Shared in D2 has been a stub for ages now.

I'm affraid it will never get implemented if we don't define any priority for it.

D2 is supposed to provide a very elegant way to handle concurency, but, in the current state of the language, it is unusable. If we want that language to succeed (and I believe it has the potential to succeed on the paper) it needs to fulfill its promesses.

On 2/23/2012 5:32 AM, deadalnix wrote:
> Shared in D2 has been a stub for ages now.
> 
> I'm affraid it will never get implemented if we don't define any priority for it.
> 
> D2 is supposed to provide a very elegant way to handle concurency, but, in the current state of the language, it is unusable. If we want that language to succeed (and I believe it has the potential to succeed on the paper) it needs to fulfill its promesses.

The under-way release is focusing on completing the regression eliminations (down to 14, pull requests appreciated) and on const/pure/immutable usability (meaning both dmd implementation issues as well as core phobos issues like Object). Shared will likely be the focus of the following release.  If not the following one, certainly one of the next few as it's definitely one of the major implementation gaps.

Le 23/02/2012 20:45, Brad Roberts a écrit :
> On 2/23/2012 5:32 AM, deadalnix wrote:
>> Shared in D2 has been a stub for ages now.
>>
>> I'm affraid it will never get implemented if we don't define any priority for it.
>>
>> D2 is supposed to provide a very elegant way to handle concurency, but, in the current state of the language, it is
>> unusable. If we want that language to succeed (and I believe it has the potential to succeed on the paper) it needs to
>> fulfill its promesses.
>
> The under-way release is focusing on completing the regression eliminations (down to 14, pull requests appreciated) and
> on const/pure/immutable usability (meaning both dmd implementation issues as well as core phobos issues like Object).
> Shared will likely be the focus of the following release.  If not the following one, certainly one of the next few as
> it's definitely one of the major implementation gaps.
>

It would be awesome. As I mentioned several time, I'm readu to help but still struggling with dmd source code.

If a guru can boostrap me, I would be happy to help.

"deadalnix" <deadalnix@gmail.com> wrote in message news:ji8i55$370$1@digitalmars.com...
>
> It would be awesome. As I mentioned several time, I'm readu to help but still struggling with dmd source code.
>
> If a guru can boostrap me, I would be happy to help.

What would you like to know about it?  Nearly everything I know about how dmd works came from stepping through and printfing the source, and reading the explanations Don somethings leaves in bugzilla comments.

I found rejects-valid bugs to be the easiest to get started on, because you can just search the source code for the error text, insert breakpoints, and use the call stack to work out how it ended up there.

Reading dmd commits can be useful too, it can tell you which parts of the code are responsible for which problems.

If you post specific questions (on the internals mailing list) I'll try to answer them, but it's a little difficult to just explain how it works.

Le 25/02/2012 12:20, Daniel Murphy a écrit :
> "deadalnix"<deadalnix@gmail.com>  wrote in message
> news:ji8i55$370$1@digitalmars.com...
>>
>> It would be awesome. As I mentioned several time, I'm readu to help but
>> still struggling with dmd source code.
>>
>> If a guru can boostrap me, I would be happy to help.
>
> What would you like to know about it?  Nearly everything I know about how
> dmd works came from stepping through and printfing the source, and reading
> the explanations Don somethings leaves in bugzilla comments.
>
> I found rejects-valid bugs to be the easiest to get started on, because you
> can just search the source code for the error text, insert breakpoints, and
> use the call stack to work out how it ended up there.
>
> Reading dmd commits can be useful too, it can tell you which parts of the
> code are responsible for which problems.
>
> If you post specific questions (on the internals mailing list) I'll try to
> answer them, but it's a little difficult to just explain how it works.
>
>

Happy to read that.

I don't want to know something very specific, just pass the entry barrier where I can start doing something useless in it. This is another topic si I'll stop here.

I'll pick up an easy bug this WE and try to solve it.

On 27 February 2012 15:09, deadalnix <deadalnix@gmail.com> wrote:
> Le 25/02/2012 12:20, Daniel Murphy a écrit :
>
>> "deadalnix"<deadalnix@gmail.com>  wrote in message news:ji8i55$370$1@digitalmars.com...
>>>
>>>
>>> It would be awesome. As I mentioned several time, I'm readu to help but still struggling with dmd source code.
>>>
>>> If a guru can boostrap me, I would be happy to help.
>>
>>
>> What would you like to know about it?  Nearly everything I know about how dmd works came from stepping through and printfing the source, and reading the explanations Don somethings leaves in bugzilla comments.
>>
>> I found rejects-valid bugs to be the easiest to get started on, because
>> you
>> can just search the source code for the error text, insert breakpoints,
>> and
>> use the call stack to work out how it ended up there.
>>
>> Reading dmd commits can be useful too, it can tell you which parts of the code are responsible for which problems.
>>
>> If you post specific questions (on the internals mailing list) I'll try to answer them, but it's a little difficult to just explain how it works.
>>
>>
>
> Happy to read that.
>
> I don't want to know something very specific, just pass the entry barrier where I can start doing something useless in it. This is another topic si I'll stop here.
>
> I'll pick up an easy bug this WE and try to solve it.

15 minute crash course worth of documentation: http://prowiki.org/wiki4d/wiki.cgi?DMDSourceGuide


-- 
Iain Buclaw

*(p < e ? p++ : p) = (c & 0x0f) + '0';