Cannot even generate sources.

Hudson can, and I can't. Grrrr.
I took the latest Pharo 1.3 from here:  https://pharo-ic.lille.inria.fr/hudson/view/Pharo/job/Pharo%201.3/lastSuccessfulBuild/artifact/Pharo-1.3.zip

I then evaluated EXACTLY what it is in latest version of file LoadVMMaker.st:
Gofer new
    squeaksource: 'MetacelloRepository';
    package: 'ConfigurationOfCog';
    load.
   
((Smalltalk at: #ConfigurationOfCog) project version: #bleedingEdge) load.


And I have the emergency console. Look the attached.

Any quotes?

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

Another case of using stuff before it initialized :)


So, when it happens? During code generation or...?

On 7 July 2011 20:32, Mariano Martinez Peck <[hidden email]> wrote:
>
> FInd also the attached PharoDebbug.log. The problem of the evaluator is in MethodNode >> printPrimitiveOn:
> where it does:
>
> Smalltalk globals
>         at: #Interpreter
>         ifPresent: [ :cls | aStream nextPutAll: ' "' , ((cls classPool at: #PrimitiveTable) at: primitive + 1) , '" ' ]
>
> because   cls classPool at: #PrimitiveTable  answers nil-
> cls is Interpreter class and there is a class variable #PrimitiveTable which is in nil.
>
> The problem is how to find the error that finally come to the emergency.
>
> thanks
>
>
> On Thu, Jul 7, 2011 at 7:25 PM, Mariano Martinez Peck <[hidden email]> wrote:
>>
>> Hudson can, and I can't. Grrrr.
>> I took the latest Pharo 1.3 from here:  https://pharo-ic.lille.inria.fr/hudson/view/Pharo/job/Pharo%201.3/lastSuccessfulBuild/artifact/Pharo-1.3.zip
>>
>> I then evaluated EXACTLY what it is in latest version of file LoadVMMaker.st:
>> Gofer new
>>     squeaksource: 'MetacelloRepository';
>>     package: 'ConfigurationOfCog';
>>     load.
>>
>> ((Smalltalk at: #ConfigurationOfCog) project version: #bleedingEdge) load.
>>
>>
>> And I have the emergency console. Look the attached.
>>
>> Any quotes?
>>
>> --
>> Mariano
>> http://marianopeck.wordpress.com
>>
>
>
>
> --
> Mariano
> http://marianopeck.wordpress.com
>
>
>

--
Best regards,
Igor Stasenko AKA sig.

Hmm.. that

((Smalltalk at: #ConfigurationOfCog) project version: #bleedingEdge) load.

actually doesn't loads latest available configuration. (Should it be so?)

Here the packages it loaded:
VMMaker-oscog-IgorStasenko.95
CMakeVMMaker-IgorStasenko.121  (and latest available config - 2.4 has 120)
VMConstruction-Plugins-OSProcessPlugin-dtl.27 (and in 2.4 its
VMConstruction-Plugins-OSProcessPlugin.oscog-eem.31)
Sound-HenrikSperreJohansen.64 (and Sound-StephaneDucasse.62 in spec)

but neverthelesss

CogMacOSConfig generateWithSources

worked just well for me.
(after fixing http://code.google.com/p/pharo/issues/detail?id=4509)

Pharo1.3
Latest update: #13283

On 8 July 2011 13:10, Mariano Martinez Peck <[hidden email]> wrote:
>
> On Fri, Jul 8, 2011 at 1:31 AM, Igor Stasenko <[hidden email]> wrote:
>>
>> Another case of using stuff before it initialized :)
>>
>>
>> So, when it happens? During code generation or...?
>>
>
> yes, sorry I forget to say....it happens while doing the #generateWithSources
>
>>
>> On 7 July 2011 20:32, Mariano Martinez Peck <[hidden email]> wrote:
>> >
>> > FInd also the attached PharoDebbug.log. The problem of the evaluator is in MethodNode >> printPrimitiveOn:
>> > where it does:
>> >
>> > Smalltalk globals
>> >         at: #Interpreter
>> >         ifPresent: [ :cls | aStream nextPutAll: ' "' , ((cls classPool at: #PrimitiveTable) at: primitive + 1) , '" ' ]
>> >
>> > because   cls classPool at: #PrimitiveTable  answers nil-
>> > cls is Interpreter class and there is a class variable #PrimitiveTable which is in nil.
>> >
>> > The problem is how to find the error that finally come to the emergency.
>> >
>> > thanks
>> >
>> >
>> > On Thu, Jul 7, 2011 at 7:25 PM, Mariano Martinez Peck <[hidden email]> wrote:
>> >>
>> >> Hudson can, and I can't. Grrrr.
>> >> I took the latest Pharo 1.3 from here:  https://pharo-ic.lille.inria.fr/hudson/view/Pharo/job/Pharo%201.3/lastSuccessfulBuild/artifact/Pharo-1.3.zip
>> >>
>> >> I then evaluated EXACTLY what it is in latest version of file LoadVMMaker.st:
>> >> Gofer new
>> >>     squeaksource: 'MetacelloRepository';
>> >>     package: 'ConfigurationOfCog';
>> >>     load.
>> >>
>> >> ((Smalltalk at: #ConfigurationOfCog) project version: #bleedingEdge) load.
>> >>
>> >>
>> >> And I have the emergency console. Look the attached.
>> >>
>> >> Any quotes?
>> >>
>> >> --
>> >> Mariano
>> >> http://marianopeck.wordpress.com
>> >>
>> >
>> >
>> >
>> > --
>> > Mariano
>> > http://marianopeck.wordpress.com
>> >
>> >
>> >
>>
>>
>>
>> --
>> Best regards,
>> Igor Stasenko AKA sig.
>
>
>
> --
> Mariano
> http://marianopeck.wordpress.com
>
>
>

--

Best regards,
Igor Stasenko AKA sig.

On 11 July 2011 10:59, Mariano Martinez Peck <[hidden email]> wrote:
>
> Hi Igor. Ok...so little more information:
>
> 1) I was wrong. The problem is when executing ((Smalltalk at: #ConfigurationOfCog) project version: #bleedingEdge) load.
> I think it is something about initialization as you said. The problem is what I said about #PrimitiveTable.
> It is completelty reproducible for me and I get the problem explained in the first mail.
>

strange, because if it won't work, hudson builds would fail.
maybe its because hudson using concrete configuration, not #bleedingEdge ?

But still, i tried it and it worked well for me.

> 2) If after the error of 1) I execute again ((Smalltalk at: #ConfigurationOfCog) project version: #bleedingEdge) load.
> it works.
>
> 3) If we need the issue  http://code.google.com/p/pharo/issues/detail?id=4509 why is not raining in Hudson world? ;)
>

sorry dont understand.
The problem you met is apparently not because of transcript, since at
the moment you reported the problem
there was no changes in transcript.

> Thanks in advance.
>
>
>
> On Fri, Jul 8, 2011 at 2:25 PM, Igor Stasenko <[hidden email]> wrote:
>>
>> Hmm.. that
>> ((Smalltalk at: #ConfigurationOfCog) project version: #bleedingEdge) load.
>>
>> actually doesn't loads latest available configuration. (Should it be so?)
>>
>> Here the packages it loaded:
>> VMMaker-oscog-IgorStasenko.95
>> CMakeVMMaker-IgorStasenko.121  (and latest available config - 2.4 has 120)
>> VMConstruction-Plugins-OSProcessPlugin-dtl.27 (and in 2.4 its
>> VMConstruction-Plugins-OSProcessPlugin.oscog-eem.31)
>> Sound-HenrikSperreJohansen.64 (and Sound-StephaneDucasse.62 in spec)
>>
>> but neverthelesss
>>
>> CogMacOSConfig generateWithSources
>>
>> worked just well for me.
>> (after fixing http://code.google.com/p/pharo/issues/detail?id=4509)
>>
>> Pharo1.3
>> Latest update: #13283
>>
>>
>> On 8 July 2011 13:10, Mariano Martinez Peck <[hidden email]> wrote:
>> >
>> > On Fri, Jul 8, 2011 at 1:31 AM, Igor Stasenko <[hidden email]> wrote:
>> >>
>> >> Another case of using stuff before it initialized :)
>> >>
>> >>
>> >> So, when it happens? During code generation or...?
>> >>
>> >
>> > yes, sorry I forget to say....it happens while doing the #generateWithSources
>> >
>> >>
>> >> On 7 July 2011 20:32, Mariano Martinez Peck <[hidden email]> wrote:
>> >> >
>> >> > FInd also the attached PharoDebbug.log. The problem of the evaluator is in MethodNode >> printPrimitiveOn:
>> >> > where it does:
>> >> >
>> >> > Smalltalk globals
>> >> >         at: #Interpreter
>> >> >         ifPresent: [ :cls | aStream nextPutAll: ' "' , ((cls classPool at: #PrimitiveTable) at: primitive + 1) , '" ' ]
>> >> >
>> >> > because   cls classPool at: #PrimitiveTable  answers nil-
>> >> > cls is Interpreter class and there is a class variable #PrimitiveTable which is in nil.
>> >> >
>> >> > The problem is how to find the error that finally come to the emergency.
>> >> >
>> >> > thanks
>> >> >
>> >> >
>> >> > On Thu, Jul 7, 2011 at 7:25 PM, Mariano Martinez Peck <[hidden email]> wrote:
>> >> >>
>> >> >> Hudson can, and I can't. Grrrr.
>> >> >> I took the latest Pharo 1.3 from here:  https://pharo-ic.lille.inria.fr/hudson/view/Pharo/job/Pharo%201.3/lastSuccessfulBuild/artifact/Pharo-1.3.zip
>> >> >>
>> >> >> I then evaluated EXACTLY what it is in latest version of file LoadVMMaker.st:
>> >> >> Gofer new
>> >> >>     squeaksource: 'MetacelloRepository';
>> >> >>     package: 'ConfigurationOfCog';
>> >> >>     load.
>> >> >>
>> >> >> ((Smalltalk at: #ConfigurationOfCog) project version: #bleedingEdge) load.
>> >> >>
>> >> >>
>> >> >> And I have the emergency console. Look the attached.
>> >> >>
>> >> >> Any quotes?
>> >> >>
>> >> >> --
>> >> >> Mariano
>> >> >> http://marianopeck.wordpress.com
>> >> >>
>> >> >
>> >> >
>> >> >
>> >> > --
>> >> > Mariano
>> >> > http://marianopeck.wordpress.com
>> >> >
>> >> >
>> >> >
>> >>
>> >>
>> >>
>> >> --
>> >> Best regards,
>> >> Igor Stasenko AKA sig.
>> >
>> >
>> >
>> > --
>> > Mariano
>> > http://marianopeck.wordpress.com
>> >
>> >
>> >
>>
>>
>>
>> --
>> Best regards,
>> Igor Stasenko AKA sig.
>
>
>
> --
> Mariano
> http://marianopeck.wordpress.com
>
>
>

--

Best regards,
Igor Stasenko AKA sig.

ouch! yes, my fault... blessed should have real versions.

I'm sorry.

El 11/07/2011, a las 1:04p.m., Igor Stasenko escribió:

>
> On 11 July 2011 17:37, Mariano Martinez Peck <[hidden email]> wrote:
>>
>>
>>
>> On Mon, Jul 11, 2011 at 1:22 PM, Igor Stasenko <[hidden email]> wrote:
>>>
>>> On 11 July 2011 10:59, Mariano Martinez Peck <[hidden email]> wrote:
>>>>
>>>> Hi Igor. Ok...so little more information:
>>>>
>>>> 1) I was wrong. The problem is when executing ((Smalltalk at: #ConfigurationOfCog) project version: #bleedingEdge) load.
>>>> I think it is something about initialization as you said. The problem is what I said about #PrimitiveTable.
>>>> It is completelty reproducible for me and I get the problem explained in the first mail.
>>>>
>>> strange, because if it won't work, hudson builds would fail.
>>> maybe its because hudson using concrete configuration, not #bleedingEdge ?
>>
>>
>> No. As far as I can see, Hudson is indeed using bleesingEdge:
>>
>> https://gitorious.org/cogvm/blessed/blobs/master/codegen-scripts/LoadVMMaker.st
>>
>
> yes, because Esteban pushed it..
> We should not push with #bleedingEdge into blessed, because then you
> cannot reproduce the builds!
>
>> Anyway, I don't think it is related.
>>
>>>
>>> But still, i tried it and it worked well for me.
>>>
>>>> 2) If after the error of 1) I execute again ((Smalltalk at: #ConfigurationOfCog) project version: #bleedingEdge) load.
>>>> it works.
>>>>
>>>> 3) If we need the issue  http://code.google.com/p/pharo/issues/detail?id=4509 why is not raining in Hudson world? ;)
>>>>
>>> sorry dont understand.
>>> The problem you met is apparently not because of transcript, since at
>>> the moment you reported the problem
>>> there was no changes in transcript.
>>
>>
>> Yes. But as I said, if I evaluate the load twice, then the load works. But after the load, when generating sources (in latest pharo 1.3) I have the problem that old transcript dnu #ensureRC
>> so....I don't understand why Hudson does not fail since such issue is not yet integrated.
>>
> maybe it picks different image.
> Check the vmmaker.image which its using.
>
> --
> Best regards,
> Igor Stasenko AKA sig.

Okay, now builds failing because of transcript...

ThreadSafeTranscript(Object)>>doesNotUnderstand: #ensureCr

eek..

--

Best regards,
Igor Stasenko AKA sig.