www.digitalmars.com         C & C++   DMDScript  

digitalmars.D - D roadmap after 1.00

reply vtp <vanh dslextreme.com> writes:
Now that 1.00 is out. When can we expect the next release? And what features
would be in? Concurrency support? GUI support? Network support?  etc.. Would be
nice to know what to expect.

vtp
Jan 04 2007
parent reply janderson <askme me.com> writes:
vtp wrote:
 Now that 1.00 is out. When can we expect the next release? And what features
would be in? Concurrency support? GUI support? Network support?  etc.. Would be
nice to know what to expect.
 
 vtp

Perhaps 1.00 should be stabilized? And then maybe some attention to performance. ie The boring house cleaning jobs. This would go a long way towards retaining D users. -Joel
Jan 04 2007
parent reply "Craig Black" <cblack ara.com> writes:
Yes.  Stabilization should be top priority since there seem to still be many 
outstanding bugs.

However, performance is a big area that isn't necessarily a "boring house 
cleaning" exercise.  For example, the most significant improvement to 
performance IMO would be to make GC precise.  This would require new 
reflection features in the compiler as it would have to know which areas of 
memory are pointers and which are not.  However, I think you may be refering 
to performance enhancements to the backend, which would also be welcome.

-Craig

"janderson" <askme me.com> wrote in message 
news:enjctk$qeq$2 digitaldaemon.com...
 vtp wrote:
 Now that 1.00 is out. When can we expect the next release? And what 
 features would be in? Concurrency support? GUI support? Network support? 
 etc.. Would be nice to know what to expect.

 vtp

Perhaps 1.00 should be stabilized? And then maybe some attention to performance. ie The boring house cleaning jobs. This would go a long way towards retaining D users. -Joel

Jan 04 2007
parent Sean Kelly <sean f4.ca> writes:
Craig Black wrote:
 Yes.  Stabilization should be top priority since there seem to still be many 
 outstanding bugs.
 
 However, performance is a big area that isn't necessarily a "boring house 
 cleaning" exercise.  For example, the most significant improvement to 
 performance IMO would be to make GC precise.  This would require new 
 reflection features in the compiler as it would have to know which areas of 
 memory are pointers and which are not.

An easy fix that should help is to key on element size. This would at least prevent blocks containing elements less than pointer size from being scanned (most notably char strings). In fact, Tango already has this behavior by default. If I had more free time I'd submit a patch to Walter for Phobos--it really isn't a tremendously difficult change aside from sorting out how the GC interface should be modified to support it. Sean
Jan 04 2007