On Wed, Oct 14, 2015 at 12:09 PM, Robert Withers <robert.w.withers@gmail.com> wrote:

On 10/14/2015 11:01 AM, Mariano Martinez Peck wrote:

Robert,

As far as I can remember, the problem of substitutions at
materialziation time was because... as you may have read, Fuel first
serializes "objects" and then the references. At materialization, it
first creates the objects and then, on a second step, sets the
references betwen objects. So the problem was WHERE to place the hook,
at which point in time. If we do it just after objects were created,
then the substitution code would NOT have access to any of the instVars
of that object. If we do it AFTER objects creation and after objects
references, then there is no easy way to "substitute" which doesn't
involve #become: (because the graph is already constructed) And you know
that #become: is still terrible slow because it scans full memory. That
will change in Spur.

The trick I learned from Gemstone is to use forwarding proxies,

Well, Spur will/does something similar called lazy become. Basically it lets a forwarding pointer object and then takes advantage of next GC pass or whatever in order to resolve such proxy in ��lazy manner.��

��

looking up into the FLobjectId dictionary (decoder>>objects?) when stitching the references. When you copy the proxies on substitution, it stitches normally at reference time.

Yes, that could work. The problem is if you need instVars of the object you want to substitute. Imagine you have a class called Client and instVar 'age'. And you want to substitute Client with instances of OldClient if 'age' is > 10.�� At that point in time, the reference to the instVar 10 has not yet been filled. Yet, you need to replace the object at graph construction time.��

��

As for Marea and Ghost,

Ghost proxies paper: https://hal.inria.fr/hal-01081236/document
Current Ghost repo: http://smalltalkhub.com/#!/~CAR/Ghost

Marea paper: http://www.jot.fm/issues/issue_2013_01/article2.pdf
Current repo: http://ss3.gemstone.com/ss/Marea.html

And finally, my PhD thesis:
https://tel.archives-ouvertes.fr/tel-00764991/document

Nicely done, sir. I'll check them out, thankyou.

Thanks, feel free to ask questions.��

��

In Marea I needed custom clusters for my proxies because the serializer
itself sends messages to the objects being serialized. My proxies would
bring back graphs from a secondary memory. So if I was serializing a
graph that had proxies already, I didn't want that. So I hooked my
custom cluster for proxies that send only a few special messages to the
proxy that these understand and answer rather than intercept those
messages.

As for how to extend Fuel for this, I recommend to check the code of
Marea. See categories�� 'Marea-Serializers'�� and 'Marea-Proxies'.
I have a Marea one click here:
https://www.dropbox.com/sh/xp8jzyypmz0898j/AACRdHno6V7UfhaJ1ofTPPXva?dl=0

Cheers,

thanks so much ^^
Robert

On Wed, Oct 14, 2015 at 11:40 AM, Robert Withers
<robert.w.withers@gmail.com <mailto:robert.w.withers@gmail.com>> wrote:

�� Good morning, Max. Thank you for the example. I got a little
�� confused, between migrations and substitutions. My issue with no-arg
�� blocks, I believe, is the inability to access my scope object to
�� maintain the object tables.

�� I'm attempting to write my own Materializer, Decoder and
�� Materialization. At the moment, I'm just going to walk the graph,
�� testing and do #becomes:. See how well that works when I get a test.

�� It's really good to know about that other list.

�� thanks so much ^^
�� Robert

�� On 10/14/2015 04:15 AM, Max Leske wrote:

�� BTW, there is a dedicated Fuel mailing list:
�� pharo-fuel@lists.gforge.inria.fr
�� <mailto:pharo-fuel@lists.gforge.inria.fr>
�� <mailto:pharo-fuel@lists.gforge.inria.fr
�� <mailto:pharo-fuel@lists.gforge.inria.fr>>

�� Max

�� On 14 Oct 2015, at 09:45, Max Leske <maxleske@gmail.com
�� <mailto:maxleske@gmail.com>
�� <mailto:maxleske@gmail.com <mailto:maxleske@gmail.com>>> wrote:

�� On 14 Oct 2015, at 04:39, Robert Withers
�� <robert.w.withers@gmail.com
�� <mailto:robert.w.withers@gmail.com>
�� <mailto:robert.w.withers@gmail.com
�� <mailto:robert.w.withers@gmail.com>>> wrote:

�� On 10/13/2015 09:43 PM, Mariano Martinez Peck wrote:

�� On Tue, Oct 13, 2015 at 10:33 PM, Robert Withers
�� <robert.w.withers@gmail.com
�� <mailto:robert.w.withers@gmail.com>
�� <mailto:robert.w.withers@gmail.com
�� <mailto:robert.w.withers@gmail.com>>
�� <mailto:robert.w.withers@gmail.com
�� <mailto:robert.w.withers@gmail.com>>> wrote:

�� Hi Mariano,

�� This presents me with a big challenge, then. I
�� read the docs and
�� explored the code and the only other aspect not
�� mentioned, beyond
�� instance creation (#fuelNew #fuelNew:) and
�� postMaterialization
�� (#fuelAfterMaterialization), is migrations.
�� However, migration only
�� allows for instanceVar mappings, no code blocks.

�� What do you mean that migrations only allows instVar
�� mappings , and no
�� code blocks? I mean, what do you mean by code blocks?

�� Sounds to me like this (see FuelOutStackDebuAction):

�� serializeTestFailureContext: aContext toFileNamed: aFilename
�� | serializer |

�� serializer := FLSerializer newDefault.
�� self encodeDebugInformationOn: serializer.
�� serializer addPostMaterializationAction: [ :materialization |
�� Smalltalk tools debugger
�� openOn: Processor activeProcess
�� context: materialization root
�� label: 'External stack'
�� contents: nil
�� fullView: false ].

�� serializer
�� " use the sender context, generally the current context is not
�� interesting"
�� serialize: aContext
�� toFileNamed: aFilename

�� This stores a block in the serialization which is evaluated
�� after
�� materialization. The only requirement is that it��s a clean
�� block (no
�� closure!).

�� We also support class renames. This is here:
�� http://rmod.inria.fr/web/software/Fuel/Version1.9/Documentation/Migration?_s=Pvs4DYqPRjfEy8sg&_k=X6ltJu7FDSxumm4X&_n&22

�� Which kind of migration example you have in mind
�� that would not be
�� supported? An example would help.

�� Well, my pics will demonstrate. I am interested in doing
�� more than
�� mappping ivars or a class rename. I want to do a total
�� substitution,
�� then a further substitution on the receiving, import side:

�� Vat1: anObject (Class A) ---> On wire: desc (Descriptor)
�� ---> Vat2:
�� aProxy (Class FarERef)

�� A desc is substituted for a PassByProxy object, then a
�� FarERef is
�� substituted for the desc.

�� #fuelAccept: is a serialization side method.

�� If Fuel supports substitution on serialization, I
�� don't understand
�� why no substitution support on materialization.

�� There was a reason, which I cannot remember
�� completely. Maybe Martin or
�� Max can remember.

�� It seems your focus was pickling to disk then back. My
�� focus is
�� distributed proxy graphs, which has different use cases.

�� I am definitely going to use the world-class Fuel
�� binary
�� serialization system. However, I find myself
�� needing to extend Fuel
�� to support substitution on materialization.
�� Perhaps the solution is
�� a custom decoder.

�� I have made custom clusters for example for my Ghost
�� proxies of Marea
�� system. It was a perfect example of how I could
�� extent Fuel besides the
�� common hooks. Fuel provides many places for
�� extending , like clusters,
�� analyzer,�� etc

�� Right on, exactly! Could you tell me more about your
�� Ghost proxies
�� and Marea, please? As well, could you mention how you
�� select a custom
�� cluster on the serialization side?

�� thanks so much ^^
�� Robert

�� No, a bit more. It looks like I need a new
�� FLSubstitutePointerObjectCluster, write them on
�� serialization with
�� the substitute, then do unsubstitution on
�� materialization, since the
�� cluster controls materialization and not the decoder.

�� Does this approach seem sound to you, from a you
�� know architecture
�� and design approach?

�� There was an issue. Hope other can remember. If not,
�� I will try to
�� explan what I remember tomorrow.

�� thanks so much ^^
�� Robert

�� On 10/13/2015 04:49 PM, Mariano Martinez Peck wrote:

�� No, unfortunately, as far as I can remember,
�� we do not have
�� that. There
�� are another hooks you may use but only in
�� certain scenarios
�� (#fuelNew,
�� #fuelAfterMaterialization, global sends,
�� etc). But everything is
�� listed
�� in
�� http://rmod.inria.fr/web/software/Fuel/Version1.9/Documentation/CustomizingGraph
�� so if you didn't find anything of help in
�� there there are
�� chances
�� there isn't anything.

�� Cheers,

�� On Tue, Oct 13, 2015 at 5:30 PM, Robert Withers
�� <robert.w.withers@gmail.com
�� <mailto:robert.w.withers@gmail.com>
�� <mailto:robert.w.withers@gmail.com
�� <mailto:robert.w.withers@gmail.com>>
�� <mailto:robert.w.withers@gmail.com
�� <mailto:robert.w.withers@gmail.com>>
�� <mailto:robert.w.withers@gmail.com
�� <mailto:robert.w.withers@gmail.com>
�� <mailto:robert.w.withers@gmail.com
�� <mailto:robert.w.withers@gmail.com>>>> wrote:

�� Yes, I meant dynamic substitution on
�� materialization, to
�� use the
�� correct terminology.

�� thanks,
�� Robert

�� On 10/13/2015 11:40 AM, Max Leske wrote:

�� On 13 Oct 2015, at 17:16, Robert
�� Withers
�� <robert.w.withers@gmail.com
�� <mailto:robert.w.withers@gmail.com>
�� <mailto:robert.w.withers@gmail.com
�� <mailto:robert.w.withers@gmail.com>>
�� <mailto:robert.w.withers@gmail.com
�� <mailto:robert.w.withers@gmail.com>>

�� <mailto:robert.w.withers@gmail.com
�� <mailto:robert.w.withers@gmail.com>
�� <mailto:robert.w.withers@gmail.com
�� <mailto:robert.w.withers@gmail.com>>>> wrote:

�� Every extra source helps, thank
�� you. I see how to do
�� non-stream substitutions on
�� materializations, but the
�� documentation did not indicate a
�� way to do non-stream
�� substitutions on serialization.
�� Is it possible?

�� I don��t understand what you mean by
�� non-stream��. Could
�� you give
�� an example?

�� thanks,
�� Robert

�� On 10/13/2015 09:00 AM, Mariano
�� Martinez Peck wrote:

�� Hi Robert,

�� As for the documentation,
�� you have LOTS of
�� tests, you
�� have the chapter
�� Torsten pasted, you have
�� this documentation:
�� http://rmod.inria.fr/web/software/Fuel

�� But also, as for internals,
�� there is a journal
�� paper we
�� wrote:
�� http://rmod.lille.inria.fr/archives/papers/Dias12a-SPE-Fuel.pdf

�� Let us know how it goes,

�� On Tue, Oct 13, 2015 at 6:00
�� AM, Torsten Bergmann
�� <astares@gmx.de
�� <mailto:astares@gmx.de> <mailto:astares@gmx.de
�� <mailto:astares@gmx.de>>
�� <mailto:astares@gmx.de <mailto:astares@gmx.de>>
�� <mailto:astares@gmx.de
�� <mailto:astares@gmx.de> <mailto:astares@gmx.de
�� <mailto:astares@gmx.de>>>
�� <mailto:astares@gmx.de
�� <mailto:astares@gmx.de> <mailto:astares@gmx.de
�� <mailto:astares@gmx.de>>
�� <mailto:astares@gmx.de
�� <mailto:astares@gmx.de> <mailto:astares@gmx.de
�� <mailto:astares@gmx.de>>>>> wrote:

�� Hi Robert,

�� Also checkout the
�� chapter on Fuel in Pharo
�� Enterprise book:

�� https://ci.inria.fr/pharo-contribution/view/Books/job/EnterprisePharoBook/lastSuccessfulBuild/artifact/book-result/EnterprisePharo-A4.pdf

�� Bye
�� Torsten

�� Gesendet: Dienstag, 13. Oktober 2015 um

�� 09:44 Uhr

�� Von: "Robert Withers"

�� <robert.w.withers@gmail.com
�� <mailto:robert.w.withers@gmail.com>
�� <mailto:robert.w.withers@gmail.com
�� <mailto:robert.w.withers@gmail.com>>
�� <mailto:robert.w.withers@gmail.com
�� <mailto:robert.w.withers@gmail.com>>

�� <mailto:robert.w.withers@gmail.com
�� <mailto:robert.w.withers@gmail.com>
�� <mailto:robert.w.withers@gmail.com
�� <mailto:robert.w.withers@gmail.com>>>

�� <mailto:robert.w.withers@gmail.com
�� <mailto:robert.w.withers@gmail.com>
�� <mailto:robert.w.withers@gmail.com
�� <mailto:robert.w.withers@gmail.com>>

�� <mailto:robert.w.withers@gmail.com
�� <mailto:robert.w.withers@gmail.com>
�� <mailto:robert.w.withers@gmail.com
�� <mailto:robert.w.withers@gmail.com>>>>>

�� An: pharo-dev@lists.pharo.org
�� <mailto:pharo-dev@lists.pharo.org>
�� <mailto:pharo-dev@lists.pharo.org
�� <mailto:pharo-dev@lists.pharo.org>>

�� <mailto:pharo-dev@lists.pharo.org
�� <mailto:pharo-dev@lists.pharo.org>>

�� <mailto:pharo-dev@lists.pharo.org
�� <mailto:pharo-dev@lists.pharo.org>
�� <mailto:pharo-dev@lists.pharo.org
�� <mailto:pharo-dev@lists.pharo.org>>>

�� <mailto:pharo-dev@lists.pharo.org
�� <mailto:pharo-dev@lists.pharo.org>
�� <mailto:pharo-dev@lists.pharo.org
�� <mailto:pharo-dev@lists.pharo.org>>

�� <mailto:pharo-dev@lists.pharo.org
�� <mailto:pharo-dev@lists.pharo.org>
�� <mailto:pharo-dev@lists.pharo.org
�� <mailto:pharo-dev@lists.pharo.org>>>>

�� Betreff: Re: [Pharo-dev] binary

�� serialization

�� Yes, I have to do object substitutions.

�� Thanks
�� for the link!

�� thanks,
�� Robert

�� On 10/13/2015 03:43 AM, Max Leske wrote:

�� On 13 Oct 2015, at 09:40, Robert Withers

�� <robert.w.withers@gmail.com
�� <mailto:robert.w.withers@gmail.com>
�� <mailto:robert.w.withers@gmail.com
�� <mailto:robert.w.withers@gmail.com>>
�� <mailto:robert.w.withers@gmail.com
�� <mailto:robert.w.withers@gmail.com>>

�� <mailto:robert.w.withers@gmail.com
�� <mailto:robert.w.withers@gmail.com>
�� <mailto:robert.w.withers@gmail.com
�� <mailto:robert.w.withers@gmail.com>>>

�� <mailto:robert.w.withers@gmail.com
�� <mailto:robert.w.withers@gmail.com>
�� <mailto:robert.w.withers@gmail.com
�� <mailto:robert.w.withers@gmail.com>>

�� <mailto:robert.w.withers@gmail.com
�� <mailto:robert.w.withers@gmail.com>
�� <mailto:robert.w.withers@gmail.com
�� <mailto:robert.w.withers@gmail.com>>>>> wrote:

�� Sven and Torsten, that's a binary

�� serialization library! It
�� will take time to learn
�� it and how to use
�� mappers.

�� What is the format; is it language

�� neutral?

�� For quick serialization you don��t

�� need to do
�� anything. It works
�� for (almost) all
�� objects. Only if you
�� want to
�� exclude things or
�� treat some objects in a
�� special way, you
�� will need
�� to do some stuff.

�� Documentation:

�� http://rmod.inria.fr/web/software/Fuel.

�� thanks,
�� Robert

�� On 10/13/2015 01:21 AM, Sven Van

�� Caekenberghe
�� wrote:

�� Yes, it is called FUEL and it is a

�� standard
�� part of the
�� image. See FLSerializer
�� and FLMaterializer.

�� On 13 Oct 2015, at 06:59, Robert

�� Withers

�� <robert.w.withers@gmail.com
�� <mailto:robert.w.withers@gmail.com>
�� <mailto:robert.w.withers@gmail.com
�� <mailto:robert.w.withers@gmail.com>>
�� <mailto:robert.w.withers@gmail.com
�� <mailto:robert.w.withers@gmail.com>>

�� <mailto:robert.w.withers@gmail.com
�� <mailto:robert.w.withers@gmail.com>
�� <mailto:robert.w.withers@gmail.com
�� <mailto:robert.w.withers@gmail.com>>>

�� <mailto:robert.w.withers@gmail.com
�� <mailto:robert.w.withers@gmail.com>
�� <mailto:robert.w.withers@gmail.com
�� <mailto:robert.w.withers@gmail.com>>

�� <mailto:robert.w.withers@gmail.com
�� <mailto:robert.w.withers@gmail.com>
�� <mailto:robert.w.withers@gmail.com
�� <mailto:robert.w.withers@gmail.com>>>>> wrote:

�� Does Pharo have stream classes to

�� binary
�� de/serialize an
�� object, such that the
�� protocol accepts an
�� object as
�� an argument and
�� converts it to a byteArray?

�� --
�� thanks,
�� Robert

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

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

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

�� <Exporting Vat.jpg><Importing Vat.jpg>

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

Mariano
http://marianopeck.wordpress.com