I'm not sure I understand your question, @hiradyazdan. In int key mode, MessagePack stores each field as an element in an array. If we didn't store a messagepack nil code in a slot of that array, what would we store instead? If we "compressed" it by simply omitting those elements then all the indexes would be off and the deserializer wouldn't know how to interpret each element.
the use case of this would be when there are two different model structures required for the same model, where one doesn’t need the elements that the other need (aiming to two different servers for example)
you have a deserializer using msgpack-csharp in c# env which can work fine without skipping nils and reindxing, and another one, on another server for example, that expects different model structure, which skips and reindex.
It sounds like you have two potential schemas you want to serialize to. You can do that in either of two ways: The simplest is to define each schemas in terms of a C# class. Serialize an instance of the class based on where you're sending the serialized result. So if you have one destination that expects 3 properties and one that expects 4, instantiate and initialize both classes, and serialize each one.
The other option is to write a custom formatter that knows to skip specific properties and serialize with that one when you want to omit them.
The other option is to write a custom formatter that knows to skip specific properties and serialize with that one when you want to omit them.
Or an Example Project?
@pmArchitect There is nothing specific about MessagePack and Unity. Use it as the first code block in the readme on the repository shows. The ONLY thing unity specific is, that it can serialize some of the types of Unity. You still need to separete serializeable and non-serialiezable data and assign the values deserialized after loading to whatever you use in Unity.
@neuecc With #589 I think we've made all the public API and behavioral changes we want to for 2.0. How do you feel about a stable package?
I'm going to go through the active PRs and maybe a few of the issues to see if any are best done before 2.0 stabilizes.
(and probably aggressively close some stale PRs)
@AArnott thanks for looking. About lint, maybe #556, during working for Unity, I've deleted Directory.Build.props.
Can MessagePack c # deserialize an array of objects lazy?
@neuecc Hello. Can MessagePack deserialize an array of objects lazy?
If you mean can it deserialize a
Lazy<int>[], I think not. Even if it could, the values inside wouldn't actually be lazy -- they would have to be constructed at deserialization time. Unless perhaps you were willing to keep the ReadOnlySequence<byte> that you deserialized from around indefinitely.
@AArnott I have a very large array of objects, when deserializing I can not load them into memory. I would like to use something like .deserialize <IEnumerable <myClass>>. And then the myClass object was deserialized only by accessing it in the foreach body
With some effort that could probably be done, again assuming you don't mind keeping the serialized form in memory while the enumerable isn't fully realized.
But in order for the contents of your enumerable to have been serialized in the first place, it would have had to all be in memory. So is this even a real thing? If they can't all fit in memory at once, how did you serialize it in the first place?
But in order for the contents of your enumerable to have been serialized in the first place, it would have had to all be in memory. So is this even a real thing? If they can't all fit in memory at once, how did you serialize it in the first place?
You had an array? So what, was it
But IMO you should probably think about this differently. If you have a very large object graph, consider serializing it separately, and index it, so that you can look up that object and deserialize it on-demand when calling a particular method to access it. That way one person enumerating your lazily constructed collection won't permanently allocate a ton of memory in your process.
Lazy<T>[] ? And now you want to deserialize it lazily using IEnumerable<T>? Regardless, I imagine there's a few ways to do it, all of which requires that you hold the ReadOnlySequence<byte> that you deserialize from (or at least parts of it) in memory indefinitely (or until all elements have been fully realized).But IMO you should probably think about this differently. If you have a very large object graph, consider serializing it separately, and index it, so that you can look up that object and deserialize it on-demand when calling a particular method to access it. That way one person enumerating your lazily constructed collection won't permanently allocate a ton of memory in your process.
@turik441 I recently came across #109 which appears to be similar to your feature request. I added a comment there that explains how we might accomplish it (and how I'm accomplishing it already via some custom formatters I use in my app).
@neuecc I propose we're ready to release 2.0 as a stable package. Thoughts or concerns?
However, if stable means not rc, we should require to publish mpc, unitypackage, too.
By the way, v2 isn't compatible with streaming reading and writing, right?
For writing, we use a chunk link that borrows a buffer from BufferPool,
It is not possible to write by Streaming.
(For IBufferWriter, Flush is not called, so a buffer is created internally)
As for reading, since it is necessary to convert to ReadOnlySequence once, it is not read by streaming.
In other words, A buffer of that size(if data is too large, require large buffer) is required for both writing and reading.
This is not a concern.
I just thought about an API that combines high-speed reading and writing with Streaming.
That hasn't been achieved yet.
Thanks, @neuecc . And yes, by going stable, I mean shedding the
-rc or any other suffix so that NuGet/SemVer considers the package stable, and (just as important) we consider the public API stable so that future changes do not break backward compatibility unless we're willing to ship a 3.0 release.
Regarding streaming, MessagePack v2.0's streaming support is to asynchronously drain the stream to memory buffers then deserialize all at once, or to serialize everything to memory buffers then asynchronously copy it all to streams. Of note though is this does not require one contiguous buffer. All the serializing or reading from a stream is to a
IMO though, buffering msgpack before (de)serialization to an async stream isn't nearly as bad when it is broken up into many smaller arrays as 2.0 does. An original or deserialized object graph is very likely to take up much more memory than the msgpack buffers that represent it, and those buffers often get recycled too. So until someone complains about not being able to serialize some very large object graph because of lack of streaming serialization, I wouldn't worry about it.
Sequence<T> which allocates many moderately sized arrays so that they don't go to the large object heap, and you don't run out of memory just because you can't allocate a large enough array. If #437 is ever done (set for 3.0 currently) it will allow asynchronous streaming deserialization without sacrificing performance.IMO though, buffering msgpack before (de)serialization to an async stream isn't nearly as bad when it is broken up into many smaller arrays as 2.0 does. An original or deserialized object graph is very likely to take up much more memory than the msgpack buffers that represent it, and those buffers often get recycled too. So until someone complains about not being able to serialize some very large object graph because of lack of streaming serialization, I wouldn't worry about it.
However, if stable means not rc, we should require to publish mpc, unitypackage, too.
How close are we to publishing mpc / unitypackage? Is there an issue tracking this? Is #593 tracking something like this?
@AArnott
Ah, I didn't set up Azure Pipelines (I was going to leave everything to you about CI)
For mpc, the rest of the work is from this comment
https://github.com/neuecc/MessagePack-CSharp/issues/617#issuecomment-551076028
It is also necessary to update the documentation.
Thanks, @neuecc. I've updated the labels on #617 to signify that it's a 2.0 ship blocker and assigned to myself.
What documentation needs to be updated? I've kept the README file up to date as I've made the various API breaking changes for 2.0, I think.
What documentation needs to be updated? I've kept the README file up to date as I've made the various API breaking changes for 2.0, I think.
