Re: GC-optional mode: does anyone care?

[Eric Wasylishen]

Hi David,

First, I wanted to say your work on GC sounds great!

In my mind, the ideal developer/user experience for GC is being able to install 
GNUstep from your OS's package manager, compile an app written for GC, and have 
it work. Or, even better, be able to install apps from the OS's package manager 
which were written for GC along with ones written using retain/release.

As I understand it, in order to support these use cases, we would either need 
OS's to ship GC-only builds of GNUstep base/gui/back, or else ship Optional GC 
builds. If the OS's shipped GC-only builds, it would force everyone to use GC 
(maybe not a problem, but there may be reasons to not want GC.) Does that sound 
accurate?

Cheers,

Eric

On 2011-05-27, at 6:11 AM, David Chisnall wrote:

> Hi Everyone,
> 
> On OS X, you can compile in one of three modes with respect to garbage 
> collection:
> 
> 1 No GC, normal retain/release semantics.
> 2GC-only, no retain/release stuff is used anywhere.
> 3 Optional GC, uses GC mode if everything that is linked supports GC, 
> otherwise uses normal retain/release semantics.
> 
> GNUstep has supported the first mode forever.  I now have the second mode 
> working quite nicely - still needs more testing, but Gorm and all of the 
> Étoilé code that I've tried seems to work.
> 
> So, my question is whether it's worth supporting the GC-optional mode.  Code 
> compiled in this mode will still send retain / release messages, but they'll 
> be ignored if you're in GC mode.  The compiler will always insert the read 
> and write barriers, but the runtime will ignore them if you're not in GC 
> mode.  
> 
> This will require some tweaking in GNUstep code (things that are currently 
> compile-time options will have to be modified to be switched at run time, for 
> example we need both versions of the autorelease pool code.  On OS X, the 
> main advantage is that you can ship binary-only frameworks that work in both 
> modes (albeit with some performance overhead).  I'm not really sure that this 
> is interesting to us, so before I waste time on a feature that no one will 
> use, I thought I'd ask if anyone cares.  
> 
> David
> 
> -- Sent from my PDP-11
> 
> 
> _______________________________________________
> Discuss-gnustep mailing list
> Discuss-gnustep@gnu.org
> https://lists.gnu.org/mailman/listinfo/discuss-gnustep