[Cross post from the D news group]

Hi All,

To gain a better understanding of how the D compiler works I am trying to implement support for bit fields in structures. I have modified the parser code to admit the following syntax (for now)...

struct A{
    uint:8 a;
    uint:5 b;
    uint:* unused; //Takes up the remaining space.
    //Raise an error when the computed size is zero.
}

You can see my changes here...https://github.com/SDX2000/dmd

I have been trying to work out a strategy for modifying the compiler code generation for quite some time now but haven't been able to make much headway in this direction.

 I would appreciate any help at this point.

Regards,
Sandeep Datta.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.puremagic.com/pipermail/dmd-internals/attachments/20120109/1cfff264/attachment.html>

Bit fields are already done in D using a library solution:

http://d.digitalmars.com/2.0/phobos/std_bitmanip.html

On 1/8/2012 7:49 AM, Sandeep Datta wrote:
> [Cross post from the D news group]
>
> Hi All,
>
> To gain a better understanding of how the D compiler works I am trying to implement support for bit fields in structures. I have modified the parser code to admit the following syntax (for now)...
>
> struct A{
>     uint:8 a;
>     uint:5 b;
>     uint:* unused; //Takes up the remaining space.
>     //Raise an error when the computed size is zero.
> }
>
> You can see my changes here...https://github.com/SDX2000/dmd
>
> I have been trying to work out a strategy for modifying the compiler code generation for quite some time now but haven't been able to make much headway in this direction.
>
> I would appreciate any help at this point.
>
> Regards,
> Sandeep Datta.
>
>
> _______________________________________________
> dmd-internals mailing list
> dmd-internals at puremagic.com
> http://lists.puremagic.com/mailman/listinfo/dmd-internals
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.puremagic.com/pipermail/dmd-internals/attachments/20120108/7fa640a4/attachment.html>

Hi Walter,

Thanks for the reply. I am aware of the mixin based solution but the main motivation for me at this point is getting to know the D compiler well. And IMHO native support for bitfields will be a welcome new change. I am trying to write an OS/kernel (https://github.com/SDX2000/Deimos) in D and as far as machine hugging code is concerned bitfields and unions are a mainstay. The mixin syntax looks jarring and alien. Please see https://github.com/SDX2000/Deimos/issues/1 for what I am trying to accomplish.

I am particularly curious about the dt_t data structure you seem to be using how does this work?

Regards,
Sandeep Datta.

On Mon, Jan 9, 2012 at 5:53 AM, Walter Bright <walter at digitalmars.com>wrote:

> **
> Bit fields are already done in D using a library solution:
>
> http://d.digitalmars.com/2.0/phobos/std_bitmanip.html
>
>
> On 1/8/2012 7:49 AM, Sandeep Datta wrote:
>
> [Cross post from the D news group]
>
>  Hi All,
>
>  To gain a better understanding of how the D compiler works I am
> trying to implement support for bit fields in structures. I have
> modified the parser code to admit the following syntax (for now)...
>
>  struct A{
>     uint:8 a;
>     uint:5 b;
>     uint:* unused; //Takes up the remaining space.
>     //Raise an error when the computed size is zero.
> }
>
>  You can see my changes here...https://github.com/SDX2000/dmd
>
>  I have been trying to work out a strategy for modifying the compiler
> code generation for quite some time now but haven't been able to
> make much headway in this direction.
>
>  I would appreciate any help at this point.
>
>  Regards,
> Sandeep Datta.
>
>
> _______________________________________________
> dmd-internals mailing listdmd-internals at puremagic.comhttp://lists.puremagic.com/mailman/listinfo/dmd-internals
>
>
> _______________________________________________
> dmd-internals mailing list
> dmd-internals at puremagic.com
> http://lists.puremagic.com/mailman/listinfo/dmd-internals
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.puremagic.com/pipermail/dmd-internals/attachments/20120109/63b72b44/attachment.html>

I'm glad you're aware of it.

Anyhow, the compiler (optimizer and back end) can already deal with bit fields as it is based on a C compiler. The bitfield ops are replaced by mask and shift in the cgelem.c code.

dt_t is simply a way to represent static data.

On 1/9/2012 3:35 AM, Sandeep Datta wrote:
> Hi Walter,
>
> Thanks for the reply. I am aware of the mixin based solution but the main motivation for me at this point is getting to know the D compiler well. And IMHO native support for bitfields will be a welcome new change. I am trying to write an OS/kernel (https://github.com/SDX2000/Deimos) in D and as far as machine hugging code is concerned bitfields and unions are a mainstay. The mixin syntax looks jarring and alien. Please see https://github.com/SDX2000/Deimos/issues/1 for what I am trying to accomplish.
>
> I am particularly curious about the dt_t data structure you seem to be using how does this work?
>
> Regards,
> Sandeep Datta.
>
> On Mon, Jan 9, 2012 at 5:53 AM, Walter Bright <walter at digitalmars.com <mailto:walter at digitalmars.com>> wrote:
>
>     Bit fields are already done in D using a library solution:
>
>     http://d.digitalmars.com/2.0/phobos/std_bitmanip.html
>
>
>     On 1/8/2012 7:49 AM, Sandeep Datta wrote:
>>     [Cross post from the D news group]
>>
>>     Hi All,
>>
>>     To gain a better understanding of how the D compiler works I am
>>     trying to implement support for bit fields in structures. I have
>>     modified the parser code to admit the following syntax (for now)...
>>
>>     struct A{
>>         uint:8 a;
>>         uint:5 b;
>>         uint:* unused; //Takes up the remaining space.
>>         //Raise an error when the computed size is zero.
>>     }
>>
>>     You can see my changes here...https://github.com/SDX2000/dmd
>>
>>     I have been trying to work out a strategy for modifying the compiler
>>     code generation for quite some time now but haven't been able to
>>     make much headway in this direction.
>>
>>     I would appreciate any help at this point.
>>
>>     Regards,
>>     Sandeep Datta.
>>
>>
>>     _______________________________________________
>>     dmd-internals mailing list
>>     dmd-internals at puremagic.com  <mailto:dmd-internals at puremagic.com>
>>     http://lists.puremagic.com/mailman/listinfo/dmd-internals
>
>     _______________________________________________
>     dmd-internals mailing list
>     dmd-internals at puremagic.com <mailto:dmd-internals at puremagic.com>
>     http://lists.puremagic.com/mailman/listinfo/dmd-internals
>
>
>
> _______________________________________________
> dmd-internals mailing list
> dmd-internals at puremagic.com
> http://lists.puremagic.com/mailman/listinfo/dmd-internals
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.puremagic.com/pipermail/dmd-internals/attachments/20120109/f0b9dcef/attachment-0001.html>

Thanks! Sorry, forgot to ask one more detail so I know which list to put him in: is he a fee-only advisor?

Thanks,

Andrei

On 1/9/12 12:10 PM, Walter Bright wrote:
> I'm glad you're aware of it.
>
> Anyhow, the compiler (optimizer and back end) can already deal with bit fields as it is based on a C compiler. The bitfield ops are replaced by mask and shift in the cgelem.c code.
>
> dt_t is simply a way to represent static data.
>
> On 1/9/2012 3:35 AM, Sandeep Datta wrote:
>> Hi Walter,
>>
>> Thanks for the reply. I am aware of the mixin based solution but the main motivation for me at this point is getting to know the D compiler well. And IMHO native support for bitfields will be a welcome new change. I am trying to write an OS/kernel (https://github.com/SDX2000/Deimos) in D and as far as machine hugging code is concerned bitfields and unions are a mainstay. The mixin syntax looks jarring and alien. Please see https://github.com/SDX2000/Deimos/issues/1 for what I am trying to accomplish.
>>
>> I am particularly curious about the dt_t data structure you seem to be using how does this work?
>>
>> Regards,
>> Sandeep Datta.
>>
>> On Mon, Jan 9, 2012 at 5:53 AM, Walter Bright <walter at digitalmars.com <mailto:walter at digitalmars.com>> wrote:
>>
>>     Bit fields are already done in D using a library solution:
>>
>>     http://d.digitalmars.com/2.0/phobos/std_bitmanip.html
>>
>>
>>     On 1/8/2012 7:49 AM, Sandeep Datta wrote:
>>>     [Cross post from the D news group]
>>>
>>>     Hi All,
>>>
>>>     To gain a better understanding of how the D compiler works I am
>>>     trying to implement support for bit fields in structures. I have
>>>     modified the parser code to admit the following syntax (for now)...
>>>
>>>     struct A{
>>>     uint:8 a;
>>>     uint:5 b;
>>>     uint:* unused; //Takes up the remaining space.
>>>     //Raise an error when the computed size is zero.
>>>     }
>>>
>>>     You can see my changes here...https://github.com/SDX2000/dmd
>>>
>>>     I have been trying to work out a strategy for modifying the compiler
>>>     code generation for quite some time now but haven't been able to
>>>     make much headway in this direction.
>>>
>>>     I would appreciate any help at this point.
>>>
>>>     Regards,
>>>     Sandeep Datta.
>>>
>>>
>>>     _______________________________________________
>>>     dmd-internals mailing list
>>>     dmd-internals at puremagic.com  <mailto:dmd-internals at puremagic.com>
>>>     http://lists.puremagic.com/mailman/listinfo/dmd-internals
>>
>>     _______________________________________________
>>     dmd-internals mailing list
>>     dmd-internals at puremagic.com <mailto:dmd-internals at puremagic.com>
>>     http://lists.puremagic.com/mailman/listinfo/dmd-internals
>>
>>
>>
>> _______________________________________________
>> dmd-internals mailing list
>> dmd-internals at puremagic.com
>> http://lists.puremagic.com/mailman/listinfo/dmd-internals
>
>
> _______________________________________________
> dmd-internals mailing list
> dmd-internals at puremagic.com
> http://lists.puremagic.com/mailman/listinfo/dmd-internals