I am in a highly dynamic context, heavily using dynamic instantiation of components from sources. Naturally I am concerned with the overhead from having to parse those sources each and every time an object is dynamically created. When the situation allows it, I am using manual caching:
readonly property var componentCache: new Object
function create(type) {
var comp = componentCache[type]
if (comp === undefined) { // "cache miss"
comp = Qt.createComponent(type)
if (comp.status !== Component.Ready) {
console.log("Component creation failed: " + comp.errorString())
return null
} else {
componentCache[type] = comp
}
}
return comp.createObject()
}
Except that this is not always applicable, for example, using a Loader with a component which needs to specify object properties using the setSource(source, properties) function. In this scenario it is not possible to use a manually cached Component as the function only takes an url. The doc does vaguely mention "caching", but it is not exactly clear whether this cache is QML engine wide for the component from that source or more likely - just for that particular Loader.
If the active property is false at the time when this function is
called, the given source component will not be loaded but the source
and initial properties will be cached. When the loader is made active,
an instance of the source component will be created with the initial
properties set.
The question is how to deal with this issue, and is it even necessary? Maybe Qt does component from source caching by default? Caching certainly would make sense in terms of avoiding redundant source loading (from disk or worse - network), parsing and component preparation, but its effects will only be prominent in the case of excessive dynamic instantiation, and the "typical" QML dynamic object creation scenarios usually involve a one-time object, in which case the caching would be useless memory overhead. Caching also doesn't make sense in the context of the possibility that the source may change in between the instantiations.
So since I don't have the time to dig through the mess that is the private implementations behind Qt APIs, if I had to guess, I'd say that component from source caching is not likely, but as a mere guess, it may as well be wrong.
Not an answer per se, I tripped into the question of component caching yesterday – and was surprised to discover that Qt appears to cache components. At least in creating dynamic components, the log statements related to createComponent only appear once in my test app. I've searched around and haven't seen any specific info in the docs about caching. I did come across a couple of interesting methods in the QQMLEngine Class. Then I came across the release notes for Qt 5.4
The component returned by Qt.createComponent() is no longer parented
to the engine. Be sure to hold a reference, or provide a parent.
So? If a parent is provided, it is cached (?) That appears to be the case. (and for 5.5+ ?) If you want to manage it yourself, don't provide a parent and retain the reference. (?)
QQmlEngine Class
QQmlEngine::clearComponentCache() Clears the engine's
internal component cache. This function causes the property
metadata of all components previously loaded by the engine to be
destroyed. All previously loaded components and the property bindings
for all extant objects created from those components will cease to
function. This function returns the engine to a state where it
does not contain any loaded component data. This may be useful in
order to reload a smaller subset of the previous component set, or to
load a new version of a previously loaded component. Once the
component cache has been cleared, components must be loaded before any
new objects can be created.
void QQmlEngine::trimComponentCache()
Trims the engine's internal component cache.
This function causes the property metadata of any loaded components
which are not currently in use to be destroyed.
A component is considered to be in use if there are any extant
instances of the component itself, any instances of other components
that use the component, or any objects instantiated by any of those
components.
After some toying with the code, it looks like caching also takes place in the "problematic case" of Loader.setSource().
Running the example code from this question, I found out that regular component creation fails to expand a tree deeper than 10 nodes, because of Qt's hard-coded limit:
// Do not create infinite recursion in object creation
static const int maxCreationDepth = 10;
This causes component instantiation to abort if there are more than nested 10 component instantiations in a regular scenario, product incorrect tree and generating a warning.
This doesn't happen when setSource() is used, the obvious reason for this would be that the component is cached and thus no component instantiation takes place.
Related
When someone talks about hydrating an object, what does that mean?
I see a Java project called Hydrate on the web that transforms data between different representations (RDMS to OOPS to XML). Is this the general meaning of object hydration; to transform data between representations? Could it mean reconstructing an object hierarchy from a stored representation?
Hydration refers to the process of filling an object with data. An object which has not yet been hydrated has been instantiated and represents an entity that does have data, but the data has not yet been loaded into the object. This is something that is done for performance reasons.
Additionally, the term hydration is used when discussing plans for loading data from databases or other data sources. Here are some examples:
You could say that an object is partially hydrated when you have only loaded some of the fields into it, but not all of them. This can be done because those other fields are not necessary for your current operations. So there's no reason to waste bandwidth and CPU cycles loading, transferring, and setting this data when it's not going to be used.
Additionally, there are some ORM's, such as Doctrine, which do not hydrate objects when they are instantiated, but only when the data is accessed in that object. This is one method that helps to not load data which is not going to be used.
With respect to the more generic term hydrate
Hydrating an object is taking an object that exists in memory, that doesn't yet contain any domain data ("real" data), and then populating it with domain data (such as from a database, from the network, or from a file system).
From Erick Robertson's comments on this answer:
deserialization == instantiation + hydration
If you don't need to worry about blistering performance, and you aren't debugging performance optimizations that are in the internals of a data access API, then you probably don't need to deal with hydration explicitly. You would typically use deserialization instead so you can write less code. Some data access APIs don't give you this option, and in those cases you'd also have to explicitly call the hydration step yourself.
For a bit more detail on the concept of Hydration, see Erick Robertson's answer on this same question.
With respect to the Java project called hydrate
You asked about this framework specifically, so I looked into it.
As best as I can tell, I don't think this project used the word "hydrate" in a very generic sense. I see its use in the title as an approximate synonym for "serialization". As explained above, this usage isn't entirely accurate:
See: http://en.wikipedia.org/wiki/Serialization
translating data structures or object state into a format that can be stored [...] and reconstructed later in the same or another computer environment.
I can't find the reason behind their name directly on the Hydrate FAQ, but I got clues to their intention. I think they picked the name "Hydrate" because the purpose of the library is similar to the popular sound-alike Hibernate framework, but it was designed with the exact opposite workflow in mind.
Most ORMs, Hibernate included, take an in-memory object-model oriented approach, with the database taking second consideration. The Hydrate library instead takes a database-schema oriented approach, preserving your relational data structures and letting your program work on top of them more cleanly.
Metaphorically speaking, still with respect to this library's name: Hydrate is like "making something ready to use" (like re-hydrating Dried Foods). It is a metaphorical opposite of Hibernate, which is more like "putting something away for the winter" (like Animal Hibernation).
The decision to name the library Hydrate, as far as I can tell, was not concerned with the generic computer programming term "hydrate".
When using the generic computer programming term "hydrate", performance optimizations are usually the motivation (or debugging existing optimizations). Even if the library supports granular control over when and how objects are populated with data, the timing and performance don't seem to be the primary motivation for the name or the library's functionality. The library seems more concerned with enabling end-to-end mapping and schema-preservation.
While it is somewhat redundant vernacular as Merlyn mentioned, in my experience it refers only to filling/populating an object, not instantiating/creating it, so it is a useful word when you need to be precise.
This is a pretty old question, but it seems that there is still confusion over the meaning of the following terms. Hopefully, this will disambiguate.
Hydrate
When you see descriptions that say things like, "an object that is waiting for data, is waiting to be hydrated", that's confusing and misleading. Objects don't wait for things, and hydration is just the act of filling an object with data.
Using JavaScript as the example:
const obj = {}; // empty object
const data = { foo: true, bar: true, baz: true };
// Hydrate "obj" with "data"
Object.assign(obj, data);
console.log(obj.foo); // true
console.log(obj.bar); // true
console.log(obj.baz); // true
Anything that adds values to obj is "hydrating" it. I'm just using Object.assign() in this example.
Since the terms "serialize" and "deserialize" were also mentioned in other answers, here are examples to help disambiguate the meaning of those concepts from hydration:
Serialize
console.log(JSON.stringify({ foo: true, bar: true, baz: true }));
Deserialize
console.log(JSON.parse('{"foo":true,"bar":true,"baz":true}'));
In PHP, you can create a new class from its name, w/o invoke constructor, like this:
require "A.php";
$className = "A";
$class = new \ReflectionClass($className);
$instance = $class->newInstanceWithoutConstructor();
Then, you can hydrate invoking setters (or public attributes)
I have an app using React + Redux + Normalizr, I would like to know the best practices to reduce the number of renders when something changes on entities.
Right if I change just one entity inside the entities, it will re-render all the components and not only the components that need that specific entity
There are several things you can do to minimize the number of renders in your app, and make updates faster.
Let's look at them from your Redux store down to your React components.
In your Redux store, you can use immutable data structures (for example, Immutable.js). This will make all subsequent optimizations faster, as you'll be able to compare changes by only checking for previous/next state slice equality rather than recursively comparing all props.
In your containers, that is in your top-level components where you inject redux state as props, ask only for the state slices you need, and use the pure option (I assume you're using react-redux) to make sure your container will be re-rendered only if the state slices returned by your mapStateToProps functions have changed.
If you need to compute derived data, that is if you inject in your containers data computed from various state slices, use a memoized function to make sure the computation is not triggered again if the input doesn't change, and to keep object equality with the value the previous call returned. Reselect is a very good library to do that.
In your dumb components use the shouldComponentUpdate lifecycle to avoid a re-render when incoming props do not change. If you do not want to implement this manually, you can use React's PureRenderMixin to check all props for you, or, for example, the pure function from the Recompose library if you need more control. A good use case at this level is rendering a list of items. If your item component implements shouldComponentUpdate only the modified items will be re-rendered. But this shouldn't be a fix-all-problems habit : a good components separation is often preferable in that it makes flow props only to those components which will need them.
As far as Normalizr is concerned, there is nothing more specific to be done.
If in some case (it should be rare) you detect performance problems that are directly related to React's rendering cycles of components, then you should implement the shouldComponentUpdate() method in the involved components (details can be found in React's docs here).
Change-detection in shouldComponentUpdate() will be particularly easy because Redux forces you to implement immutable state:
shouldComponentUpdate(nextProps, nextState) {
return nextProps.dataObject !== this.props.dataObject;
// true when dataObject has become a new object,
// which happens if (and only if) its data has changed,
// thanks to immutability
}
I'm using DataMapper as ORM framework after many years of experience with AR. For that reason I sometimes try to find a specific DM function that mirrors some behaviour from AR. Sometimes I'm lucky, sometimes I'm not. With the #reload directive, I'm kind of in a limbo. The method exists, but somehow doesn't do what I expected it to. Basically, instead of the AR behaviour in which the instance attributes would be updated looking up to the DB, DM somehow marks every attributes from the instance as "not loaded".
Can somebody tell me if this is possible to achieve using DM?
DataMapper marks the attributes as not loaded and will load them on the next access.
This is a result from support of lazy loading groups. DM-1 will wait to see what attribute is accessed next to load only a limited set of attributes.
Per default all attributes are in the :default group, so most likely all attributes are loaded once you hit one.
In case this lazy behavior is not wanted you can do the following:
resource = YourClass.first(:some => :stuff)
# full non lazy reload (make sure you do not have
# a reference to old somewhere that causes confusion
resource = resource.model.get(resource.id)
Slogging through MVC+EF and trying to focus on doing things the right way. Right now I'm looking to add a dropdown to a form but I'd like to avoid hitting the database every time the page loads so I'd like to store the data in the app level. I figure creating an application level variable isn't the best approach. I've read about using the cache and static utility functions but surprisingly, nothing has sounded terribly definitive. (Static classes bad for unit testing, caching bad
So I have two scenarios that I'm curious about, I'm not sure if the approach would differ between the two.
1) A basic lookup, let's say the fifty states. Small, defined, will never change. Load at application startup. (Not looking for a hard coded solution but retrieval from the database.)
2) A lookup that will very rarely change and only via an admin-like screen. Let's say, cities/stores where your product is being sold. So data would be stored
in the model but would be relatively static unless someone made changes via the application. So not looking to hit the database every time I need to populate a dropdown/listbox.
Seems like basic stuff but it's basically the same as this topic that was never answered:
Is it good to use a static EF object context in an MVC application for better perf?
Any help is appreciated.
I will address you question in a few parts. First off, is it inherently bad to use static variables or caching patterns in MVC. The answer is simply no. As long as your architecture supports them it is OK. Just put your cache in the right place and design for testability as I will explain later.
The second part is what is the "right" way to have this type of persisted data stored so you don't have to make round trips to the DB to populate common UI items. For this, I don't recommend storing EF objects. I would create POCO objects (View models or similar) that you cache. So in the example of your 50 states you might have something like this:
public class State
{
public string Abbreviation { get; set; }
public string Name { get; set; }
}
Then you would do something like this to create your cached list:
List<State> states = Context.StateData.Select(s => new State { Abbreviation = s.Abbreviation, Name = s.Name}).ToList();
Finally, whatever your caching solution is, it should implement an interface so you can mock that caching method for testing.
To do this without running into circular references or using reflection, you will need at least 3 assemblies:
Your MVC application
A class library to define your POCO objects and interfaces
A class library do perform your data access and caching (this can obviously be split into 2 libraries if that makes it easier to maintain and/or test)
That way you could have something like this in your MVC code:
ICache myCache = CacheFactory.CreateCache();
List<State> states = myCache.ListStates();
// populate your view model with states
Where ICache and State are in one library and your actual implementation of ICache is in another.
This is what I do for my standard architecture: splitting POCO objects and interfacees which are data access agnostic into a separate library from data access which is the separate from my MVC app.
Look into using a Dependency Injection tool such as unity, ninject, structuremap, etc. These will allow for the application level control you are looking for by implementing a kernel which holds on to objects in a very similar way to what you seem to be describing.
I have two classes that each need an instance of each other to function. Ordinarily if an object needs another object to run, I like to pass it in the constructor. But I can't do that in this case, because one object has to be instantiated before the other, and so therefore the second object does not exist to be passed to the first object's constructor.
I can resolve this by passing the first object to the second object's constructor, then calling a setter on the first object to pass the second object to it, but that seems a little clunky, and I'm wondering if there's a better way:
backend = new Backend();
panel = new Panel(backend);
backend.setPanel();
I've never put any study into MVC; I suppose I'm dealing with a model here (the Backend), and a view or a controller (the Panel). Any insights here I can gain from MVC?
It's time to take a look at MVC. :-) When you have a model-view-controller situation, the consensus is that the model shouldn't be aware of the view-controller (MVC often plays out as M-VC), but the view is invariably aware of the model.
If the model needs to tell the view something, it does so by notifying its listeners, of which it may have multiples. Your view should be one of them.
In a circular construction scenario I'd use a factory class/factory method. I would normally make the construction logic private to the factory (using friend construct, package level protection or similar), to en sure that no-one could construct instances without using the factory.
The use of setter/constructor is really a part of the contract between the two classes and the factory, so I'd just use whichever's convenient.
As has been pointed out, you really should try to find a non-circular solution.
First of all, contrary to what others has said here, there's no inherent problem with circular references. For example, an Order object would be expected to have a reference to the Customer object of the person who placed the Order. Similarly, it would be natural for the Customer object to have a list of Orders he has placed.
In a refernce-based language (like Java or C#) there's no problem, at all. In a value-based language (like C++), you have to take care in designing them.
That said, you design of:
backend = new Backend();
panel = new Panel(backend);
backend.setPanel(panel);
It pretty much the only way to do it.
It's better to avoid circular references. I would personally try to rethink my objects.
panel = new Panel(backend);
You do this in this routine something like
Public Sub Panel(ByVal BackEnd as BackEnd)
Me.MyBackEnd = BackEnd
BackEnd.MyPanel = Me
End Sub
You don't need BackEnd.SetPanel
It is better to use Proxies. A proxy links one object to another through raising a Event. The parent hands the child a proxy. When the child needs the parent it calls a GetRef method on the proxy. The proxy then raises a event which the parent uses to return itself to the proxy which then hands it to the child.
The use of the Event/Delegate mechanism avoids any circular reference problems.
So you have (assuming that the backend is the 'parent' here)
Public Sub Panel(ByVal BackEnd as BackEnd)
Me.MyBackEnd = BackEnd.Proxy
BackEnd.MyPanel = Me
End Sub
Public Property MyBackEnd() as BackEnd
Set (ByVal Value as BackEnd)
priBackEndProxy = BackEnd.Proxy
End Set
Get
Return priBackEndProxy.GetRef
End Get
End Property
Here is a fuller discussion on the problem of circular references. Although it is focused on fixing it in Visual Basic 6.0.
Dynamic Memory Allocation
Also another solution is aggregating Panel and BackEnd into another object. This is common if both elements are UI Controls and need to behave in a coordinated manner.
Finally as far as MVC goes I recommend using a a Model View Presenter approach instead.
Basically you have your Form Implement a IPanelForm interface. It registers itself with a class called Panel which does all the UI logic. BackEnd should have events that Panel can hook into for when the model changes. Panel handles the event and updates the form through the IPanelForm interface.
User clicks a button
The form passes to Panel that the user clicked a button
Panel handles the button and retrieves the data from the backend
Panel formats the data.
Panel uses IPanelForm Interface to show the data on the Form.
I've been delaying implementing the lessons learned here, giving me plenty of time to think about the exact right way to do it. As other people said, having a clear separation where the backend objects have listeners for when their properties change is definitely the way to go. Not only will it resolve the specific issue I was asking about in this question, it is going to make a lot of other bad design smells in this code look better. There are actually a lot of different Backend classes (going by the generic class names I used in my example), each with their own corresponding Panel class. And there's even a couple of places where some things can be moved around to separate other pairs of classes into Backend/Panel pairs following the same pattern and reducing a lot of passing junk around as parameters.
The rest of this answer is going to get language specific, as I am using Java.
I've not worried a whole lot about "JavaBeans," but I have found that following basic JavaBean conventions has been very helpful for me in the past: basically, using standard getters and setters for properties. Turns out there's a JavaBean convention I was unaware of which is really going to help here: bound properties. Bound properties are properties available through standard getters and setters which fire PropertyChangeEvents when they change. [I don't know for sure, but the JavaBeans standard may specify that all properties are supposed to be "bound properties." Not relevant to me, at this point. Be aware also that "standard" getters and setters can be very non-standard through the use of BeanInfo classes to define a JavaBean's exact interface, but I never use that, either.] (The main other JavaBean convention that I choose to follow or not as appropriate in each situation is a no-argument constructor; I'm already following it in this project because each of these Backend objects has to be serializable.)
I've found this blog entry, which was very helpful in cluing me into the bound properties/PropertyChangeEvents issue and helping me construct a plan for how I'm going to rework this code.
Right now all of my backend objects inherit from a common class called Model, which provides a couple of things every backend in this system needs including serialization support. I'm going to create an additional class JavaBean as a superclass of Model which will provide the PropertyChangeEvent support that I need, inherited by every Model. I'll update the setters in each Model to fire a PropertyChangeEvent when called. I may also have JavaBean inherited by a couple of classes which aren't technically Models in the same sense as these but which could also benefit from having other classes registered as listeners for them. The JavaBean class may not fully implement the JavaBean spec; as I've said, there are several details I don't care about. But it's good enough for this project. It sounds like I could get all this by inheriting from java.awt.Component, but these aren't components in any sense that I can justify, so I don't want to do that. (I also don't know what overhead it might entail.)
Once every Model is a JavaBean, complete with PropertyChangeEvent support, I'll do a lot of code cleanup: Models that are currently keeping references to Panels will be updated and the Panels will register themselves as listeners. So much cleaner! The Model won't have to know (and shouldn't have known in the first place) what methods the Panel should call on itself when the property updates.