[Mono-devel-list] Binary JVM to MSIL conversion

Gaurav Vaish gvaish at adobe.com
Tue Aug 26 00:44:59 EDT 2003


Thanks James for a great mail.
    I would be eagerly looking forward to the extensions. If you need any
help, just drop by.

Well,
    Don't copy 'From' to your addressbook. Instead keep this one -
gvaish_mono~AT~lycos~DOT~com :-)


Cheers,
Gaurav
http://gvaish.virtualave.net
--------------------------------


----- Original Message ----- 
From: "James Michael DuPont" <mdupont777 at yahoo.com>
To: "Gaurav Vaish" <gvaish at adobe.com>; "Mono list"
<mono-devel-list at lists.ximian.com>
Sent: Monday, August 25, 2003 15:47
Subject: Re: [Mono-devel-list] Binary JVM to MSIL conversion


>
> --- Gaurav Vaish <gvaish at adobe.com> wrote:
> > > At the risk of talking about something I have not researched into
> > > detail yet, but have thought about, I will have to look in this in
> > more
> > > detail.
> > > We are talking about attributes at the run time ,based on the IL?
> >
> >     I am not sure that you mean by attributes based on the IL but
> > what I
> > meant was simply 'having metadata about classes / properties /
> > methods /
> > events etc'.
> >
> > > I was thinking about this in the context of porting the pnet
> > runtime to
> > > the gcc. My idea was to embed a good chunk of the runtime as a lib
> > > inside of the target program, or as a lib. This would also include
> > the
> > > attribute handling.
> >
> >     Do you mean .a / .so? And then if we plan to use attributes
> > features,
> > we'll need to link to the library.
>
> Yes, you will have a lib for the attribute funcitionality
> [attribute-base] and a lib containing all the attributes-data
> [attribute-instance] for one module. the attribute instance will use
> the attribute-base to bootstrap.
>
> >
> > > Well, I would start by porting a implementation of the c#
> > attributes to
> > > java. Creating an set of classes to represent them in memory. And
> > then
> > > making some way to instanciate them at start up and run time.
> >
> >     That's right.
> >
> > > I would have to look into this in detail. Do you know of any
> > > difficulties that might occur? How important is this for you?
> > > Would you like a more detailed analysis, is it of relevance to the
> > > mission of mono?
> >
> >     Well - I'm not sure of any difficulties (or success probablities
> > <
> > 100.0%).
> >     Guess what - I was having some areas to think over and have got
> > the
> > following idea while I am writing this mail :-)
> >
> >     1. Create necesary classes in Java, NotImplementedAttribute,
> > DefaultEventAttribute etc etc etc.
> >     2. While compiling a class, add the members to the class:
> >         - static attrib_DefaultEvent_Class (= new
> > attrib_DefaultEventAttribute(...)) and then in the static
> > constructor:
> >
> >         static ClassName
> >         {
> >             attrib_DefaultEvent.setProperty(someName) = someValue;
> >         }
> >
> >         This is how to add property at class level. For any property:
> >
> >         - static attrib_NotImplemented_Property_propertyName = new
> > ....
> > followed by
> >
> >         static ClassName
> >         {
> >
> > attrib_NotImplemented_Property_propertyName.setProperty(someName) =
> > someValue;
> >         }
> >
> >     For me - Attributes are quite important - and they play a major
> > role.
> > See for example ASP.Net webserver where there the attributes like
> > DefaultEvent, DefaultProperty, ToolBoxData, DefaultValue, Bindable,
> > TypeConverter - play a very critical role (and Gonzalo should agree
> > with me
> > on this).
>
> Yes of course, such attributes should be possible to load in a run
> time, I dont know how well you wil be able to
>
> >
> >     btw, have you started some work on this... or still in Phase 1?
>
> I am currently looking into representing all of the programs source
> code and metadata data as a library. Right now I have a working version
> on the codification of the GCC ASTS into statements of logic. Then
> current project is to augment the pnet/treecc(tm) compiler so that it
> will support that type of dumping for all treecc applications. The pnet
> team has a (buggy) java compiler in the works.
>
> You can see my project (b)log here :
> http://www.advogato.org/article/696.html
>
> C# Attributes support will be coming along later when the pnet/cscc
> (tm) c# compiler has been infected by the introspector virus.
>
> > > ps :
> > > please feel to ban me from the list for off topic discussion, for
> > > distraction, for  wasting you bytes and time, or for absolutly no
> > > reason at all :)
> >
> >     Well, I find all this pretty interesting.
> Thats great. I just added that so that the mono people will have an
> easy time to ban me. I guess they wont right now, but there are some
> other lists where my crazy ideas and usless comments have gotten me
> banned on.
>
> mike
>
>
> =====
> James Michael DuPont
> http://introspector.sourceforge.net/
>
> __________________________________
> Do you Yahoo!?
> Yahoo! SiteBuilder - Free, easy-to-use web site design software
> http://sitebuilder.yahoo.com
> _______________________________________________
> Mono-devel-list mailing list
> Mono-devel-list at lists.ximian.com
> http://lists.ximian.com/mailman/listinfo/mono-devel-list
>




More information about the Mono-devel-list mailing list