A month ago, I announced msgpack4d at D.announce.
After that, I fixed bugs and improved some features.

Additional list:
- real support (suggested by Lars)
- enum support (reported by Hoshi)
- direct-conversion deserializer (suggested by Michel).
- DeflateFilter (originally DeflationBuffer)

A community review settled down.
I think current status is practicable quality.
So, I would like to add this to Phobos.

Do you see any problems?


Masahiro

I think we should do a review first, and establish this practice for future proposals as well. Could you please send the code and the documentation?

Andrei

On 06/04/2010 02:19 PM, Masahiro Nakagawa wrote:
> A month ago, I announced msgpack4d at D.announce.
> After that, I fixed bugs and improved some features.
>
> Additional list:
> - real support (suggested by Lars)
> - enum support (reported by Hoshi)
> - direct-conversion deserializer (suggested by Michel).
> - DeflateFilter (originally DeflationBuffer)
>
> A community review settled down.
> I think current status is practicable quality.
> So, I would like to add this to Phobos.
>
> Do you see any problems?
>
>
> Masahiro
> _______________________________________________
> phobos mailing list
> phobos at puremagic.com
> http://lists.puremagic.com/mailman/listinfo/phobos

I resend this mail. Somehow, I didn't reply to Phobos ML.

code and documentation is here:

code: https://dl.dropbox.com/u/374829/std/msgpack.d
doc:  https://dl.dropbox.com/u/374829/std/msgpack.html
zip:  https://dl.dropbox.com/u/374829/std_msgpack.zip (archives code and
doc)

Masahiro


On Sat, 05 Jun 2010 05:24:53 +0900, Andrei Alexandrescu <andrei at erdani.com> wrote:

> I think we should do a review first, and establish this practice for future proposals as well. Could you please send the code and the documentation?
>
> Andrei
>
> On 06/04/2010 02:19 PM, Masahiro Nakagawa wrote:
>> A month ago, I announced msgpack4d at D.announce.
>> After that, I fixed bugs and improved some features.
>>
>> Additional list:
>> - real support (suggested by Lars)
>> - enum support (reported by Hoshi)
>> - direct-conversion deserializer (suggested by Michel).
>> - DeflateFilter (originally DeflationBuffer)
>>
>> A community review settled down.
>> I think current status is practicable quality.
>> So, I would like to add this to Phobos.
>>
>> Do you see any problems?
>>
>>
>> Masahiro
>> _______________________________________________
>> phobos mailing list
>> phobos at puremagic.com
>> http://lists.puremagic.com/mailman/listinfo/phobos

Hi again,


I think a solid serialization mechanism is vital for communicating data across threads effectively.

MessagePack is a solid binary offering, but it doesn't offer the upper layers needed for a complete serialization solution. As it stands, it is a good transport mechanism. I confess, however, I wished we adopted a more widely used binary protocol such as Thrift.

Anyway, I don't see why MessagePack could not be one of the available transport mechanisms in a wider serialization library that has pluggable transports.

Thoughts?


Andrei

On 6/5/10 0:57 PDT, Masahiro Nakagawa wrote:
> I resend this mail. Somehow, I didn't reply to Phobos ML.
>
> code and documentation is here:
>
> code: https://dl.dropbox.com/u/374829/std/msgpack.d
> doc: https://dl.dropbox.com/u/374829/std/msgpack.html
> zip: https://dl.dropbox.com/u/374829/std_msgpack.zip (archives code and
> doc)
>
> Masahiro
>
>
> On Sat, 05 Jun 2010 05:24:53 +0900, Andrei Alexandrescu <andrei at erdani.com> wrote:
>
>> I think we should do a review first, and establish this practice for future proposals as well. Could you please send the code and the documentation?
>>
>> Andrei
>>
>> On 06/04/2010 02:19 PM, Masahiro Nakagawa wrote:
>>> A month ago, I announced msgpack4d at D.announce.
>>> After that, I fixed bugs and improved some features.
>>>
>>> Additional list:
>>> - real support (suggested by Lars)
>>> - enum support (reported by Hoshi)
>>> - direct-conversion deserializer (suggested by Michel).
>>> - DeflateFilter (originally DeflationBuffer)
>>>
>>> A community review settled down.
>>> I think current status is practicable quality.
>>> So, I would like to add this to Phobos.
>>>
>>> Do you see any problems?
>>>
>>>
>>> Masahiro
>>> _______________________________________________
>>> phobos mailing list
>>> phobos at puremagic.com
>>> http://lists.puremagic.com/mailman/listinfo/phobos
> _______________________________________________
> phobos mailing list
> phobos at puremagic.com
> http://lists.puremagic.com/mailman/listinfo/phobos

One more thought - MessagePack sounds perfect for the inter-thread protocol because it's compact and fast. Maybe we could tune it for that usage.

Andrei

On 6/5/10 0:57 PDT, Masahiro Nakagawa wrote:
> I resend this mail. Somehow, I didn't reply to Phobos ML.
>
> code and documentation is here:
>
> code: https://dl.dropbox.com/u/374829/std/msgpack.d
> doc: https://dl.dropbox.com/u/374829/std/msgpack.html
> zip: https://dl.dropbox.com/u/374829/std_msgpack.zip (archives code and
> doc)
>
> Masahiro
>
>
> On Sat, 05 Jun 2010 05:24:53 +0900, Andrei Alexandrescu <andrei at erdani.com> wrote:
>
>> I think we should do a review first, and establish this practice for future proposals as well. Could you please send the code and the documentation?
>>
>> Andrei
>>
>> On 06/04/2010 02:19 PM, Masahiro Nakagawa wrote:
>>> A month ago, I announced msgpack4d at D.announce.
>>> After that, I fixed bugs and improved some features.
>>>
>>> Additional list:
>>> - real support (suggested by Lars)
>>> - enum support (reported by Hoshi)
>>> - direct-conversion deserializer (suggested by Michel).
>>> - DeflateFilter (originally DeflationBuffer)
>>>
>>> A community review settled down.
>>> I think current status is practicable quality.
>>> So, I would like to add this to Phobos.
>>>
>>> Do you see any problems?
>>>
>>>
>>> Masahiro
>>> _______________________________________________
>>> phobos mailing list
>>> phobos at puremagic.com
>>> http://lists.puremagic.com/mailman/listinfo/phobos
> _______________________________________________
> phobos mailing list
> phobos at puremagic.com
> http://lists.puremagic.com/mailman/listinfo/phobos

My serialization library Orange is still available: http://dsource.org/projects/orange/
I've added an example of serializing through base class references. I can make the necessary changes for inclusion in Phobos if that is of any interest.


/Jacob Carlborg

On 27 aug 2010, at 07:57, Andrei Alexandrescu wrote:

> Hi again,
> 
> 
> I think a solid serialization mechanism is vital for communicating data across threads effectively.
> 
> MessagePack is a solid binary offering, but it doesn't offer the upper layers needed for a complete serialization solution. As it stands, it is a good transport mechanism. I confess, however, I wished we adopted a more widely used binary protocol such as Thrift.
> 
> Anyway, I don't see why MessagePack could not be one of the available transport mechanisms in a wider serialization library that has pluggable transports.
> 
> Thoughts?
> 
> 
> Andrei
> 
> On 6/5/10 0:57 PDT, Masahiro Nakagawa wrote:
>> I resend this mail. Somehow, I didn't reply to Phobos ML.
>> 
>> code and documentation is here:
>> 
>> code: https://dl.dropbox.com/u/374829/std/msgpack.d
>> doc: https://dl.dropbox.com/u/374829/std/msgpack.html
>> zip: https://dl.dropbox.com/u/374829/std_msgpack.zip (archives code and
>> doc)
>> 
>> Masahiro
>> 
>> 
>> On Sat, 05 Jun 2010 05:24:53 +0900, Andrei Alexandrescu <andrei at erdani.com> wrote:
>> 
>>> I think we should do a review first, and establish this practice for future proposals as well. Could you please send the code and the documentation?
>>> 
>>> Andrei
>>> 
>>> On 06/04/2010 02:19 PM, Masahiro Nakagawa wrote:
>>>> A month ago, I announced msgpack4d at D.announce.
>>>> After that, I fixed bugs and improved some features.
>>>> 
>>>> Additional list:
>>>> - real support (suggested by Lars)
>>>> - enum support (reported by Hoshi)
>>>> - direct-conversion deserializer (suggested by Michel).
>>>> - DeflateFilter (originally DeflationBuffer)
>>>> 
>>>> A community review settled down.
>>>> I think current status is practicable quality.
>>>> So, I would like to add this to Phobos.
>>>> 
>>>> Do you see any problems?
>>>> 
>>>> 
>>>> Masahiro
>>>> _______________________________________________
>>>> phobos mailing list
>>>> phobos at puremagic.com
>>>> http://lists.puremagic.com/mailman/listinfo/phobos
>> _______________________________________________
>> phobos mailing list
>> phobos at puremagic.com
>> http://lists.puremagic.com/mailman/listinfo/phobos
> _______________________________________________
> phobos mailing list
> phobos at puremagic.com
> http://lists.puremagic.com/mailman/listinfo/phobos

Sorry. I don't know why, but I sent this post to Andrei directly :(

On Fri, 27 Aug 2010 14:57:15 +0900, Andrei Alexandrescu <andrei at erdani.com> wrote:

>
> I think a solid serialization mechanism is vital for communicating data across threads effectively.

I agree.

> MessagePack is a solid binary offering, but it doesn't offer the upper layers needed for a complete serialization solution. As it stands, it is a good transport mechanism. I confess, however, I wished we adopted a more widely used binary protocol such as Thrift.

What are the upper layers? IDL?
If so, the RPC of MessagePack supports Thrift-compatible IDL.

The serialization of MessagePack doesn't depend on IDL or schema.
This way is the usable for stand-alone serialization.
(Avro, BERT, BSON are similar mechanism).

> Anyway, I don't see why MessagePack could not be one of the available transport mechanisms in a wider serialization library that has pluggable transports.

I don't know the detail, but he(msgpack author) wanted the
high-performance RPC framework.
Unfortunately, existing frameworks falled short of his requirements.


Masahiro