I am using MemChk to solve GC in Delphi ;-) I think no overhead. Have a nice day
Regards Leigh www.smootharm.com -----Original Message----- From: delphi-boun...@delphi.org.nz [mailto:delphi-boun...@delphi.org.nz]on Behalf Of Todd Martin Sent: Wednesday, 17 June 2009 3:16 p.m. To: NZ Borland Developers Group - Delphi List Subject: Re: [DUG] Embarcadero article Hi Sean I agree with you, GC is good to have at times. What I am suggesting would allow Delphi to switch on GC, for those that want it, or off, for those that don't. Or even better, provide GC (or not) for a subset of registered classes. Can anyone complain about that? If Delphi provided a notification system which dispatched an event whenever a reference was added or released on a TObject instance, there would be no need for IInterface/IUnknown. TObject would not need any extra methods and therefore would not incur any overhead. Todd. > I don't want to get to deep into a discussion of this, the last time I got into a discussion on GC in non technical, I lost days of my life. However: > - Suggesting that I move to prism to get a gc makes as much sense as suggesting that you move to c++ to get a 64 bit compiler > - As Phil said "It's horse for courses". In the areas you work in, gc is obviously of no use. In the areas I mostly work in, it would make my life much easier. > - Personally I get tired of the "GC = laziness" argument. It's dismissive without any real thought. > - I do know how to use reference counted interfaces and smart pointers. I use them for some things. However reference counting doesn't work for circular references without a lot of additional work. Adding to the opf I use would take more time than it would save. > > Regards > > Sean Cross > CIO > Catalyst Risk Management > > > -----Original Message----- > From: delphi-boun...@delphi.org.nz [mailto:delphi-boun...@delphi.org.nz] On Behalf Of Jolyon Smith > Sent: Wednesday, 17 June 2009 2:28 p.m. > To: 'NZ Borland Developers Group - Delphi List' > Subject: Re: [DUG] Embarcadero article > > I think the vitriol stems from the idea that people believe (and I think > rightly) that the sort of garbage collection that people are asking for is > largely an all or nothing affair. > > i.e. if it's added to the language to please those who want GC (but who - > for some reason - aren't inclined to simply use a runtime environment > provides exactly what they say they want) then that this necessarily > pollutes the language for those who feel that Garbage Collection comes at > too high a price and is just pandering to the inherent laziness in all of us > (myself included - I'd love to be able to think less, but I know that I'd > make more mistakes if I did, not less). > > The distrust of GC as a technology amongst experienced practitioners (as it > typically is) of the black arts of software development stems from: > > - less efficient use of memory (*) > - unpredictably distributed performance across application usage (*) > - loss of determinism (debugging complexity) > > (*) GC assumes a vast over supply of RAM and an availability of application > idle cycles in which to recover wastefully over-allocated RAM. > > > The beauty of the GC framework (ref counted lifetime managed via interfaces) > that Delphi ALREADY OFFERS (!) is that: > > - as you say with 'D', it's optional :) > - it remains deterministic > > > So don't you already have what you want? > All you have to do now is use it. :) > > -- > "Smile", they said. "it could be worse!" > So I did. And it was. > > > -----Original Message----- > From: delphi-boun...@delphi.org.nz [mailto:delphi-boun...@delphi.org.nz] On > Behalf Of Sean Cross > Sent: Wednesday, 17 June 2009 14:02 > To: NZ Borland Developers Group - Delphi List > Subject: Re: [DUG] Embarcadero article > > I do use C# where it makes sense. It just doesn't make sense for the sort > of shareware and desktop apps I mostly do. > > I have never understood the fanatical hatred of garbage collection that some > Delphi developers have (not aimed at anyone in particular but if you either > try discussing gc in non-technical, you will get a large amount of vitriol > heading your way). It makes some things much easier, and some things > harder. Optional GC such as in D seems reasonable to me, but apparently not > to everyone. > > Regards > > Sean Cross > CIO > Catalyst Risk Management > > > -----Original Message----- > From: delphi-boun...@delphi.org.nz [mailto:delphi-boun...@delphi.org.nz] On > Behalf Of Phil Scadden > Sent: Wednesday, 17 June 2009 1:33 p.m. > To: NZ Borland Developers Group - Delphi List > Subject: Re: [DUG] Embarcadero article > > >> Garbage collection is what I really want, but it's a long way down the > list of what will ever get added to Delphi :( >> > Then use C#. I want Delphi as a C++ replacement with as few compromises > as possible, not C#. That said, the C# programming model has things I > wouldnt mind in the language - data bound objects, decent object > inspector and the dynamic interface thingies. Of course, if you are > doing 64 bit, would MPI support be too much to ask as well? > _______________________________________________ NZ Borland Developers Group - Delphi mailing list Post: delphi@delphi.org.nz Admin: http://delphi.org.nz/mailman/listinfo/delphi Unsubscribe: send an email to delphi-requ...@delphi.org.nz with Subject: unsubscribe _______________________________________________ NZ Borland Developers Group - Delphi mailing list Post: delphi@delphi.org.nz Admin: http://delphi.org.nz/mailman/listinfo/delphi Unsubscribe: send an email to delphi-requ...@delphi.org.nz with Subject: unsubscribe
