www.aidaweb.si
_______________________________________________
Aida mailing list
[hidden email]
http://lists.aidaweb.si/mailman/listinfo/aida

Janko,

Actually the package name should be something like:

  Aida6.0-jm.3.mcz

That way the package version will sort greater than 'Aida6.0-beta.2.mcz':)

Dale
----- "Dale Henrichs" <[hidden email]> wrote:

| Janko,
|
| Yes that should do the trick. As soon as you've posted the new mcz
| file, I'll update the configuration.
|
| Dale
| ----- "Janko Mivšek" <[hidden email]> wrote:
|
| | Dale,
| |
| | so the quick and dirty solution would be to repost Aida as
| | Aida6.0-jm.mcz? Let me try, I need to repost anyway, because the
| | connection to the parent version was somehow lost...
| |
| | Best regards
| | Janko
| |
| |
| | On 28. 01. 2010 07:02, Dale Henrichs wrote:
| | > Janko,
| | >
| | > This isn't a waste of time:)
| | >
| | > I agree that there are problems with the Monticello naming
| | convention. It _isn't_ enforced by anything but the tools and the
| sort
| | order has changed over time even within the same tool set...
| | >
| | > Part of the rationale behind Metacello is to break the link
| between
| | the mcz file names and versions. In Metacello, there is an explicit
| | version number class. I've taken the version numbers from your
| | packages, converted them to metacello version numbers and sorted the
| | list:
| | >
| | > (#('5.4.1' '5.4' '5.6' '5.4-beta1' '6.0' '6.0-beta.1'
| '6.0-beta.2')
| |
| | >   collect: [:str | str asMetacelloVersionNumber ]) sort: [:a :b |
| a
| | <= b ]
| | >
| | >   5.4-beta1
| | >   5.4
| | >   5.4.1
| | >   5.6
| | >   6.0-beta.1
| | >   6.0-beta.2
| | >   6.0
| | >
| | > Which I believe is the correct sort order for the versions. As an
| | aside numbers are sorted in numeric order so 9.0 sorts before 10.0.
| | Strings are sorted using string order so -alpha sorts before -beta.
| | Finally 6.0-anything sorts before 6.0.
| | >
| | > Gofer is used to load packages in Metacello and the Gofer sort
| order
| | based on the package names is:
| | >
| | >   '5.4.1' '5.4' '5.6' '5.4-beta1' '6.0' '6.0-beta.1' '6.0-beta.2'
| | >
| | > to top it off the name 'Aida6.0' looks like a package name and
| Gofer
| | endeavors to load the latest version of the package which is
| | 6.0-beta.2:(
| | >
| | > I'm not sure that I can force Gofer to do things differently
| without
| | breaking a lot of other behavior. I will look a little bit more
| | tomorrow, but I think that we are caught between a rock and a hard
| | place.
| | >
| | > If you could create packages with mcz file names that conform to
| the
| | packagename-author.pkgversionnumber, then the Metacello
| configuration
| | could be used to represent each of the unique product versions and
| the
| | package name would be less important a development artifact rather
| | than a version artifact. I see from the package history the you use
| | the "standard" naming convention for development so Metacello could
| | point into your development repository...
| | >
| | > As I said, I will look a little bit more tomorrow, but I'm not
| | optimistic ... the computer has to have some algorithm for
| determining
| | sort order and I think that the name Aida6.0 is too ambiguous (is it
| a
| | base package name or an mcz file name)....
| | >
| | > Hopefully we'll be able to figure something out.
| | >
| | > Dale
| | >
| | > ----- "Janko Mivšek" <[hidden email]> wrote:
| | >
| | > | Hi Dale,
| | > |
| | > | On 27. 01. 2010 19:12, Dale Henrichs wrote:
| | > |
| | > | > I will check into to see what might be going on ... I think
| that
| | the
| | > | naming convention that you use for Aida, Swazoo and Sport
| doesn't
| | > | quite fit the "standard" naming conventions which is confusing
| | > | SqueakSource and Gofer (which is used as the loader for
| | Metacello).
| | > |
| | > | Thanks a lot and sorry for wasting your time due to naming
| | convention.
| | > |
| | > | Which is not enforced and I always regarded as unwritten rule,
| but
| | not
| | > | very handy specially for official released versions, where
| author
| | > | certainly is not so important to be in the name. The MC package
| | > | contains
| | > | the author among its properties anway.
| | > |
| | > | Also, from usability and newbies standpoint the released package
| | name
| | > | should be short and a new long-name author convention is direct
| | > | oposite
| | > | of that rule.  While for development ones it is good to have all
| | three
| | > | attributes in the name.
| | > |
| | > | Just few thoughts
| | > | Janko
| | > |
| | > |
| | > | >
| | > | > Dale
| | > | > ----- "Janko Mivšek" <[hidden email]> wrote:
| | > | >
| | > | > | Hi Dale,
| | > | > |
| | > | > | Thanks a lot for that configuration! I just tried it and it
| | loads
| | > | > | nicelly in Pharo RC2, just that it loads Aida6.0-beta.2
| | instead of
| | > | > | final
| | > | > | Aida6.0. From looking at class ConfigurationOfAida, method
| | > | > | #version60:,
| | > | > | I didn's see any problem. Is that Metacello wrongly
| determine
| | the
| | > | > | version to load?
| | > | > |
| | > | > | Ok, I just "released" and "blessed" Aida6.0.mcz on
| | SqueakSource,
| | > | > | maybe
| | > | > | this will help... no, it still loads 6.0-beta.2 instead.
| | > | > | http://www.squeaksource.com/Aida.html
| | > | > |
| | > | > | Funny thing is also that on SqueakSource 6.0 is now among
| | Released
| | > | > | and
| | > | > | Blessed, but not among Versions, neither among Latest. What
| | is
| | > | wrong?
| | > | > | I'm not very proficient with MC as you see :)
| | > | > |
| | > | > | Best regards
| | > | > | Janko
| | > | > |
| | > | > |
| | > | > | On 27. 01. 2010 04:20, Dale Henrichs wrote:
| | > | > | >
| | > | > | > ----- "Geert Claes" <[hidden email]> wrote:
| | > | > | >
| | > | > | > | Is there a metacello (ConfigurationOfAIDAWeb type of
| | thing)
| | > | > | install
| | > | > | > | for AIDA
| | > | > | > | Web?
| | > | > | > | --
| | > | > | > | View this message in context:
| | > | > | > | http://n4.nabble.com/Metacello-tp997116p997116.html
| | > | > | > | Sent from the AIDA/Web mailing list archive at
| | Nabble.com.
| | > | > | > | _______________________________________________
| | > | > | > | Aida mailing list
| | > | > | > | [hidden email]
| | > | > | > | http://lists.aidaweb.si/mailman/listinfo/aida
| | > | > | >
| | > | > | > Geert,
| | > | > | >
| | > | > | > I've just committed ConfigurationOfAida to the
| | > | MetacelloRepository.
| | > | > | >
| | > | > | > "on Pharo"
| | > | > | > Gofer new
| | > | > | > squeaksource: 'MetacelloRepository';
| | > | > | > package: 'ConfigurationOfAida';
| | > | > | > load.
| | > | > | >
| | > | > | > ((Smalltalk at: #ConfigurationOfAida) project version:
| | '6.0')
| | > | load.
| | > | > | >
| | > | > | > "on Squeak trunk"
| | > | > | > (Installer ss project: 'MetacelloRepository')
| | > | > | >     install: 'ConfigurationOfAida'.
| | > | > | > ((Smalltalk at: #ConfigurationOfAida) project version:
| | '6.0')
| | > | load.
| | > | > | >
| | > | > | > Adding GemStone to the configuration is on my todo list.
| | > | > | >
| | > | > | > Dale
| | > | > | > _______________________________________________
| | > | > | > Aida mailing list
| | > | > | > [hidden email]
| | > | > | > http://lists.aidaweb.si/mailman/listinfo/aida
| | > | > | >
| | > | > |
| | > | > | --
| | > | > | Janko Mivšek
| | > | > | AIDA/Web
| | > | > | Smalltalk Web Application Server
| | > | > | http://www.aidaweb.si
| | > | >
| | > |
| | > | --
| | > | Janko Mivšek
| | > | AIDA/Web
| | > | Smalltalk Web Application Server
| | > | http://www.aidaweb.si
| | > | _______________________________________________
| | > | Aida mailing list
| | > | [hidden email]
| | > | http://lists.aidaweb.si/mailman/listinfo/aida
| | > _______________________________________________
| | > Aida mailing list
| | > [hidden email]
| | > http://lists.aidaweb.si/mailman/listinfo/aida
| |
| | --
| | Janko Mivšek
| | AIDA/Web
| | Smalltalk Web Application Server
| | http://www.aidaweb.si
| _______________________________________________
| Aida mailing list
| [hidden email]
| http://lists.aidaweb.si/mailman/listinfo/aida
_______________________________________________
Aida mailing list
[hidden email]
http://lists.aidaweb.si/mailman/listinfo/aida

Dale,

Ok, after few failed attempts I finally managed to publish as

        Aida6.0-final.1000.mcz

with a comment: Wierd Monticello/Gofer naming conventions, will you be
now satisfied???

And I must admit I lost quite some confidence of that branch of version
control systems, specially comparing to VW Store. And be sure, here is
the problem most people have with Squeak/Pharo etc and will have in the
future too, if you guys don't seat down and rethink some basics.

Just some well intended thoughts...

Best regards
Janko

On 29. 01. 2010 20:21, Dale Henrichs wrote:
--
Janko Mivšek
AIDA/Web
Smalltalk Web Application Server
http://www.aidaweb.si
_______________________________________________
Aida mailing list
[hidden email]
http://lists.aidaweb.si/mailman/listinfo/aida

Janko,

I've updated ConfigurationOfAida and the correct package is loaded. Thanks for your effort.

The problem with Store is that it is proprietary and cannot be used by other Smalltalk dialects ... if it had been opened up in the beginning, we probably wouldn't be having this conversation:)

As it is, we make do with the systems that are available to us. Gofer and Metacello are attempts at filling in gaps that Monticello was not designed to address.

Dale

----- "Janko Mivšek" <[hidden email]> wrote:

| Dale,
|
| Ok, after few failed attempts I finally managed to publish as
|
| Aida6.0-final.1000.mcz
|
| with a comment: Wierd Monticello/Gofer naming conventions, will you
| be
| now satisfied???
|
| And I must admit I lost quite some confidence of that branch of
| version
| control systems, specially comparing to VW Store. And be sure, here
| is
| the problem most people have with Squeak/Pharo etc and will have in
| the
| future too, if you guys don't seat down and rethink some basics.
|
| Just some well intended thoughts...
|
| Best regards
| Janko
|
| On 29. 01. 2010 20:21, Dale Henrichs wrote:
| > Janko,
| >
| > We're talking quick and dirty here. Aida6.0-jm.mcz will solve the
| first problem, in that the correct package will be loaded by
| Metacello. However, "load latest package" is a feature of Metacello
| and if you use Aida6.0-jm.mcz as the name, Aida6.0-beta.2.mcz will
| sort as the "latest version" of the Aida6.0 package:( Which is my I
| suggested using Aida6.0-jm.3.mcz.
| >
| > Here is the detail on the parsing/sorting of mcz file names used by
| Gofer (which is used by Metacello):
| >
| >   The parsing the mcz file name using the following basic template:
| >
| >     packagename-author.branch.versionnumber
| >
| >   packagename field can contain many $- characters (and many $.
| characters as
| >   well), with the _last_ $- delimiting the author. If the last
| character is the
| >   string is not a digit, the packagename is parsed differently (too
| wierd to
| >   describe:).
| >
| >   The author field  is delimited by the _first_ $. following the
| last $-.
| >
| >   The branch field consists of the characters between the $.
| following the
| >   author filed and the last $.. The branch field may be empty.
| >
| >   The versionnumber field consists of the characters from the _last_
| $. the end of  
| >   the string. If the versionnumber field is empty, the versionnumber
| defaults to 0.
| >
| >   Sort order for a list of mcz files is based first on packagename,
| then branch,
| >   then versionnumber and then author.
| >
| >   The latest version of a package is last package in the sorted list
| of packages in
| >   a repository (i.e, files whose packagenames are equal).
| >
| > Whether you agree with the naming convention or not, it is _is_
| necessary to have a consistent convention when talking about sort
| order and for better or worse, the above algorithms are used across
| all platforms that Gofer/Metacello are supported on.
| >
| > With that said, I think it is possible for you to name your package
| names that make sense and fit within the above naming convention, I
| had to monkey with a bunch of different name combos, but these names
| appear to be the "best of the lot":
| >
| > 'Aida6.0-0.mcz'
| > 'Aida6.0-beta.2-0.mcz'
| > 'Aida6.0-beta.1-0.mcz'
| > 'Aida5.6-0.mcz'
| > 'Aida5.4.1-0.mcz'
| > 'Aida5.4-0.mcz'
| > 'Aida5.4-beta1-0.mcz'
| >
| > Not that the trick to "fooling" the parser is th e'-0.mcz' sequence.
| The last character preceding the .mcz must be a digit and the $- must
| be precede the digit. You could use '-jm0.mcz'or other sequences in
| the 'author' field.
| >
| > The advantage is the latest package feature won't cause older
| versions of the software to be loaded.
| >
| > If you renamed all of the packages in the Aida repository things
| would be fine, however, the existing packages are still problematic
| and would still result in the wrong packages being loaded by "latest
| package." An alternate repository could be used as well.
| >
| > I understand your distaste with the ugly names, but version control
| systems have never been known for having human understandable names..
| >
| > Dale
| >
| > ----- "Janko Mivšek" <[hidden email]> wrote:
| >
| > | On 28. 01. 2010 20:34, Dale Henrichs wrote:
| > |
| > | > Actually the package name should be something like:
| > | >
| > | >   Aida6.0-jm.3.mcz
| > | >
| > | > That way the package version will sort greater than
| > | 'Aida6.0-beta.2.mcz':)
| > |
| > | Hmm, this is even worse for released versions, which shall be as
| > | accessible as possible also by the simplicity of the name.
| > |
| > | What if we rather add the command to load the 'exact' versions
| with
| > | exact filename? In both Metacello anf Gopher, to skip automatisms
| in
| > | such cases. Something like:
| > |
| > | Gofer new
| > |    squeaksource: 'Aida';
| > |    packageExact: 'Aida6.0.mcz';
| > |    load.
| > | Or not with a new method but just #package: 'Aida6.0.mcz'. If you
| > | have
| > | therefore a .mcz extension, the package is exactly specified,
| while
| > | without .mcz you use the utomatism.
| > |
| > | In any cade in the future this naming problem need to be
| rethought,
| > | because it is obviously a cause of many confusions and not
| > | appropriate
| > | for released stuff. While for development versions it is ok, I
| agree
| > | too.
| > |
| > | Best regards
| > | Janko
| > |
| > |
| > | >
| > | > Dale
| > | > ----- "Dale Henrichs" <[hidden email]> wrote:
| > | >
| > | > | Janko,
| > | > |
| > | > | Yes that should do the trick. As soon as you've posted the
| new
| > | mcz
| > | > | file, I'll update the configuration.
| > | > |
| > | > | Dale
| > | > | ----- "Janko Mivšek" <[hidden email]> wrote:
| > | > |
| > | > | | Dale,
| > | > | |
| > | > | | so the quick and dirty solution would be to repost Aida as
| > | > | | Aida6.0-jm.mcz? Let me try, I need to repost anyway,
| because
| > | the
| > | > | | connection to the parent version was somehow lost...
| > | > | |
| > | > | | Best regards
| > | > | | Janko
| > | > | |
| > | > | |
| > | > | | On 28. 01. 2010 07:02, Dale Henrichs wrote:
| > | > | | > Janko,
| > | > | | >
| > | > | | > This isn't a waste of time:)
| > | > | | >
| > | > | | > I agree that there are problems with the Monticello
| naming
| > | > | | convention. It _isn't_ enforced by anything but the tools
| and
| > | the
| > | > | sort
| > | > | | order has changed over time even within the same tool
| set...
| > | > | | >
| > | > | | > Part of the rationale behind Metacello is to break the
| link
| > | > | between
| > | > | | the mcz file names and versions. In Metacello, there is an
| > | explicit
| > | > | | version number class. I've taken the version numbers from
| your
| > | > | | packages, converted them to metacello version numbers and
| sorted
| > | the
| > | > | | list:
| > | > | | >
| > | > | | > (#('5.4.1' '5.4' '5.6' '5.4-beta1' '6.0' '6.0-beta.1'
| > | > | '6.0-beta.2')
| > | > | |
| > | > | | >   collect: [:str | str asMetacelloVersionNumber ]) sort:
| [:a
| > | :b |
| > | > | a
| > | > | | <= b ]
| > | > | | >
| > | > | | >   5.4-beta1
| > | > | | >   5.4
| > | > | | >   5.4.1
| > | > | | >   5.6
| > | > | | >   6.0-beta.1
| > | > | | >   6.0-beta.2
| > | > | | >   6.0
| > | > | | >
| > | > | | > Which I believe is the correct sort order for the
| versions. As
| > | an
| > | > | | aside numbers are sorted in numeric order so 9.0 sorts
| before
| > | 10.0.
| > | > | | Strings are sorted using string order so -alpha sorts
| before
| > | -beta.
| > | > | | Finally 6.0-anything sorts before 6.0.
| > | > | | >
| > | > | | > Gofer is used to load packages in Metacello and the Gofer
| > | sort
| > | > | order
| > | > | | based on the package names is:
| > | > | | >
| > | > | | >   '5.4.1' '5.4' '5.6' '5.4-beta1' '6.0' '6.0-beta.1'
| > | '6.0-beta.2'
| > | > | | >
| > | > | | > to top it off the name 'Aida6.0' looks like a package
| name
| > | and
| > | > | Gofer
| > | > | | endeavors to load the latest version of the package which
| is
| > | > | | 6.0-beta.2:(
| > | > | | >
| > | > | | > I'm not sure that I can force Gofer to do things
| differently
| > | > | without
| > | > | | breaking a lot of other behavior. I will look a little bit
| more
| > | > | | tomorrow, but I think that we are caught between a rock and
| a
| > | hard
| > | > | | place.
| > | > | | >
| > | > | | > If you could create packages with mcz file names that
| conform
| > | to
| > | > | the
| > | > | | packagename-author.pkgversionnumber, then the Metacello
| > | > | configuration
| > | > | | could be used to represent each of the unique product
| versions
| > | and
| > | > | the
| > | > | | package name would be less important a development artifact
| > | rather
| > | > | | than a version artifact. I see from the package history the
| you
| > | use
| > | > | | the "standard" naming convention for development so
| Metacello
| > | could
| > | > | | point into your development repository...
| > | > | | >
| > | > | | > As I said, I will look a little bit more tomorrow, but
| I'm
| > | not
| > | > | | optimistic ... the computer has to have some algorithm for
| > | > | determining
| > | > | | sort order and I think that the name Aida6.0 is too
| ambiguous
| > | (is it
| > | > | a
| > | > | | base package name or an mcz file name)....
| > | > | | >
| > | > | | > Hopefully we'll be able to figure something out.
| > | > | | >
| > | > | | > Dale
| > | > | | >
| > | > | | > ----- "Janko Mivšek" <[hidden email]> wrote:
| > | > | | >
| > | > | | > | Hi Dale,
| > | > | | > |
| > | > | | > | On 27. 01. 2010 19:12, Dale Henrichs wrote:
| > | > | | > |
| > | > | | > | > I will check into to see what might be going on ... I
| > | think
| > | > | that
| > | > | | the
| > | > | | > | naming convention that you use for Aida, Swazoo and
| Sport
| > | > | doesn't
| > | > | | > | quite fit the "standard" naming conventions which is
| > | confusing
| > | > | | > | SqueakSource and Gofer (which is used as the loader for
| > | > | | Metacello).
| > | > | | > |
| > | > | | > | Thanks a lot and sorry for wasting your time due to
| naming
| > | > | | convention.
| > | > | | > |
| > | > | | > | Which is not enforced and I always regarded as
| unwritten
| > | rule,
| > | > | but
| > | > | | not
| > | > | | > | very handy specially for official released versions,
| where
| > | > | author
| > | > | | > | certainly is not so important to be in the name. The MC
| > | package
| > | > | | > | contains
| > | > | | > | the author among its properties anway.
| > | > | | > |
| > | > | | > | Also, from usability and newbies standpoint the
| released
| > | package
| > | > | | name
| > | > | | > | should be short and a new long-name author convention
| is
| > | direct
| > | > | | > | oposite
| > | > | | > | of that rule.  While for development ones it is good to
| have
| > | all
| > | > | | three
| > | > | | > | attributes in the name.
| > | > | | > |
| > | > | | > | Just few thoughts
| > | > | | > | Janko
| > | > | | > |
| > | > | | > |
| > | > | | > | >
| > | > | | > | > Dale
| > | > | | > | > ----- "Janko Mivšek" <[hidden email]> wrote:
| > | > | | > | >
| > | > | | > | > | Hi Dale,
| > | > | | > | > |
| > | > | | > | > | Thanks a lot for that configuration! I just tried it
| and
| > | it
| > | > | | loads
| > | > | | > | > | nicelly in Pharo RC2, just that it loads
| Aida6.0-beta.2
| > | > | | instead of
| > | > | | > | > | final
| > | > | | > | > | Aida6.0. From looking at class ConfigurationOfAida,
| > | method
| > | > | | > | > | #version60:,
| > | > | | > | > | I didn's see any problem. Is that Metacello wrongly
| > | > | determine
| > | > | | the
| > | > | | > | > | version to load?
| > | > | | > | > |
| > | > | | > | > | Ok, I just "released" and "blessed" Aida6.0.mcz on
| > | > | | SqueakSource,
| > | > | | > | > | maybe
| > | > | | > | > | this will help... no, it still loads 6.0-beta.2
| > | instead.
| > | > | | > | > | http://www.squeaksource.com/Aida.html
| > | > | | > | > |
| > | > | | > | > | Funny thing is also that on SqueakSource 6.0 is now
| > | among
| > | > | | Released
| > | > | | > | > | and
| > | > | | > | > | Blessed, but not among Versions, neither among
| Latest.
| > | What
| > | > | | is
| > | > | | > | wrong?
| > | > | | > | > | I'm not very proficient with MC as you see :)
| > | > | | > | > |
| > | > | | > | > | Best regards
| > | > | | > | > | Janko
| > | > | | > | > |
| > | > | | > | > |
| > | > | | > | > | On 27. 01. 2010 04:20, Dale Henrichs wrote:
| > | > | | > | > | >
| > | > | | > | > | > ----- "Geert Claes" <[hidden email]>
| wrote:
| > | > | | > | > | >
| > | > | | > | > | > | Is there a metacello (ConfigurationOfAIDAWeb
| type
| > | of
| > | > | | thing)
| > | > | | > | > | install
| > | > | | > | > | > | for AIDA
| > | > | | > | > | > | Web?
| > | > | | > | > | > | --
| > | > | | > | > | > | View this message in context:
| > | > | | > | > | > |
| http://n4.nabble.com/Metacello-tp997116p997116.html
| > | > | | > | > | > | Sent from the AIDA/Web mailing list archive at
| > | > | | Nabble.com.
| > | > | | > | > | > | _______________________________________________
| > | > | | > | > | > | Aida mailing list
| > | > | | > | > | > | [hidden email]
| > | > | | > | > | > | http://lists.aidaweb.si/mailman/listinfo/aida
| > | > | | > | > | >
| > | > | | > | > | > Geert,
| > | > | | > | > | >
| > | > | | > | > | > I've just committed ConfigurationOfAida to the
| > | > | | > | MetacelloRepository.
| > | > | | > | > | >
| > | > | | > | > | > "on Pharo"
| > | > | | > | > | > Gofer new
| > | > | | > | > | > squeaksource: 'MetacelloRepository';
| > | > | | > | > | > package: 'ConfigurationOfAida';
| > | > | | > | > | > load.
| > | > | | > | > | >
| > | > | | > | > | > ((Smalltalk at: #ConfigurationOfAida) project
| > | version:
| > | > | | '6.0')
| > | > | | > | load.
| > | > | | > | > | >
| > | > | | > | > | > "on Squeak trunk"
| > | > | | > | > | > (Installer ss project: 'MetacelloRepository')
| > | > | | > | > | >     install: 'ConfigurationOfAida'.
| > | > | | > | > | > ((Smalltalk at: #ConfigurationOfAida) project
| > | version:
| > | > | | '6.0')
| > | > | | > | load.
| > | > | | > | > | >
| > | > | | > | > | > Adding GemStone to the configuration is on my
| todo
| > | list.
| > | > | | > | > | >
| > | > | | > | > | > Dale
| > | > | | > | > | > _______________________________________________
| > | > | | > | > | > Aida mailing list
| > | > | | > | > | > [hidden email]
| > | > | | > | > | > http://lists.aidaweb.si/mailman/listinfo/aida
| > | > | | > | > | >
| > | > | | > | > |
| > | > | | > | > | --
| > | > | | > | > | Janko Mivšek
| > | > | | > | > | AIDA/Web
| > | > | | > | > | Smalltalk Web Application Server
| > | > | | > | > | http://www.aidaweb.si
| > | > | | > | >
| > | > | | > |
| > | > | | > | --
| > | > | | > | Janko Mivšek
| > | > | | > | AIDA/Web
| > | > | | > | Smalltalk Web Application Server
| > | > | | > | http://www.aidaweb.si
| > | > | | > | _______________________________________________
| > | > | | > | Aida mailing list
| > | > | | > | [hidden email]
| > | > | | > | http://lists.aidaweb.si/mailman/listinfo/aida
| > | > | | > _______________________________________________
| > | > | | > Aida mailing list
| > | > | | > [hidden email]
| > | > | | > http://lists.aidaweb.si/mailman/listinfo/aida
| > | > | |
| > | > | | --
| > | > | | Janko Mivšek
| > | > | | AIDA/Web
| > | > | | Smalltalk Web Application Server
| > | > | | http://www.aidaweb.si
| > | > | _______________________________________________
| > | > | Aida mailing list
| > | > | [hidden email]
| > | > | http://lists.aidaweb.si/mailman/listinfo/aida
| > | >
| > |
| > | --
| > | Janko Mivšek
| > | Svetovalec za informatiko
| > | Eranova d.o.o.
| > | Ljubljana, Slovenija
| > | www.eranova.si
| > | tel:  01 514 22 55
| > | faks: 01 514 22 56
| > | gsm: 031 674 565
| >
|
| --
| Janko Mivšek
| AIDA/Web
| Smalltalk Web Application Server
| http://www.aidaweb.si
_______________________________________________
Aida mailing list
[hidden email]
http://lists.aidaweb.si/mailman/listinfo/aida