[Glass] Problem with UTF8 encoding and Monticello after GLASS update

Mariano,

It could be because I am jet-lagged and trying to read code ... but when i read the UTF8Encoding  code ( Utf8Encoding.310-dkh.31), I don't see how you are getting the ArgumentTypeError,

so I guess I need to see a full stack ... also verify that you have Utf8Encoding.310-dkh.31 loaded in both OLD and NEW ... 

Dale

On Mon, Aug 25, 2014 at 1:47 PM, Mariano Martinez Peck <[hidden email]> wrote:

Hi guys. 

I have a strange situation. I have 2 stones. Both using 3.1.0.5. One stone (let's call it the OLD), was created several months ago. The other site, let's call it the NEW, was built a week ago. 

I have just updated GLASS in both of them, using this expression:

Metacello new

  baseline: 'GLASS1';

  repository: 'github://glassdb/glass:master/repository';

  onConflict: [ :ex | ex allow ];

  onWarning: [ :ex |

        Transcript

          cr;

          show: ex description.

        ex resume ];

  load: 'default'.

After that... the NEW works ok. But the LAST, I cannot read any more packages from Monticello, since I am getting a UTF error. I debug it and I arrive to this difference:

In NEW stone (the one that works), this works: 

(UTF8Encoding newEncoder encode: 'ConfigurationOfIAM-MarianoMartinezPeck.68.mcz') asString 

In OLD stone, it yields a:   "a ArgumentTypeError occurred (error 2719), for aCollection  expected a anArray( String, ByteArray), arg is a Utf8"  when it tries to do the #asString.

The difference is that NEW #encode: answers a string. While the OLD answers a Utf8 instance. 

So....any workaround?

Thanks, 

--
Mariano
http://marianopeck.wordpress.com

_______________________________________________
Glass mailing list
[hidden email]
http://lists.gemtalksystems.com/mailman/listinfo/glass

Uffff...then I ignore for a minute the utf8 stuff (I let it without encoding) and yet another problem...

MCMczReader >> zip

zip ifNil:

[zip _ ZipArchive new.

zip readFrom: stream].

^ zip

Fails because ZipArchive doesn't exist. WTF??   Now I am comparing Core-Squeak and in the OLD stone I have lots of classes and in new stone only a few... In fact...it looks like I miss all Zip classes in the new stone. Maybe they were moved to a separate package yet that package is not included in the conf? 

On Mon, Aug 25, 2014 at 7:19 PM, Mariano Martinez Peck <[hidden email]> wrote:

On Mon, Aug 25, 2014 at 6:40 PM, Dale Henrichs <[hidden email]> wrote:

Mariano,

It could be because I am jet-lagged and trying to read code ... but when i read the UTF8Encoding  code ( Utf8Encoding.310-dkh.31), I don't see how you are getting the ArgumentTypeError,

No, the error is not there. The error is because in the OLD (broken)

(UTF8Encoding newEncoder encode: 'ConfigurationOfIAM-MarianoMartinezPeck.68.mcz')  

answers a Utf8 instance rather than a string (which is what happens in the new). 

Then,..the type error is in the #asString.

(UTF8Encoding newEncoder encode: 'ConfigurationOfIAM-MarianoMartinezPeck.68.mcz')  asString

 

Btw,...when loading from MC, the error is from #encodeForHTTPWithTextEncoding: encodingName conditionBlock: conditionBlock

where it does:

((GRCodec forEncoding: encodingName) encode: self) asString 

so I guess I need to see a full stack ... also verify that you have Utf8Encoding.310-dkh.31 loaded in both OLD and NEW ... 

How can I get a string full stack from GemTools?

Thanks 

 

Dale

On Mon, Aug 25, 2014 at 1:47 PM, Mariano Martinez Peck <[hidden email]> wrote:

Hi guys. 

I have a strange situation. I have 2 stones. Both using 3.1.0.5. One stone (let's call it the OLD), was created several months ago. The other site, let's call it the NEW, was built a week ago. 

I have just updated GLASS in both of them, using this expression:

Metacello new

  baseline: 'GLASS1';

  repository: 'github://glassdb/glass:master/repository';

  onConflict: [ :ex | ex allow ];

  onWarning: [ :ex |

        Transcript

          cr;

          show: ex description.

        ex resume ];

  load: 'default'.

After that... the NEW works ok. But the LAST, I cannot read any more packages from Monticello, since I am getting a UTF error. I debug it and I arrive to this difference:

In NEW stone (the one that works), this works: 

(UTF8Encoding newEncoder encode: 'ConfigurationOfIAM-MarianoMartinezPeck.68.mcz') asString 

In OLD stone, it yields a:   "a ArgumentTypeError occurred (error 2719), for aCollection  expected a anArray( String, ByteArray), arg is a Utf8"  when it tries to do the #asString.

The difference is that NEW #encode: answers a string. While the OLD answers a Utf8 instance. 

So....any workaround?

Thanks, 

--
Mariano
http://marianopeck.wordpress.com

_______________________________________________
Glass mailing list
[hidden email]
http://lists.gemtalksystems.com/mailman/listinfo/glass

--
Mariano
http://marianopeck.wordpress.com

--
Mariano
http://marianopeck.wordpress.com

On Mon, Aug 25, 2014 at 3:19 PM, Mariano Martinez Peck <[hidden email]> wrote:

On Mon, Aug 25, 2014 at 6:40 PM, Dale Henrichs <[hidden email]> wrote:

Mariano,

It could be because I am jet-lagged and trying to read code ... but when i read the UTF8Encoding  code ( Utf8Encoding.310-dkh.31), I don't see how you are getting the ArgumentTypeError,

No, the error is not there. The error is because in the OLD (broken)

(UTF8Encoding newEncoder encode: 'ConfigurationOfIAM-MarianoMartinezPeck.68.mcz')  

answers a Utf8 instance rather than a string (which is what happens in the new). 

Then,..the type error is in the #asString.

(UTF8Encoding newEncoder encode: 'ConfigurationOfIAM-MarianoMartinezPeck.68.mcz')  asString

 

Btw,...when loading from MC, the error is from #encodeForHTTPWithTextEncoding: encodingName conditionBlock: conditionBlock

where it does:

((GRCodec forEncoding: encodingName) encode: self) asString 

so I guess I need to see a full stack ... also verify that you have Utf8Encoding.310-dkh.31 loaded in both OLD and NEW ... 

How can I get a string full stack from GemTools?

There's a 'copy stack' menu item ... when the popup comes up use the 'copy' menu item to copy the text ... then paste should contain the stack... 

The correct Utf8>>asString method is in the package Core31x-dkh.4 ... double check that you've got that package loaded

Thanks 

 

Dale

On Mon, Aug 25, 2014 at 1:47 PM, Mariano Martinez Peck <[hidden email]> wrote:

Hi guys. 

I have a strange situation. I have 2 stones. Both using 3.1.0.5. One stone (let's call it the OLD), was created several months ago. The other site, let's call it the NEW, was built a week ago. 

I have just updated GLASS in both of them, using this expression:

Metacello new

  baseline: 'GLASS1';

  repository: 'github://glassdb/glass:master/repository';

  onConflict: [ :ex | ex allow ];

  onWarning: [ :ex |

        Transcript

          cr;

          show: ex description.

        ex resume ];

  load: 'default'.

After that... the NEW works ok. But the LAST, I cannot read any more packages from Monticello, since I am getting a UTF error. I debug it and I arrive to this difference:

In NEW stone (the one that works), this works: 

(UTF8Encoding newEncoder encode: 'ConfigurationOfIAM-MarianoMartinezPeck.68.mcz') asString 

In OLD stone, it yields a:   "a ArgumentTypeError occurred (error 2719), for aCollection  expected a anArray( String, ByteArray), arg is a Utf8"  when it tries to do the #asString.

The difference is that NEW #encode: answers a string. While the OLD answers a Utf8 instance. 

So....any workaround?

Thanks, 

--
Mariano
http://marianopeck.wordpress.com

_______________________________________________
Glass mailing list
[hidden email]
http://lists.gemtalksystems.com/mailman/listinfo/glass

--
Mariano
http://marianopeck.wordpress.com

On Mon, Aug 25, 2014 at 10:50 PM, Dale Henrichs <[hidden email]> wrote:

On Mon, Aug 25, 2014 at 3:19 PM, Mariano Martinez Peck <[hidden email]> wrote:

On Mon, Aug 25, 2014 at 6:40 PM, Dale Henrichs <[hidden email]> wrote:

Mariano,

It could be because I am jet-lagged and trying to read code ... but when i read the UTF8Encoding  code ( Utf8Encoding.310-dkh.31), I don't see how you are getting the ArgumentTypeError,

No, the error is not there. The error is because in the OLD (broken)

(UTF8Encoding newEncoder encode: 'ConfigurationOfIAM-MarianoMartinezPeck.68.mcz')  

answers a Utf8 instance rather than a string (which is what happens in the new). 

Then,..the type error is in the #asString.

(UTF8Encoding newEncoder encode: 'ConfigurationOfIAM-MarianoMartinezPeck.68.mcz')  asString

 

Btw,...when loading from MC, the error is from #encodeForHTTPWithTextEncoding: encodingName conditionBlock: conditionBlock

where it does:

((GRCodec forEncoding: encodingName) encode: self) asString 

so I guess I need to see a full stack ... also verify that you have Utf8Encoding.310-dkh.31 loaded in both OLD and NEW ... 

How can I get a string full stack from GemTools?

There's a 'copy stack' menu item ... when the popup comes up use the 'copy' menu item to copy the text ... then paste should contain the stack... 

The correct Utf8>>asString method is in the package Core31x-dkh.4 ... double check that you've got that package loaded

Hi Dale,

Thanks for your answer. Core31x package is loaded in NEW stone, but not in  the OLD (failing). Do you know if it should be loaded automatically?

I guess yes...but then I am not sure what it hasn't been loaded.

BTW...we agree we are talking about gemstone 3.1 right? and not seaside 3.1? because I am still using seaside 3.0.9.

Thanks

 

Thanks 

 

Dale

On Mon, Aug 25, 2014 at 1:47 PM, Mariano Martinez Peck <[hidden email]> wrote:

Hi guys. 

I have a strange situation. I have 2 stones. Both using 3.1.0.5. One stone (let's call it the OLD), was created several months ago. The other site, let's call it the NEW, was built a week ago. 

I have just updated GLASS in both of them, using this expression:

Metacello new

  baseline: 'GLASS1';

  repository: 'github://glassdb/glass:master/repository';

  onConflict: [ :ex | ex allow ];

  onWarning: [ :ex |

        Transcript

          cr;

          show: ex description.

        ex resume ];

  load: 'default'.

After that... the NEW works ok. But the LAST, I cannot read any more packages from Monticello, since I am getting a UTF error. I debug it and I arrive to this difference:

In NEW stone (the one that works), this works: 

(UTF8Encoding newEncoder encode: 'ConfigurationOfIAM-MarianoMartinezPeck.68.mcz') asString 

In OLD stone, it yields a:   "a ArgumentTypeError occurred (error 2719), for aCollection  expected a anArray( String, ByteArray), arg is a Utf8"  when it tries to do the #asString.

The difference is that NEW #encode: answers a string. While the OLD answers a Utf8 instance. 

So....any workaround?

Thanks, 

--
Mariano
http://marianopeck.wordpress.com

_______________________________________________
Glass mailing list
[hidden email]
http://lists.gemtalksystems.com/mailman/listinfo/glass

--
Mariano
http://marianopeck.wordpress.com

--
Mariano
http://marianopeck.wordpress.com

Hi Dale,

Thanks for your answer. Core31x package is loaded in NEW stone, but not in  the OLD (failing). Do you know if it should be loaded automatically?

I guess yes...but then I am not sure what it hasn't been loaded.

BTW...we agree we are talking about gemstone 3.1 right? and not seaside 3.1? because I am still using seaside 3.0.9.

what is in the curl-tmpgithubglassdbglassmasterzip.err file?

On Mon, Aug 25, 2014 at 7:11 PM, Mariano Martinez Peck <[hidden email]> wrote:

Ufff...of course...I should have get the baseline first:

Metacello new

  baseline: 'GLASS1';

  repository: 'github://glassdb/glass:master/repository';

  get.

Metacello new

  baseline: 'GLASS1';

  repository: 'github://glassdb/glass:master/repository';

  onConflict: [ :ex | ex allow ];

  onWarning: [ :ex |

        Transcript

          cr;

          show: ex description.

        ex resume ];

  load: 'default'.

But now I get this error:

...RETRY->BaselineOfGLASS1

...RETRY->BaselineOfGLASS1

gofer repository error: 'a GoferRepositoryError occurred (error 2710)'...ignoring

...FAILED->BaselineOfGLASS1

The repository is aMCGitHubRepository(github://glassdb/glass:master/repository) .... 

And a  curl-tmpgithubglassdbglassmasterzip.err is writte in /tmp

any idea?

On Mon, Aug 25, 2014 at 10:56 PM, Mariano Martinez Peck <[hidden email]> wrote:

On Mon, Aug 25, 2014 at 10:55 PM, Mariano Martinez Peck <[hidden email]> wrote:

On Mon, Aug 25, 2014 at 10:50 PM, Dale Henrichs <[hidden email]> wrote:

On Mon, Aug 25, 2014 at 3:19 PM, Mariano Martinez Peck <[hidden email]> wrote:

On Mon, Aug 25, 2014 at 6:40 PM, Dale Henrichs <[hidden email]> wrote:

Mariano,

It could be because I am jet-lagged and trying to read code ... but when i read the UTF8Encoding  code ( Utf8Encoding.310-dkh.31), I don't see how you are getting the ArgumentTypeError,

No, the error is not there. The error is because in the OLD (broken)

(UTF8Encoding newEncoder encode: 'ConfigurationOfIAM-MarianoMartinezPeck.68.mcz')  

answers a Utf8 instance rather than a string (which is what happens in the new). 

Then,..the type error is in the #asString.

(UTF8Encoding newEncoder encode: 'ConfigurationOfIAM-MarianoMartinezPeck.68.mcz')  asString

 

Btw,...when loading from MC, the error is from #encodeForHTTPWithTextEncoding: encodingName conditionBlock: conditionBlock

where it does:

((GRCodec forEncoding: encodingName) encode: self) asString 

so I guess I need to see a full stack ... also verify that you have Utf8Encoding.310-dkh.31 loaded in both OLD and NEW ... 

How can I get a string full stack from GemTools?

There's a 'copy stack' menu item ... when the popup comes up use the 'copy' menu item to copy the text ... then paste should contain the stack... 

The correct Utf8>>asString method is in the package Core31x-dkh.4 ... double check that you've got that package loaded

Hi Dale,

Thanks for your answer. Core31x package is loaded in NEW stone, but not in  the OLD (failing). Do you know if it should be loaded automatically?

as I said in my first email, this was the script I used to update GLAS:

Metacello new

  baseline: 'GLASS1';

  repository: 'github://glassdb/glass:master/repository';

  onConflict: [ :ex | ex allow ];

  onWarning: [ :ex |

        Transcript

          cr;

          show: ex description.

        ex resume ];

  load: 'default'.

 

I guess yes...but then I am not sure what it hasn't been loaded.

BTW...we agree we are talking about gemstone 3.1 right? and not seaside 3.1? because I am still using seaside 3.0.9.

Thanks

 

Thanks 

 

Dale

On Mon, Aug 25, 2014 at 1:47 PM, Mariano Martinez Peck <[hidden email]> wrote:

Hi guys. 

I have a strange situation. I have 2 stones. Both using 3.1.0.5. One stone (let's call it the OLD), was created several months ago. The other site, let's call it the NEW, was built a week ago. 

I have just updated GLASS in both of them, using this expression:

Metacello new

  baseline: 'GLASS1';

  repository: 'github://glassdb/glass:master/repository';

  onConflict: [ :ex | ex allow ];

  onWarning: [ :ex |

        Transcript

          cr;

          show: ex description.

        ex resume ];

  load: 'default'.

After that... the NEW works ok. But the LAST, I cannot read any more packages from Monticello, since I am getting a UTF error. I debug it and I arrive to this difference:

In NEW stone (the one that works), this works: 

(UTF8Encoding newEncoder encode: 'ConfigurationOfIAM-MarianoMartinezPeck.68.mcz') asString 

In OLD stone, it yields a:   "a ArgumentTypeError occurred (error 2719), for aCollection  expected a anArray( String, ByteArray), arg is a Utf8"  when it tries to do the #asString.

The difference is that NEW #encode: answers a string. While the OLD answers a Utf8 instance. 

So....any workaround?

Thanks, 

--
Mariano
http://marianopeck.wordpress.com

_______________________________________________
Glass mailing list
[hidden email]
http://lists.gemtalksystems.com/mailman/listinfo/glass

--
Mariano
http://marianopeck.wordpress.com

--
Mariano
http://marianopeck.wordpress.com

--
Mariano
http://marianopeck.wordpress.com

--
Mariano
http://marianopeck.wordpress.com

Btw, once you've loaded GLASS1 you do not want to try to load GLASS 1.0-beta.9.3 ... 

On Mon, Aug 25, 2014 at 7:12 PM, Dale Henrichs <[hidden email]> wrote:

On Mon, Aug 25, 2014 at 6:55 PM, Mariano Martinez Peck <[hidden email]> wrote:

Hi Dale,

Thanks for your answer. Core31x package is loaded in NEW stone, but not in  the OLD (failing). Do you know if it should be loaded automatically?

I guess yes...but then I am not sure what it hasn't been loaded.

BTW...we agree we are talking about gemstone 3.1 right? and not seaside 3.1? because I am still using seaside 3.0.9.

I believe that you have run into a Metacello bug in the OLD stone... If you didn't update to GLASS 1.0-beta9.3 before loading GLASS1 (probably loaded by the new version of ZINC) then it is possible that you have hit a bug ...

I would like to understand the current state of your OLD (broken) stone ... could you send me a copy of your 'Version Report' and a screen shot of the package versions loaded from your Monticello Browser ... with that information I can recommend a course of action for getting things repaired ...

Dale

Mariano,

After a couple of hours of sleep, I have realized a couple of things that I should have mentioned earlier:

  - GLASS 1.0-beta.9.3 is a prerequisite to loading GLASS1. 

  - 1.0-beta.9.3 loads the MetacelloPreview "in a single step" and replaces the explicit (and ugly)

     multi-step MetacelloPreview load sequence

  - once the #stable version of MetacelloPreview is loaded projects may directly depend upon GLASS1 

    and it should become unnecessary to explicitly load GLASS1 as a prerequisite for loading another 

    project

The READMEs for both the Metacello[1] and GLASS1[2] projects are currently out-of-date and need to be updated.

The following script should be equivalent to upgrading to GLASS1.0-beta9.3:

"Load the Preview version of Metacello from GitHub"

  | previewLoaded |

  previewLoaded := false.

  Smalltalk

    at: #'BaselineOfMetacello'

    ifPresent: [ :ignored | 

      Smalltalk

        at: #'MetacelloProjectRegistration'

        ifPresent: [ :cls | 

          (cls registrationForClassNamed: 'BaselineOfMetacello' ifAbsent: [  ])

            ifNotNil: [ :registration | previewLoaded := registration loadedInImage ] ] ].

  previewLoaded

    ifTrue: [ 

      (Smalltalk at: #'Metacello') new

        configuration: 'MetacelloPreview';

        version: #'stable';

        repository: 'github://dalehenrich/metacello-work:configuration';

        get ]

    ifFalse: [ 

      "Bootstrap Metacello Preview, using mcz files (#'previewBootstrap' symbolic version"

      Gofer new

        gemsource: 'metacello';

        package: 'ConfigurationOfMetacello';

        load.

      GsDeployer

        deploy: [ 

          ((Smalltalk at: #'ConfigurationOfMetacello') project

            version: #'previewBootstrap') load ] ].

  GsDeployer

    deploy: [ 

      "load the Metacello Preview"

      (Smalltalk at: #'Metacello') new

        configuration: 'MetacelloPreview';

        version: #'stable';

        repository: 'github://dalehenrich/metacello-work:configuration';

        load ]

I've tested this script on a virgin 3.1.0.5 extent and it loads cleanly ... then a GLASS1 load using the following script:

GsDeployer deploy: [Metacello new

  baseline: 'GLASS1';

  repository: 'github://glassdb/glass:master/repository';

  onConflict: [ :ex | ex allow ];

  onWarning: [ :ex |

        Transcript

          cr;

          show: ex description.

        ex resume ];

  load: 'default'.]

also loads cleanly so with luck we should be out of the woods.

I'm still not quite sure what state your image is in,so if you still have trouble loading GLASS1 then I'll probably need to need the detailed package info to try to reproduce the problems myself ...

I'm sorry that I was so dense earlier ... I really shouldn't have tried to answer any emails given how jet-lagged I was ....

Dale

[1] https://github.com/dalehenrich/metacello-work#install-preview-version

[2] https://github.com/glassdb/glass#installation

On Mon, Aug 25, 2014 at 7:25 PM, Dale Henrichs <[hidden email]> wrote:

In the process of getting Seaside3.2 and Zinc working, I discovered some issues in how Metacello and GLASS1 were being loaded and after working through the issues I determined that for best results GLASS1.0-beta.9.3 needed to be loaded before attempting to load GLASS1 ...

So your load expression could result in some issues ... I am very foggy right now and don't recall all of the details, because of my jet lag ... 

If I can look at your list of packages and projects loaded into the OLD image, I should be able to recommend a course of action .... tomorrow ...

Dale

On Mon, Aug 25, 2014 at 7:16 PM, Mariano Martinez Peck <[hidden email]> wrote:

On Mon, Aug 25, 2014 at 11:14 PM, Dale Henrichs <[hidden email]> wrote:

Btw, once you've loaded GLASS1 you do not want to try to load GLASS 1.0-beta.9.3 ... 

What do you mean exactly?   I thought the following will grab the latest GLASS:

Metacello new

  baseline: 'GLASS1';

  repository: 'github://glassdb/glass:master/repository';

  get.

Metacello new

  baseline: 'GLASS1';

  repository: 'github://glassdb/glass:master/repository';

  onConflict: [ :ex | ex allow ];

  onWarning: [ :ex |

        Transcript

          cr;

          show: ex description.

        ex resume ];

  load: 'default'.

isn't it the case?

 

On Mon, Aug 25, 2014 at 7:12 PM, Dale Henrichs <[hidden email]> wrote:

On Mon, Aug 25, 2014 at 6:55 PM, Mariano Martinez Peck <[hidden email]> wrote:

Hi Dale,

Thanks for your answer. Core31x package is loaded in NEW stone, but not in  the OLD (failing). Do you know if it should be loaded automatically?

I guess yes...but then I am not sure what it hasn't been loaded.

BTW...we agree we are talking about gemstone 3.1 right? and not seaside 3.1? because I am still using seaside 3.0.9.

I believe that you have run into a Metacello bug in the OLD stone... If you didn't update to GLASS 1.0-beta9.3 before loading GLASS1 (probably loaded by the new version of ZINC) then it is possible that you have hit a bug ...

I would like to understand the current state of your OLD (broken) stone ... could you send me a copy of your 'Version Report' and a screen shot of the package versions loaded from your Monticello Browser ... with that information I can recommend a course of action for getting things repaired ...

Dale

--
Mariano
http://marianopeck.wordpress.com

On Tue, Aug 26, 2014 at 7:06 AM, Dale Henrichs <[hidden email]> wrote:

Mariano,

After a couple of hours of sleep, I have realized a couple of things that I should have mentioned earlier:

  - GLASS 1.0-beta.9.3 is a prerequisite to loading GLASS1. 

  - 1.0-beta.9.3 loads the MetacelloPreview "in a single step" and replaces the explicit (and ugly)

     multi-step MetacelloPreview load sequence

  - once the #stable version of MetacelloPreview is loaded projects may directly depend upon GLASS1 

    and it should become unnecessary to explicitly load GLASS1 as a prerequisite for loading another 

    project

The READMEs for both the Metacello[1] and GLASS1[2] projects are currently out-of-date and need to be updated.

OK, thanks Dale for taking care. I answer below. 

 

The following script should be equivalent to upgrading to GLASS1.0-beta9.3:

"Load the Preview version of Metacello from GitHub"

  | previewLoaded |

  previewLoaded := false.

  Smalltalk

    at: #'BaselineOfMetacello'

    ifPresent: [ :ignored | 

      Smalltalk

        at: #'MetacelloProjectRegistration'

        ifPresent: [ :cls | 

          (cls registrationForClassNamed: 'BaselineOfMetacello' ifAbsent: [  ])

            ifNotNil: [ :registration | previewLoaded := registration loadedInImage ] ] ].

  previewLoaded

    ifTrue: [ 

      (Smalltalk at: #'Metacello') new

        configuration: 'MetacelloPreview';

        version: #'stable';

        repository: 'github://dalehenrich/metacello-work:configuration';

        get ]

    ifFalse: [ 

      "Bootstrap Metacello Preview, using mcz files (#'previewBootstrap' symbolic version"

      Gofer new

        gemsource: 'metacello';

        package: 'ConfigurationOfMetacello';

        load.

      GsDeployer

        deploy: [ 

          ((Smalltalk at: #'ConfigurationOfMetacello') project

            version: #'previewBootstrap') load ] ].

  GsDeployer

    deploy: [ 

      "load the Metacello Preview"

      (Smalltalk at: #'Metacello') new

        configuration: 'MetacelloPreview';

        version: #'stable';

        repository: 'github://dalehenrich/metacello-work:configuration';

        load ]

I've tested this script on a virgin 3.1.0.5 extent and it loads cleanly ... then a GLASS1 load using the following script:

GsDeployer deploy: [Metacello new

  baseline: 'GLASS1';

  repository: 'github://glassdb/glass:master/repository';

  onConflict: [ :ex | ex allow ];

  onWarning: [ :ex |

        Transcript

          cr;

          show: ex description.

        ex resume ];

  load: 'default'.]

also loads cleanly so with luck we should be out of the woods.

OK, executing that I do get things updated and I can load my code. So it seems this fixes my problem. Now...from what I understand above, I am loading GLASS 1.0-beta.9.3  as a side effect of loading latest Metacello. 

For my stones, I don't want to rely on that. I want to load an explicit version of glass so that I know and control what do I load. 

So..for the moment, it's fine, the above script is OK.

But then, what I would appreciate the most, is an updated https://github.com/glassdb/glass#installation 

From what I understood from your words, using the script from that page but simply changing to GLASS 1.0-beta.9.3  would not have the same case as the lines provided above. 

In fact, maybe tomorrow I re-run the lines provided above, and I may get GLASS 1.0-beta.9.4. And I don't want that. I want a script to load GLASS 1.0-beta.9.XXXX plus + Metacello + GLASS1.  :)

 

I'm still not quite sure what state your image is in,so if you still have trouble loading GLASS1 then I'll probably need to need the detailed package info to try to reproduce the problems myself ...

I'm sorry that I was so dense earlier ... I really shouldn't have tried to answer any emails given how jet-lagged I was ....

Np and thanks for answering. I was also very frustrated because I was not able to update my code and everything I tried ended up in another problem. 

For my stones, I don't want to rely on that. I want to load an explicit version of glass so that I know and control what do I load. 

So..for the moment, it's fine, the above script is OK.

If you are building a new stone from scratch, you should upgrade to GLASS1.0-beta.9.3 before loading anything else ... that way you have the proper GLASS/Metacello foundation for loading any other packages.

I will open another thread for this. 

For my stones, I don't want to rely on that. I want to load an explicit version of glass so that I know and control what do I load. 

So..for the moment, it's fine, the above script is OK.

If you are building a new stone from scratch, you should upgrade to GLASS1.0-beta.9.3 before loading anything else ... that way you have the proper GLASS/Metacello foundation for loading any other packages.

I will open another thread for this. 

OK. Say I have a fresh extent (not the seaside one) that I grab from /opt/gemstone/product/bin/extent0.dbf  in GemStone 3.1.0.5.  Until now, I was installing GLASS1.0-beta.... with this script that you share with me:

su -m $GEMSTONE_USER -c "$GEMSTONE/bin/topaz -l -T200000 <<EOF

 

set user SystemUser pass $2

login

run

Stream installLegacyStreamImplementation.

(Globals at: #'PositionableStream_position') == #'ANSI'

   ifTrue: [

     Globals at: #'PositionableStream_position' put: #'Legacy'.

     PositionableStream compilePositionMethods ].

true

%

commit

logout

EOF

"

su -m $GEMSTONE_USER -c "$GEMSTONE/bin/topaz -l -T200000 <<EOF

login

run

System myUserProfile _createDictionary: #GLASS at: 1.

UserGlobals at: #BootstrapSymbolDictionaryName put: #GLASS.

true

%

commit

errorCount

input $GEMSTONE/upgrade/installMaster30.topaz

commit

errorCount

logout

EOF

"

As you can see...this actually executes  $GEMSTONE/upgrade/installMaster30.topaz  (which has a hardcoded configurationVersionString: '1.0-beta.9'; in my case)  and that calls another script and another..etc

So... I wonder, is this still the correct way to install GLASS in a FRESH extent?

Or something like an updated https://github.com/glassdb/glass#installation should be enough?

Or the idea is that I first do the above to get the "supposed" GLASS version for my gemstone version and then after perform an upgrade to get latest GLASS? 

But then, what I would appreciate the most, is an updated https://github.com/glassdb/glass#installation