I'm using the qdbusxml2cpp tool to generate a D-Bus adaptor class for my D-Bus server. However, it has the following drawbacks:
Code is generated once, and then you're not supposed to modify it. But what if we have to make changes (see below) and then the XML changes (in a backwards compatible way, of course)?
It is assumed that the "adaptee" has the exact same functions and signatures as the D-Bus interface. In my case, that's not exactly true, e.g. some methods are named differently. Because the generated code uses QMetaObject::invokeMethod, this is only detected at runtime. And we can't sensibly modify the generated code if we might need to regenerate it in the future.
It would be much nicer, in my opinion, if qdbusxml2cpp would generate an abstract class, just a header, where all methods are pure virtual. Then I can write an implementation of that class that simply calls the right methods on the adaptee, without going through the Qt metatype system. This solves both problems:
If the XML changes, we just regenerate the header. Now the compiler will complain until we implement the new interface correctly.
We have the freedom to call whatever functions we like on the "adaptee" class, instead of maintaining the exact same signature as in the public D-Bus interface.
I couldn't find any tool or qdbusxml2cpp fork that does the above. Before I write it myself, are there any problems with the above approach that I might be overlooking, design-wise or technical? Perhaps limitations of the metatype system related to abstract classes or pure virtual functions?
Note that I need this to work not just with methods, but with properties and signals as well.
I've also considered writing an "intermediate" adaptor that wraps the "adaptee" and offers the exact interface that the D-Bus adaptor expects, but the D-Bus adaptor would still be using the metatype system and runtime checks. Surely we can do better.
As you've discovered, there's no way to do what you want with qdbusxml2cpp directly. Which means that we have a few options here, some that you've already listed out:
It would be much nicer, in my opinion, if qdbusxml2cpp would generate an abstract class, just a header, where all methods are pure virtual.
There doesn't seem to be anything terribly wrong with this, although some tools/IDEs may not play nicely with it. One downside to this is that whenever something changes, you would have to ensure that you update all parts of the C++ source as well, not just your header, e.g. if there is boilerplate that has to change whenever a new method is added.
Code is generated once, and then you're not supposed to modify it. But what if we have to make changes (see below) and then the XML changes (in a backwards compatible way, of course)?
Depending on how well the code generator works, I've sometimes found it easier to simply use the generated code as a starting point and then simply modifying it from there. Most of the chages are generally pretty simple.
One other option that you could do is to use a different library to do the DBus communications.
dbus-cxx
dbus-cpp
dbus-cplusplus
I use(and maintain) dbus-cxx; there is a tool included(dbus-cxx-xml2cpp) that generates adaptee classes that are similar to the output of qdbusxml2cpp in that the adaptee class simply calls a different class that handles the actual response. The downside is that the xml2cpp tool is not that smart, and will not always output correct code. And to use dbus-cxx in a Qt application, you need to turn off the Qt keywords. However, it does have the advantage of using templated functions, so if your signature is incorrect you will get a compile error.
Unfortunately, there isn't really a good "right" way to do this, so I'm afraid that I don't have a "Do it this way" answer.
Related
Poco supports HMACEngine with different hash functions. For example to create a HMAC-SHA1 function I can use HMACEngine<SHA1Engine> hmca_sha1("secret");
This works well with DigestEngine like SHA1Engine and MD5Engine that have a constructor without argument.
Is it possible to use HMACEngine with hash functions from OpenSSL as they are provided by Poco::Crypto::DigestEngine?
The problem is that these functions take a string parameter in their constructor that specifies which hashing algorithm to use. This means HMACEngine<Poco::Crypto:DigestEngine> hmca_xxx("secret"); wouldn't work.
Any idea how to do this?
I think it will require some coding on Poco side. HMAC is accepting any class as template argument, and hoping it is a Digest class ans using methods like digest() that may not even exist. The HMACEngine that instantiates the class as template, so you can´t passa anything to its constructor as it is. In fact I don´t think that is a good use of templates, while I´m not very familiar with using this C++ resource yet, I think one should not expect anything to be available from templated class.
Probably a new HMACEngine for the Crypto module (instead of Foundation module) would be good. But looks like an approach more like the Poco::Crypto::RSADigestEngine would be better than the current template based one. I´ve recently modified Poco::Crypto::RSADigestEngine to be based on Poco::Crypto::DigestEngine instead off old Foundation Poco::DigestEngine, and therefore for next release it will also support all hashes that OpenSSL support. Basically RSADigestEngine creates an instance of a Poco::Crypto::DigestEngine as base and uses it.
I could help on a patch proposal if you create one at github. Poco::Crypto does need some updates :)
I'm working in a tool that is supposed to generate some Java Code to accelerate part of the development based in a swing input dialog...there is no need to get any further with it so I'm going to my problem...
I need to retrieve all the attributes from a class to check whenever it is necessary to add a new one. I tried to use reflection but things started getting complicated. In order to use reflection I need to compile the class I want to get the attributes as it does not work directly from .java file, .class is required for it.
The problem is that many of the classes has a lot of dependencies! Due to some design flaws some classes are a high coupled, so if I am supposed to dynamic use a class loader to compile a class A I would have to retrieve and compile all its dependencies! And then retrieve all the possible dependencies from the class A dependency classes!
I made a test running an existing ant file to compile to whole project instead of the above approach but it takes about 9 minutes to finish! From the final user perspective waiting 9 minutes every run is not accetable!
Does any one here knows a better solution???
If you want to avoid working with reflection and bytecode, it means that you will have to parse the .java files yourself with a grammar and, well, a parser based on this grammar. It is possible (especially if you do not implement the whole grammar, because many java features might be useless in your project perimeter), but I reckon this is no easy task.
There is an Apache commons Sandbox package called ClassScan. It is capable of doing the kind of source parsing you appear to require. http://commons.apache.org/sandbox/commons-classscan/. Note that it is in the Sandbox, so not part of the Commons Proper.
Plone Dexterity supports the definition of the content-type schema either through an interface (using zope.schema for the definition) or through an XML file. What is the preferred/recommended way?
In addition: is there documentation of the XML dialect used for defining a schema (models/mytype.xml) ?
This presentation appears close but not complete.
I personally much prefer the zope.schema route; I can, if I really wanted to, vary the interface attributes dynamically with python, while the XML definition is of course static.
Also, note that to register adapters and views against an XML-defined schema, you need to pull it into python code anyway:
from plone.dexterity import api
class IMyXMLDefinedType(api.Schema):
api.model('my_xml_defined_type.xml')
The XML dialect is part of plone.supermodel package; I was not able to locate any documentation beyond the source code.
I prefer an interface over an xml model. Partly that is because I prefer Python over XML. Partly it is because you cannot do some things with the XML. For example, if you want to register a field as searchable, with collective.dexteritytextindexer, you (currently) cannot set this in the Plone interface, so you will have to use Python code and therefore an interface. But Martijn shows in his answer that you can use api.model in an interface to refer to an xml file, so maybe that would be a way around it if you really want to.
I'm going to contribute to the mess by saying there is no hard and fast answer.
With simpler content types, or early in the development of more complex ones, I'm often oriented towards the supermodel XML because of how closely it works with the dexterity TTW editor. It allows me to work with a client with very rapid feedback on what they want from their content type.
Sometimes I'll even move into file system development of some features while still having the fields defined in the FTI via supermodel.
However, with more complex content types, you're nearly certainly going to hit something you can't do via supermodel alone. At that point, I usually translate to schema — and that's typically pretty easy to do.
Ideally, if you're doing a lot of dexterity development, you should probably be able to shift pretty easily back and forth. They're just different ways of representing the same objects and attributes.
I try to set up a web based application using spring and xslt. Since i always use xslt in a pipelining style, i would like to use calabash. Is there a possibility to call calabash from Java? I read thru the documentation on http://xmlcalabash.com but there is only a description how to use it from command line. I also tired to find some javadoc on githup but wasn't successful. Obviously, there is the Main class with the main() method and i could supply the command line parameters as a string array...
I wonder if there is a better way to do it.
I looked into this recently too. I took a pragmatic approach where I call Main.run(), and pass in a string array that I generate from a (File)Properties object. It doesn't allow passing in file inputs as streams or sources however, they must reside on the file-system.
Likely there are nicer ways. You could for instance look into http://expath.org/ . There should be sources of that project. The webapp modules (formerly known as servlex?) seems to provide XMLCalabash integration.
HTH!
I am VERY new to ASP.NET. I come from a VB6 / ASP (classic) / SQL Server 2000 background. I am reading a lot about Visual Studio 2008 (have installed it and am poking around). I have read about "reflection" and would like someone to explain, as best as you can to an older developer of the technologies I've written above, what exactly Reflection is and why I would use it... I am having trouble getting my head around that. Thanks!
Reflection is how you can explore the internals of different Types, without normally having access (ie. private, protected, etc members).
It's also used to dynamically load DLL's and get access to types and methods defined in them without statically compiling them into your project.
In a nutshell: Reflection is your toolkit for peeking under the hood of a piece of code.
As to why you would use it, it's generally only used in complex situations, or code analysis. The other common use is for loading precompiled plugins into your project.
Reflection lets you programmatically load an assembly, get a list of all the types in an assembly, get a list of all the properties and methods in these types, etc.
As an example:
myobject.GetType().GetProperty("MyProperty").SetValue(myobject, "wicked!", null)
It allows the internals of an object to be reflected to the outside world (code that is using said objects).
A practical use in statically typed languages like C# (and Java) is to allow invocation of methods/members at runtime via a string (eg the name of the method - perhaps you don't know the name of the method you will use at compile time).
In the context of dynamic languages I haven't heard the term as much (as generally you don't worry about the above), other then perhaps to iterate through a list of methods/members etc...
Reflection is .Net's means to manipulate or extract information of an assembly, class or method at run time. For example, you can create a class at runtime, including it's methods. As stated by monoxide, reflection is used to dynamically load assembly as plugins, or in advance cases, it is used to create .Net compiler targeting .Net, like IronPython.
Updated: You may refer to the topic on metaprogramming and its related topics for more details.
When you build any assembly in .NET (ASP.NET, Windows Forms, Command line, class library etc), a number of meta-data "definition tables" are also created within the assembly storing information about methods, fields and types corresponding to the types, fields and methods you wrote in your code.
The classes in System.Reflection namespace in .NET allow you to enumerate and interate over these tables, providing an "object model" for you to query and access items in these tables.
One common use of Reflection is providing extensibility (plug-ins) to your application. For example, Reflection allows you to load an assembly dynamically from a file path, query its types for a specific useful type (such as an Interface your application can call) and then actually invoke a method on this external assembly.
Custom Attributes also go hand in hand with reflection. For example the NUnit unit testing framework allows you to indicate a testing class and test methods by adding [Test] {TestFixture] attributes to your own code.
However then the NUnit test runner must use Reflection to load your assembly, search for all occurrences of methods that have the test attribute and then actually call your test.
This is simplifying it a lot, however it gives you a good practical example of where Reflection is essential.
Reflection certainly is powerful, however be ware that it allows you to completely disregard the fundamental concept of access modifiers (encapsulation) in object oriented programming.
For example you can easily use it to retrieve a list of Private methods in a class and actually call them. For this reason you need to think carefully about how and where you use it to avoid bypassing encapsulation and very tightly coupling (bad) code.
Reflection is the process of inspecting the metadata of an application. In other words,When reading attributes, you’ve already looked at some of the functionality that reflection
offers. Reflection enables an application to collect information about itself and act on this in-
formation. Reflection is slower than normally executing static code. It can, however, give you
a flexibility that static code can’t provide