www.digitalmars.com         C & C++   DMDScript  

digitalmars.D - Array operations in D for 1.0

reply "Walter" <newshound digitalmars.com> writes:
A long standing unimplemented feature in D are the array operations. I want
to get 1.0 out, but Norbert Nemec has convinced me that the array operations
feature needs to be rethought out. I want to make it work right so it will
provide better performance opportunity than FORTRAN arrays do, and Norbert
has proposed some intriguing features it might have that could put it well
ahead of FORTRAN and C++.

So, I'm going to drop the array ops for 1.0, and pick up the torch again
after it's out.
Feb 05 2005
next sibling parent zwang <nehzgnaw gmail.com> writes:
There won't be any major changes in D 1.0 except bug-fixes, will there?
I'm also curious about the release date of 1.0 (if scheduled.)


Walter wrote:
 A long standing unimplemented feature in D are the array operations. I want
 to get 1.0 out, but Norbert Nemec has convinced me that the array operations
 feature needs to be rethought out. I want to make it work right so it will
 provide better performance opportunity than FORTRAN arrays do, and Norbert
 has proposed some intriguing features it might have that could put it well
 ahead of FORTRAN and C++.
 
 So, I'm going to drop the array ops for 1.0, and pick up the torch again
 after it's out.
 
 

Feb 05 2005
prev sibling next sibling parent reply "Matthew" <admin stlsoft.dot.dot.dot.dot.org> writes:
Wow! I assumed there were several outstanding important issues, 
including at least the GC/dynamic libs one, to be resolve first.

I understand a release point must come at some arbitrarily chosen point, 
but just not expected it so soon.

"Walter" <newshound digitalmars.com> wrote in message 
news:cu3tnh$2e8$1 digitaldaemon.com...
A long standing unimplemented feature in D are the array operations. I 
want
 to get 1.0 out, but Norbert Nemec has convinced me that the array 
 operations
 feature needs to be rethought out. I want to make it work right so it 
 will
 provide better performance opportunity than FORTRAN arrays do, and 
 Norbert
 has proposed some intriguing features it might have that could put it 
 well
 ahead of FORTRAN and C++.

 So, I'm going to drop the array ops for 1.0, and pick up the torch 
 again
 after it's out.

 

Feb 05 2005
parent reply "Walter" <newshound digitalmars.com> writes:
Isn't the gc DLL issue resolved? (I know the shared library on linux isn't.)

And yes, 1.0 will probably just likely be an arbitrary stake in the ground
at this point.

"Matthew" <admin stlsoft.dot.dot.dot.dot.org> wrote in message
news:cu44ch$71j$1 digitaldaemon.com...
 Wow! I assumed there were several outstanding important issues,
 including at least the GC/dynamic libs one, to be resolve first.

 I understand a release point must come at some arbitrarily chosen point,
 but just not expected it so soon.

Feb 05 2005
parent reply "Matthew" <admin stlsoft.dot.dot.dot.dot.org> writes:
"Walter" <newshound digitalmars.com> wrote in message 
news:cu4550$7km$2 digitaldaemon.com...
 Isn't the gc DLL issue resolved? (I know the shared library on linux 
 isn't.)

Last I heard you'd had an idea, and were going away to mull on it. We lap-witted plebians are eagerly awaiting your prognostications. Or have I missed something?
 And yes, 1.0 will probably just likely be an arbitrary stake in the 
 ground
 at this point.

Feb 05 2005
parent reply "Walter" <newshound digitalmars.com> writes:
"Matthew" <admin stlsoft.dot.dot.dot.dot.org> wrote in message
news:cu45a1$7qn$1 digitaldaemon.com...
 "Walter" <newshound digitalmars.com> wrote in message
 news:cu4550$7km$2 digitaldaemon.com...
 Isn't the gc DLL issue resolved? (I know the shared library on linux
 isn't.)

Last I heard you'd had an idea, and were going away to mull on it. We lap-witted plebians are eagerly awaiting your prognostications. Or have I missed something?

The 0.112 release added support for sharing a single gc instance among multiple DLLs. If there was something else I promised, I can't remember what it was <g>.
Feb 05 2005
parent "Matthew" <admin stlsoft.dot.dot.dot.dot.org> writes:
"Walter" <newshound digitalmars.com> wrote in message 
news:cu47lk$9bb$1 digitaldaemon.com...
 "Matthew" <admin stlsoft.dot.dot.dot.dot.org> wrote in message
 news:cu45a1$7qn$1 digitaldaemon.com...
 "Walter" <newshound digitalmars.com> wrote in message
 news:cu4550$7km$2 digitaldaemon.com...
 Isn't the gc DLL issue resolved? (I know the shared library on 
 linux
 isn't.)

Last I heard you'd had an idea, and were going away to mull on it. We lap-witted plebians are eagerly awaiting your prognostications. Or have I missed something?

The 0.112 release added support for sharing a single gc instance among multiple DLLs. If there was something else I promised, I can't remember what it was <g>.

Gotcha
Feb 05 2005
prev sibling next sibling parent Derek <derek psych.ward> writes:
On Sat, 5 Feb 2005 17:44:15 -0800, Walter wrote:

 A long standing unimplemented feature in D are the array operations. I want
 to get 1.0 out, but Norbert Nemec has convinced me that the array operations
 feature needs to be rethought out. I want to make it work right so it will
 provide better performance opportunity than FORTRAN arrays do, and Norbert
 has proposed some intriguing features it might have that could put it well
 ahead of FORTRAN and C++.
 
 So, I'm going to drop the array ops for 1.0, and pick up the torch again
 after it's out.

Sounds like a good approach. In the meantime, array operations can be written out in long hand ;-) -- Derek Melbourne, Australia
Feb 05 2005
prev sibling next sibling parent Stewart Gordon <smjg_1998 yahoo.com> writes:
Walter wrote:
 A long standing unimplemented feature in D are the array operations. 
 I want to get 1.0 out, but Norbert Nemec has convinced me that the 
 array operations feature needs to be rethought out.

Rewriting the spec so that it makes more sense seems straightforward, and with the spec I had in mind (and imagined was what you really meant in the first place), I'd've thought it a matter of just a handful of lines of code.
 I want to make it work right so it will provide better performance 
 opportunity than FORTRAN arrays do,

Indeed, optimising the performance in the general case would take a bit of work. But doesn't the second paragraph of http://www.digitalmars.com/d/faq.html#q7_3 apply here too?
 and Norbert has proposed some intriguing features it might have that 
 could put it well ahead of FORTRAN and C++.
 
 So, I'm going to drop the array ops for 1.0, and pick up the torch 
 again after it's out.

In that case, I look forward to seeing the spec updated to reflect this.... Stewart. -- My e-mail is valid but not my primary mailbox. Please keep replies on the 'group where everyone may benefit.
Feb 07 2005
prev sibling parent Norbert Nemec <Norbert Nemec-online.de> writes:
Walter wrote:

 A long standing unimplemented feature in D are the array operations. I
 want to get 1.0 out, but Norbert Nemec has convinced me that the array
 operations feature needs to be rethought out. I want to make it work right
 so it will provide better performance opportunity than FORTRAN arrays do,
 and Norbert has proposed some intriguing features it might have that could
 put it well ahead of FORTRAN and C++.
 
 So, I'm going to drop the array ops for 1.0, and pick up the torch again
 after it's out.

Thanks a lot. Now I just hope that I will manage to write down my ideas sometimes soon, so the future of arrays after 1.0 might become a bit clearer.
Feb 07 2005