- - | 2 Nov 2009 19:44
Picon

how to use gitfs?

http://code.google.com/p/gitfs/

how i can create repo?
how to work with it?

more detailed article than the readme would be desirable

Anthony Sorace | 8 Nov 2009 01:04
Picon

inferno acme programs under plan9 acme?

anyone ever glued the pieces together right to get an inferno acme
program running using the namespace exported by plan9's acme?
i suppose going the other direction would be interesting as well,
although it feels somehow less natural.

roger peppe | 8 Nov 2009 04:03
Picon
Gravatar

Re: inferno acme programs under plan9 acme?

at first thought, it should be pretty trivial - just do the
bind of #U/mnt/acme before running your program.
the rest of the p9 namespace won't be in the usual
place though, but that might not be a problem depending
on which program you want to run.

2009/11/7 Anthony Sorace <anothy@...>:
> anyone ever glued the pieces together right to get an inferno acme
> program running using the namespace exported by plan9's acme?
> i suppose going the other direction would be interesting as well,
> although it feels somehow less natural.
>

Venkatesh Srinivas | 9 Nov 2009 22:04
Gravatar

Great Compiler Shootout: thread-ring

Hiya,

I just implemented a solution to the thread-ring
(http://shootout.alioth.debian.org/u32q/benchmark.php?test=threadring&lang=all)
benchmark in Limbo. On very similar hardware to the test machine (Core
2 6700 instead of the 6600 they used), Limbo finished the test in ~9.5
sec, which would put it third, after C with the LWP library and
Haskell / GHC.

Just thought people would be interested.
-- vs

Noah Evans | 9 Nov 2009 22:33
Picon

Re: Great Compiler Shootout: thread-ring

The thread ring benchmark doesn't allow coroutines does it?

On Mon, Nov 9, 2009 at 10:04 PM, Venkatesh Srinivas
<me@...> wrote:
> Hiya,
>
> I just implemented a solution to the thread-ring
> (http://shootout.alioth.debian.org/u32q/benchmark.php?test=threadring&lang=all)
> benchmark in Limbo. On very similar hardware to the test machine (Core
> 2 6700 instead of the 6600 they used), Limbo finished the test in ~9.5
> sec, which would put it third, after C with the LWP library and
> Haskell / GHC.
>
> Just thought people would be interested.
> -- vs
>

Venkatesh Srinivas | 10 Nov 2009 02:37

Polymorphism & value types

Is there any particular reason polymorphism isn't allowed on value types?  Is
there a chance this will be changed in the future?

I have a written up a simple Rendezvous[T], but the most common type to use it
with would be an int... which is impossible.

Thanks,
-- vs

roger peppe | 10 Nov 2009 10:36
Picon
Gravatar

Re: Polymorphism & value types

it's difficult to do polymorphism on value types without
generating type-specific code on the fly.
this is possible (.NET does it for example) but quite
a bit of work to do. until that work is done, polymorphism
is on pointer types only, i'm afraid.

note that most languages that provide polymorphism
on arbitrary types have a universal 1-word representation
of a value.

2009/11/9 Venkatesh Srinivas <me@...>:
> Is there any particular reason polymorphism isn't allowed on value types?
>  Is
> there a chance this will be changed in the future?
>
> I have a written up a simple Rendezvous[T], but the most common type to use
> it
> with would be an int... which is impossible.
>
> Thanks,
> -- vs
>

roger peppe | 10 Nov 2009 10:40
Picon
Gravatar

Re: Great Compiler Shootout: thread-ring

2009/11/9 Noah Evans <noah.evans@...>:
> The thread ring benchmark doesn't allow coroutines does it?

i think inferno processes would probably qualify
as lightwight threads. i don't see them specifying
that the threads must run on multiple processors
simultaneously.

roger peppe | 10 Nov 2009 12:09
Picon
Gravatar

Re: Great Compiler Shootout: thread-ring

venkatesh mentioned he was looking at the Chameneos test,
(http://bit.ly/37uXTm) and i'm ashamed to say i couldn't resist
implementing a version
of it (attached). it runs in 24 seconds on my macbook with N = 6000000.

that's not fast by the standards of the other languages
in the test, but it might be noteworthy that i spent
almost no time debugging the concurrent aspects of the
code (i think i recompiled three times, and they were all
for trivial and easy to find mistakes).

any takers for a better implementation?
Attachment (chameneos-redux.b): application/octet-stream, 2795 bytes
Michaelian Ennis | 10 Nov 2009 17:32
Picon

iota vs. enum

Does anyone (reading) know the reasoning for the decision to use iota
instead of enum?  I thought perhaps it has something to do with the
DIS architecture.

Ian


Gmane