InterpreterError 2079

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
8 messages Options
Reply | Threaded
Open this post in threaded view
|

InterpreterError 2079

NorbertHartl
Hi,

I get an error

"InterpreterError 2079: GemStone Smalltalk execution could not return from the current Activation. Home context of block to return from may no longer be active."

and I'm not sure what causes it.

It happens in my component in method

add
        | grape comp editor |
        editor := HRGrape new asComponent addValidatedForm.
        comp := self context command viewComponent.
        grape := comp call: editor.
        grape ifNil: [ ^ nil ].
        (PRPathLookup
                start: PRCurrentContext value structure
                path: '/')
                        addChild: (HRGrapePage named: grape title asLowercase)."
        HRGrape add: grape

The stack is attached to this mail.

If I put the adChild stuff into a separate method and trigger it via link after the add method has been called than it works. I would suspect any kind of collision between the continuation and the dynamic variable. But then it is just wild guessing.

Maybe you know what happens here.

Norbert

stack.txt (28K) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: InterpreterError 2079

Dale
Norbert,

The error means that an attempt was made to return from a block that was not on the stack, so there was nowhere to return to ...

I see that PRCurrentContext is involved, so I will have to try to reproduce the problem here to get to the bottom of it (unless code reading yields something obvious).

What version of Pier do you have loaded?

Dale

----- "Norbert Hartl" <[hidden email]> wrote:

| Hi,
|
| I get an error
|
| "InterpreterError 2079: GemStone Smalltalk execution could not return
| from the current Activation. Home context of block to return from may
| no longer be active."
|
| and I'm not sure what causes it.
|
| It happens in my component in method
|
| add
| | grape comp editor |
| editor := HRGrape new asComponent addValidatedForm.
| comp := self context command viewComponent.
| grape := comp call: editor.
| grape ifNil: [ ^ nil ].
| (PRPathLookup
| start: PRCurrentContext value structure
| path: '/')
| addChild: (HRGrapePage named: grape title asLowercase)."
| HRGrape add: grape
|
| The stack is attached to this mail.
|
| If I put the adChild stuff into a separate method and trigger it via
| link after the add method has been called than it works. I would
| suspect any kind of collision between the continuation and the dynamic
| variable. But then it is just wild guessing.
|
| Maybe you know what happens here.
|
| Norbert
Reply | Threaded
Open this post in threaded view
|

Re: InterpreterError 2079

NorbertHartl
Dale,

I'm still on GLASS.230-dkh.231.I did not update anything manually.

This seems to be pier 1.1.1, right?

Norbert

On 01.03.2010, at 22:30, Dale Henrichs wrote:

> Norbert,
>
> The error means that an attempt was made to return from a block that was not on the stack, so there was nowhere to return to ...
>
> I see that PRCurrentContext is involved, so I will have to try to reproduce the problem here to get to the bottom of it (unless code reading yields something obvious).
>
> What version of Pier do you have loaded?
>
> Dale
>
> ----- "Norbert Hartl" <[hidden email]> wrote:
>
> | Hi,
> |
> | I get an error
> |
> | "InterpreterError 2079: GemStone Smalltalk execution could not return
> | from the current Activation. Home context of block to return from may
> | no longer be active."
> |
> | and I'm not sure what causes it.
> |
> | It happens in my component in method
> |
> | add
> | | grape comp editor |
> | editor := HRGrape new asComponent addValidatedForm.
> | comp := self context command viewComponent.
> | grape := comp call: editor.
> | grape ifNil: [ ^ nil ].
> | (PRPathLookup
> | start: PRCurrentContext value structure
> | path: '/')
> | addChild: (HRGrapePage named: grape title asLowercase)."
> | HRGrape add: grape
> |
> | The stack is attached to this mail.
> |
> | If I put the adChild stuff into a separate method and trigger it via
> | link after the add method has been called than it works. I would
> | suspect any kind of collision between the continuation and the dynamic
> | variable. But then it is just wild guessing.
> |
> | Maybe you know what happens here.
> |
> | Norbert

Reply | Threaded
Open this post in threaded view
|

Re: InterpreterError 2079

NorbertHartl
In reply to this post by Dale
Dale,

I'm still on GLASS.230-dkh.231.I did not update anything manually.

This seems to be pier 1.1.1, right?

Norbert

On 01.03.2010, at 22:30, Dale Henrichs wrote:

> Norbert,
>
> The error means that an attempt was made to return from a block that was not on the stack, so there was nowhere to return to ...
>
> I see that PRCurrentContext is involved, so I will have to try to reproduce the problem here to get to the bottom of it (unless code reading yields something obvious).
>
> What version of Pier do you have loaded?
>
> Dale
>
> ----- "Norbert Hartl" <[hidden email]> wrote:
>
> | Hi,
> |
> | I get an error
> |
> | "InterpreterError 2079: GemStone Smalltalk execution could not return
> | from the current Activation. Home context of block to return from may
> | no longer be active."
> |
> | and I'm not sure what causes it.
> |
> | It happens in my component in method
> |
> | add
> | | grape comp editor |
> | editor := HRGrape new asComponent addValidatedForm.
> | comp := self context command viewComponent.
> | grape := comp call: editor.
> | grape ifNil: [ ^ nil ].
> | (PRPathLookup
> | start: PRCurrentContext value structure
> | path: '/')
> | addChild: (HRGrapePage named: grape title asLowercase)."
> | HRGrape add: grape
> |
> | The stack is attached to this mail.
> |
> | If I put the adChild stuff into a separate method and trigger it via
> | link after the add method has been called than it works. I would
> | suspect any kind of collision between the continuation and the dynamic
> | variable. But then it is just wild guessing.
> |
> | Maybe you know what happens here.
> |
> | Norbert

Reply | Threaded
Open this post in threaded view
|

Re: InterpreterError 2079

Dale
In reply to this post by NorbertHartl
Dale,

GLASS.230-dkh.231 is good enough for me ... for 1.1.1 I haven't seen anything odd going on, so if you've got some pier code that I can load into my own image, I should be able to figure it out ... if you want, you can sent the mcz file by private mail...

Dale

----- "Norbert Hartl" <[hidden email]> wrote:

| Dale,
|
| I'm still on GLASS.230-dkh.231.I did not update anything manually.
|
| This seems to be pier 1.1.1, right?
|
| Norbert
|
| On 01.03.2010, at 22:30, Dale Henrichs wrote:
|
| > Norbert,
| >
| > The error means that an attempt was made to return from a block that
| was not on the stack, so there was nowhere to return to ...
| >
| > I see that PRCurrentContext is involved, so I will have to try to
| reproduce the problem here to get to the bottom of it (unless code
| reading yields something obvious).
| >
| > What version of Pier do you have loaded?
| >
| > Dale
| >
| > ----- "Norbert Hartl" <[hidden email]> wrote:
| >
| > | Hi,
| > |
| > | I get an error
| > |
| > | "InterpreterError 2079: GemStone Smalltalk execution could not
| return
| > | from the current Activation. Home context of block to return from
| may
| > | no longer be active."
| > |
| > | and I'm not sure what causes it.
| > |
| > | It happens in my component in method
| > |
| > | add
| > | | grape comp editor |
| > | editor := HRGrape new asComponent addValidatedForm.
| > | comp := self context command viewComponent.
| > | grape := comp call: editor.
| > | grape ifNil: [ ^ nil ].
| > | (PRPathLookup
| > | start: PRCurrentContext value structure
| > | path: '/')
| > | addChild: (HRGrapePage named: grape title asLowercase)."
| > | HRGrape add: grape
| > |
| > | The stack is attached to this mail.
| > |
| > | If I put the adChild stuff into a separate method and trigger it
| via
| > | link after the add method has been called than it works. I would
| > | suspect any kind of collision between the continuation and the
| dynamic
| > | variable. But then it is just wild guessing.
| > |
| > | Maybe you know what happens here.
| > |
| > | Norbert
Reply | Threaded
Open this post in threaded view
|

Re: InterpreterError 2079

NorbertHartl
Dale,

here you can find it.

MCHttpRepository
        location: 'http://source.selfish.org/mc'
        user: 'small'
        password: 'talk'

Package PierTestComponent.

Norbert

On 02.03.2010, at 02:16, Dale Henrichs wrote:

> Dale,
>
> GLASS.230-dkh.231 is good enough for me ... for 1.1.1 I haven't seen anything odd going on, so if you've got some pier code that I can load into my own image, I should be able to figure it out ... if you want, you can sent the mcz file by private mail...
>
> Dale
>
> ----- "Norbert Hartl" <[hidden email]> wrote:
>
> | Dale,
> |
> | I'm still on GLASS.230-dkh.231.I did not update anything manually.
> |
> | This seems to be pier 1.1.1, right?
> |
> | Norbert
> |
> | On 01.03.2010, at 22:30, Dale Henrichs wrote:
> |
> | > Norbert,
> | >
> | > The error means that an attempt was made to return from a block that
> | was not on the stack, so there was nowhere to return to ...
> | >
> | > I see that PRCurrentContext is involved, so I will have to try to
> | reproduce the problem here to get to the bottom of it (unless code
> | reading yields something obvious).
> | >
> | > What version of Pier do you have loaded?
> | >
> | > Dale
> | >
> | > ----- "Norbert Hartl" <[hidden email]> wrote:
> | >
> | > | Hi,
> | > |
> | > | I get an error
> | > |
> | > | "InterpreterError 2079: GemStone Smalltalk execution could not
> | return
> | > | from the current Activation. Home context of block to return from
> | may
> | > | no longer be active."
> | > |
> | > | and I'm not sure what causes it.
> | > |
> | > | It happens in my component in method
> | > |
> | > | add
> | > | | grape comp editor |
> | > | editor := HRGrape new asComponent addValidatedForm.
> | > | comp := self context command viewComponent.
> | > | grape := comp call: editor.
> | > | grape ifNil: [ ^ nil ].
> | > | (PRPathLookup
> | > | start: PRCurrentContext value structure
> | > | path: '/')
> | > | addChild: (HRGrapePage named: grape title asLowercase)."
> | > | HRGrape add: grape
> | > |
> | > | The stack is attached to this mail.
> | > |
> | > | If I put the adChild stuff into a separate method and trigger it
> | via
> | > | link after the add method has been called than it works. I would
> | > | suspect any kind of collision between the continuation and the
> | dynamic
> | > | variable. But then it is just wild guessing.
> | > |
> | > | Maybe you know what happens here.
> | > |
> | > | Norbert

Reply | Threaded
Open this post in threaded view
|

Re: InterpreterError 2079

Dale
In reply to this post by NorbertHartl
Norbert,

It looks like you might have run into a bug in either the ANSI exception handling implementation in Gemstone/S version 2.3 or the continuation implementation.

The block with the non-local return is a resumeBlock inside a GemStone exceptionHandler. I will need to dig a little deeper into this problem:

  1) does it fail with later versions of Pier?
  2) does it fail in GemStone/S 2.4?
  3) is it a bug in exceptions, continuations or something else?

Dale
----- "Norbert Hartl" <[hidden email]> wrote:

| Dale,
|
| here you can find it.
|
| MCHttpRepository
| location: 'http://source.selfish.org/mc'
| user: 'small'
| password: 'talk'
|
| Package PierTestComponent.
|
| Norbert
|
| On 02.03.2010, at 02:16, Dale Henrichs wrote:
|
| > Dale,
| >
| > GLASS.230-dkh.231 is good enough for me ... for 1.1.1 I haven't seen
| anything odd going on, so if you've got some pier code that I can load
| into my own image, I should be able to figure it out ... if you want,
| you can sent the mcz file by private mail...
| >
| > Dale
| >
| > ----- "Norbert Hartl" <[hidden email]> wrote:
| >
| > | Dale,
| > |
| > | I'm still on GLASS.230-dkh.231.I did not update anything
| manually.
| > |
| > | This seems to be pier 1.1.1, right?
| > |
| > | Norbert
| > |
| > | On 01.03.2010, at 22:30, Dale Henrichs wrote:
| > |
| > | > Norbert,
| > | >
| > | > The error means that an attempt was made to return from a block
| that
| > | was not on the stack, so there was nowhere to return to ...
| > | >
| > | > I see that PRCurrentContext is involved, so I will have to try
| to
| > | reproduce the problem here to get to the bottom of it (unless
| code
| > | reading yields something obvious).
| > | >
| > | > What version of Pier do you have loaded?
| > | >
| > | > Dale
| > | >
| > | > ----- "Norbert Hartl" <[hidden email]> wrote:
| > | >
| > | > | Hi,
| > | > |
| > | > | I get an error
| > | > |
| > | > | "InterpreterError 2079: GemStone Smalltalk execution could
| not
| > | return
| > | > | from the current Activation. Home context of block to return
| from
| > | may
| > | > | no longer be active."
| > | > |
| > | > | and I'm not sure what causes it.
| > | > |
| > | > | It happens in my component in method
| > | > |
| > | > | add
| > | > | | grape comp editor |
| > | > | editor := HRGrape new asComponent addValidatedForm.
| > | > | comp := self context command viewComponent.
| > | > | grape := comp call: editor.
| > | > | grape ifNil: [ ^ nil ].
| > | > | (PRPathLookup
| > | > | start: PRCurrentContext value structure
| > | > | path: '/')
| > | > | addChild: (HRGrapePage named: grape title asLowercase)."
| > | > | HRGrape add: grape
| > | > |
| > | > | The stack is attached to this mail.
| > | > |
| > | > | If I put the adChild stuff into a separate method and trigger
| it
| > | via
| > | > | link after the add method has been called than it works. I
| would
| > | > | suspect any kind of collision between the continuation and
| the
| > | dynamic
| > | > | variable. But then it is just wild guessing.
| > | > |
| > | > | Maybe you know what happens here.
| > | > |
| > | > | Norbert
Reply | Threaded
Open this post in threaded view
|

Re: InterpreterError 2079

Dale
In reply to this post by NorbertHartl
Norbert,

Results of my tests:

  - GemStone/S 2.3, Pier 1.1.1 [fails]
  - GemStone/S 2.3, Pier 1.2.1.3 [fails]

  - GemStone/S 2.4, Pier 1.2.1.3 [works]

The stack looks identical for 2.3 and 2.4 and changes were made to both the continuation code and the exception code, so I'm still not sure which one is the root cause. I'm guessing that the root cause is in the continuation code, since the ExceptionHandler logic looks to be the same...

see Issue 59 (http://code.google.com/p/glassdb/issues/detail?id=59) for tracking.

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

| Norbert,
|
| It looks like you might have run into a bug in either the ANSI
| exception handling implementation in Gemstone/S version 2.3 or the
| continuation implementation.
|
| The block with the non-local return is a resumeBlock inside a GemStone
| exceptionHandler. I will need to dig a little deeper into this
| problem:
|
|   1) does it fail with later versions of Pier?
|   2) does it fail in GemStone/S 2.4?
|   3) is it a bug in exceptions, continuations or something else?
|
| Dale
| ----- "Norbert Hartl" <[hidden email]> wrote:
|
| | Dale,
| |
| | here you can find it.
| |
| | MCHttpRepository
| | location: 'http://source.selfish.org/mc'
| | user: 'small'
| | password: 'talk'
| |
| | Package PierTestComponent.
| |
| | Norbert
| |
| | On 02.03.2010, at 02:16, Dale Henrichs wrote:
| |
| | > Dale,
| | >
| | > GLASS.230-dkh.231 is good enough for me ... for 1.1.1 I haven't
| seen
| | anything odd going on, so if you've got some pier code that I can
| load
| | into my own image, I should be able to figure it out ... if you
| want,
| | you can sent the mcz file by private mail...
| | >
| | > Dale
| | >
| | > ----- "Norbert Hartl" <[hidden email]> wrote:
| | >
| | > | Dale,
| | > |
| | > | I'm still on GLASS.230-dkh.231.I did not update anything
| | manually.
| | > |
| | > | This seems to be pier 1.1.1, right?
| | > |
| | > | Norbert
| | > |
| | > | On 01.03.2010, at 22:30, Dale Henrichs wrote:
| | > |
| | > | > Norbert,
| | > | >
| | > | > The error means that an attempt was made to return from a
| block
| | that
| | > | was not on the stack, so there was nowhere to return to ...
| | > | >
| | > | > I see that PRCurrentContext is involved, so I will have to
| try
| | to
| | > | reproduce the problem here to get to the bottom of it (unless
| | code
| | > | reading yields something obvious).
| | > | >
| | > | > What version of Pier do you have loaded?
| | > | >
| | > | > Dale
| | > | >
| | > | > ----- "Norbert Hartl" <[hidden email]> wrote:
| | > | >
| | > | > | Hi,
| | > | > |
| | > | > | I get an error
| | > | > |
| | > | > | "InterpreterError 2079: GemStone Smalltalk execution could
| | not
| | > | return
| | > | > | from the current Activation. Home context of block to
| return
| | from
| | > | may
| | > | > | no longer be active."
| | > | > |
| | > | > | and I'm not sure what causes it.
| | > | > |
| | > | > | It happens in my component in method
| | > | > |
| | > | > | add
| | > | > | | grape comp editor |
| | > | > | editor := HRGrape new asComponent addValidatedForm.
| | > | > | comp := self context command viewComponent.
| | > | > | grape := comp call: editor.
| | > | > | grape ifNil: [ ^ nil ].
| | > | > | (PRPathLookup
| | > | > | start: PRCurrentContext value structure
| | > | > | path: '/')
| | > | > | addChild: (HRGrapePage named: grape title asLowercase)."
| | > | > | HRGrape add: grape
| | > | > |
| | > | > | The stack is attached to this mail.
| | > | > |
| | > | > | If I put the adChild stuff into a separate method and
| trigger
| | it
| | > | via
| | > | > | link after the add method has been called than it works. I
| | would
| | > | > | suspect any kind of collision between the continuation and
| | the
| | > | dynamic
| | > | > | variable. But then it is just wild guessing.
| | > | > |
| | > | > | Maybe you know what happens here.
| | > | > |
| | > | > | Norbert