xmlvm-users

[xmlvm-users] NSObject dispatchObject

[Panayotis K. <pan...@pa...>]

I have two questions:

1)  what is the reason of this variable

static NSObject* dispatchObject = nil;

in the NSObject class of the C backend?


2) I have a JAVA_OBJECT , say "o1"
This object, (in java) it has a specific method, let's say
public void myMethod(Object val) {..}

How do I call this method from inside C, given I have the pointer o1 of this object?

Re: [xmlvm-users] NSObject dispatchObject

[Arno P. <ar...@pu...>]

On 5/19/11 4:35 PM, Panayotis Katsaloulis wrote:
> I have two questions:
>
> 1)  what is the reason of this variable
>
> static NSObject* dispatchObject = nil;
>
> in the NSObject class of the C backend?

The C wrapper use this object to call the performSelectorOnMainThread: 
method on a real Objective-C object.

> 2) I have a JAVA_OBJECT , say "o1"
> This object, (in java) it has a specific method, let's say
> public void myMethod(Object val) {..}
>
> How do I call this method from inside C, given I have the pointer o1 of this object?

It depends if the call is made through an interface type or class type. 
I suggest that you write a mini-Java program that does this call (in 
Java) and then cross-compile it with --target=posix. Then you can look 
at the generated code.

Arno

Re: [xmlvm-users] NSObject dispatchObject

[Panayotis K. <pan...@pa...>]

First of all, thank you Arno for being for one more time helpful.


On May 20, 2011, at 4:18 AM, Arno Puder wrote:

> 
> 
> On 5/19/11 4:35 PM, Panayotis Katsaloulis wrote:
>> I have two questions:
>> 
>> 1)  what is the reason of this variable
>> 
>> static NSObject* dispatchObject = nil;
>> 
>> in the NSObject class of the C backend?
> 
> The C wrapper use this object to call the performSelectorOnMainThread: 
> method on a real Objective-C object.

The reason I ask is because I can not find it anywhere in the code, and probably it is a leftover.


>> 2) I have a JAVA_OBJECT , say "o1"
>> This object, (in java) it has a specific method, let's say
>> public void myMethod(Object val) {..}
>> 
>> How do I call this method from inside C, given I have the pointer o1 of this object?
> 
> It depends if the call is made through an interface type or class type. 
> I suggest that you write a mini-Java program that does this call (in 
> Java) and then cross-compile it with --target=posix. Then you can look 
> at the generated code.

I want to do this in the compatibility library for arbitrary objects, in order to implement a method of an interface, so I don't know in advance what this would be.
Do you believe I should use reflection then?


And a last probably silly question:
How do I convert an ObjC String to java_lang_String ?
(and for completeness and back?)

Re: [xmlvm-users] NSObject dispatchObject

[Arno P. <ar...@pu...>]

On 5/19/11 11:47 PM, Panayotis Katsaloulis wrote:
>>> 1)  what is the reason of this variable
>>>
>>> static NSObject* dispatchObject = nil;
>>>
>>> in the NSObject class of the C backend?
>>
>> The C wrapper use this object to call the performSelectorOnMainThread:
>> method on a real Objective-C object.
>
> The reason I ask is because I can not find it anywhere in the code, and probably it is a leftover.

ah, ok, you are right. It is indeed a left-over. The dispatcher object 
is created in performSelectorOnMainThread.

>>> 2) I have a JAVA_OBJECT , say "o1"
>>> This object, (in java) it has a specific method, let's say
>>> public void myMethod(Object val) {..}
>>>
>>> How do I call this method from inside C, given I have the pointer o1 of this object?
>>
>> It depends if the call is made through an interface type or class type.
>> I suggest that you write a mini-Java program that does this call (in
>> Java) and then cross-compile it with --target=posix. Then you can look
>> at the generated code.
>
> I want to do this in the compatibility library for arbitrary objects, in order to implement a method of an interface, so I don't know in advance what this would be.
> Do you believe I should use reflection then?

hmm. Do you want to be able to call arbitrary *methods* or a certain 
method on arbitrary *objects*? For the former you need reflected, for 
the latter not. Actually, take a look at [DispatcherObject run] (in 
NSObject.m) to get an idea how you can invoke a Java method from C code.

> And a last probably silly question:
> How do I convert an ObjC String to java_lang_String ?
> (and for completeness and back?)

NSString* toNSString(JAVA_OBJECT o)
JAVA_OBJECT toJavaString(NSString* str)

both defined in NSString.m

Arno

Re: [xmlvm-users] NSObject dispatchObject

[Panayotis K. <pan...@pa...>]

On May 20, 2011, at 9:56 AM, Arno Puder wrote:
> 
> hmm. Do you want to be able to call arbitrary *methods* or a certain 
> method on arbitrary *objects*? For the former you need reflected, for 
> the latter not. Actually, take a look at [DispatcherObject run] (in 
> NSObject.m) to get an idea how you can invoke a Java method from C code.


certain method, arbitrary objects.
I did, I actually found these two parts:
java_lang_reflect_Method* method = (*(JAVA_OBJECT (*)(JAVA_OBJECT, JAVA_OBJECT, JAVA_OBJECT)) targetClass->tib->vtable[XMLVM_VTABLE_IDX_java_lang_Class_getMethod___java_lang_String_java_lang_Class_1ARRAY])(targetClass, methodName, paramTypes);
and
 (*(JAVA_OBJECT (*)(JAVA_OBJECT, JAVA_OBJECT, JAVA_OBJECT)) method->tib->vtable[XMLVM_VTABLE_IDX_java_lang_reflect_Method_invoke___java_lang_Object_java_lang_Object_1ARRAY])(method, target, params);
but I am a bit confused how to decode it.
Moreover it seems that they use reflection - that's why I asked.


>> And a last probably silly question:
>> How do I convert an ObjC String to java_lang_String ?
>> (and for completeness and back?)
> 
> NSString* toNSString(JAVA_OBJECT o)
> JAVA_OBJECT toJavaString(NSString* str)
> 
> both defined in NSString.m
> 

Oh , I missed that part - I was looking under java_lang_String :)

Re: [xmlvm-users] NSObject dispatchObject

[Paul P. <bay...@gm...>]

I'm going to go off on a tangent here, not really directed at this
conversation, but somewhat relevant.

My opinion is that reflection is a necessary evil that should be avoided
whenever possible.  One of the key words in that is "necessary."  A great
use of reflection is for web services or other generated code.  And that
equivalent is appropriately used in XMLVM.  I believe it is okay in this
type of situation because all parts using reflection are generated & not
typed by hand.  E.g. a web service is generated & a client is generated, so
if an update occurs, we just regenerate.

The "evil" part, in my opinion, comes when developers are manually typing
method names & class names into Strings, thus avoiding the compiler.  This
is unfortunately not a rarity in the real world, such as in configuration
files.  My hope is that as annotations gain muster, they will help squash
that practice where they can.

I realize there are exceptions to what I just said, but I still try to avoid
it at all costs.  Since reflection is learned at a more advanced stage, it
seems it is sometimes considered advantageous & advanced to use it, but I
believe that to be a misconception.

Anyways, I am not trying to contradict anything in this email chain, but
rather just get on my soapbox about "evil" reflection & have a conversation
in general.  I'm interested if anyone has some other thoughts on this.

Paul



On Fri, May 20, 2011 at 2:10 AM, Panayotis Katsaloulis <
pan...@pa...> wrote:

>
> On May 20, 2011, at 9:56 AM, Arno Puder wrote:
>
>
> hmm. Do you want to be able to call arbitrary *methods* or a certain
> method on arbitrary *objects*? For the former you need reflected, for
> the latter not. Actually, take a look at [DispatcherObject run] (in
> NSObject.m) to get an idea how you can invoke a Java method from C code.
>
>
>
> certain method, arbitrary objects.
> I did, I actually found these two parts:
> java_lang_reflect_Method* method = (*(JAVA_OBJECT (*)(JAVA_OBJECT,
> JAVA_OBJECT, JAVA_OBJECT)) targetClass->tib->vtable[
> XMLVM_VTABLE_IDX_java_lang_Class_getMethod___java_lang_String_java_lang_Class_1ARRAY])(targetClass,
> methodName, paramTypes);
> and
>  (*(JAVA_OBJECT (*)(JAVA_OBJECT, JAVA_OBJECT, JAVA_OBJECT)) method->tib->
> vtable[
> XMLVM_VTABLE_IDX_java_lang_reflect_Method_invoke___java_lang_Object_java_lang_Object_1ARRAY])(method,
> target, params);
> but I am a bit confused how to decode it.
> Moreover it seems that they use reflection - that's why I asked.
>
>
> And a last probably silly question:
>
> How do I convert an ObjC String to java_lang_String ?
>
> (and for completeness and back?)
>
>
> NSString* toNSString(JAVA_OBJECT o)
> JAVA_OBJECT toJavaString(NSString* str)
>
> both defined in NSString.m
>
>
> Oh , I missed that part - I was looking under java_lang_String :)
>
>
>
> ------------------------------------------------------------------------------
> What Every C/C++ and Fortran developer Should Know!
> Read this article and learn how Intel has extended the reach of its
> next-generation tools to help Windows* and Linux* C/C++ and Fortran
> developers boost performance applications - including clusters.
> http://p.sf.net/sfu/intel-dev2devmay
> _______________________________________________
> xmlvm-users mailing list
> xml...@li...
> https://lists.sourceforge.net/lists/listinfo/xmlvm-users
>
>

Re: [xmlvm-users] NSObject dispatchObject

[Panayotis K. <pan...@pa...>]

On May 25, 2011, at 5:55 PM, Paul Poley wrote:

> I'm going to go off on a tangent here, not really directed at this conversation, but somewhat relevant.
> 
> My opinion is that reflection is a necessary evil that should be avoided whenever possible.  One of the key words in that is "necessary."  A great use of reflection is for web services or other generated code.  And that equivalent is appropriately used in XMLVM.  I believe it is okay in this type of situation because all parts using reflection are generated & not typed by hand.  E.g. a web service is generated & a client is generated, so if an update occurs, we just regenerate.
> 
> The "evil" part, in my opinion, comes when developers are manually typing method names & class names into Strings, thus avoiding the compiler.  This is unfortunately not a rarity in the real world, such as in configuration files.  My hope is that as annotations gain muster, they will help squash that practice where they can.
> 
> I realize there are exceptions to what I just said, but I still try to avoid it at all costs.  Since reflection is learned at a more advanced stage, it seems it is sometimes considered advantageous & advanced to use it, but I believe that to be a misconception.
> 
> Anyways, I am not trying to contradict anything in this email chain, but rather just get on my soapbox about "evil" reflection & have a conversation in general.  I'm interested if anyone has some other thoughts on this.
> 
> Paul

Exactly, this is the spirit of this change - to get rid of reflection. :)

Re: [xmlvm-users] NSObject dispatchObject

[Paul P. <bay...@gm...>]

That's good!  I'll get off my soapbox :)

On Thu, May 26, 2011 at 11:11 AM, Panayotis Katsaloulis <
pan...@pa...> wrote:

>
> On May 25, 2011, at 5:55 PM, Paul Poley wrote:
>
> > I'm going to go off on a tangent here, not really directed at this
> conversation, but somewhat relevant.
> >
> > My opinion is that reflection is a necessary evil that should be avoided
> whenever possible.  One of the key words in that is "necessary."  A great
> use of reflection is for web services or other generated code.  And that
> equivalent is appropriately used in XMLVM.  I believe it is okay in this
> type of situation because all parts using reflection are generated & not
> typed by hand.  E.g. a web service is generated & a client is generated, so
> if an update occurs, we just regenerate.
> >
> > The "evil" part, in my opinion, comes when developers are manually typing
> method names & class names into Strings, thus avoiding the compiler.  This
> is unfortunately not a rarity in the real world, such as in configuration
> files.  My hope is that as annotations gain muster, they will help squash
> that practice where they can.
> >
> > I realize there are exceptions to what I just said, but I still try to
> avoid it at all costs.  Since reflection is learned at a more advanced
> stage, it seems it is sometimes considered advantageous & advanced to use
> it, but I believe that to be a misconception.
> >
> > Anyways, I am not trying to contradict anything in this email chain, but
> rather just get on my soapbox about "evil" reflection & have a conversation
> in general.  I'm interested if anyone has some other thoughts on this.
> >
> > Paul
>
> Exactly, this is the spirit of this change - to get rid of reflection. :)
>
> ------------------------------------------------------------------------------
> vRanger cuts backup time in half-while increasing security.
> With the market-leading solution for virtual backup and recovery,
> you get blazing-fast, flexible, and affordable data protection.
> Download your free trial now.
> http://p.sf.net/sfu/quest-d2dcopy1
> _______________________________________________
> xmlvm-users mailing list
> xml...@li...
> https://lists.sourceforge.net/lists/listinfo/xmlvm-users
>