Re: Garbage collection
- From: "m. Th." <a@xxxxx>
- Date: Tue, 25 Sep 2007 10:09:59 +0300
Peter Morris wrote:
The whole point is that the objects don't have to do anything, they are simply destroyed when they are no longer referenced from some root point such as
A: A global variable
B: Something else that is ultimately referenced from a global variable.
Pete
Yes, I think that I understand this. But my point was to not _enforce_ the need of a GC for all programming elements. As you know very well, a GC has advantages and drawbacks, both technical and human ("no free beer" (TM) ), so, imho, there are sometimes situations in which a GC is recommended, but sometimes not. An on/off global switch per project is in fact rather a one-way road (tell me, what would you do if you have a project with GC enabled and you decide that for eg. speed reasons, avoiding sloppy code aso. you want to disable it? You will traverse the entire code to put .Free everywhere is needed?). That's why I tried to propose an approach to allow the developer to explicitly declare what's GCed or not by using both inclusion and exclusion filtering declarative approach. Imho, the GC situation now is something similar with 'var' declaration explicit in Pascal, implicit in VB. Imho, an 'option explicit' will be a nice thing.
just my 2c,
--
m. th.
.
- Follow-Ups:
- Re: Garbage collection
- From: Peter Morris
- Re: Garbage collection
- References:
- Garbage collection
- From: Peter Morris
- Re: Garbage collection
- From: m. Th.
- Re: Garbage collection
- From: Peter Morris
- Garbage collection
- Prev by Date: Re: Making case for D7->RAD Studio 2007 upgrade
- Next by Date: Re: rant
- Previous by thread: Re: Garbage collection
- Next by thread: Re: Garbage collection
- Index(es):
Relevant Pages
|