Walter Bright Wrote:

> Some new stuff to make writing templates easier.
> 
> http://www.digitalmars.com/d/1.0/changelog.html http://ftp.digitalmars.com/dmd.1.031.zip
> 
> http://www.digitalmars.com/d/2.0/changelog.html http://ftp.digitalmars.com/dmd.2.015.zip

Walter thanks for the updates!

In testing my D v1.0 project source code, I found that two functions were removed between v1.028 and the v1.029 releases, the bool isPosZero(real) and bool isNegZero(real) functions that were in std.math. Since D v1.0 shouldn't be changing much, and just mainly bug fixes for the most part, I normally wait a few releases before testing my code against them. Thus looking at std.math I see a lot of nice improvements, but feel the non-private labeled functions should remain in now that D v1.0 as been reached.

A two simple wrapper functions should be put back in so existing D v1.0 code doesn't break...in this case of course I could simply add in replacement functions for these in my own code. But the point being that functions should not be removed at this point.

I think the following code could replace these two functions using the new changes found in std.math:

bool isPosZero(real x)
{ return isIdentical(x, 0.0); }

bool isNegZero(real x)
{ return isIdentical(x, -0.0); }

Thanks, keep up the good work!
David L. Davis

David L. Davis wrote:
> A two simple wrapper functions should be put back in so existing D
> v1.0 code doesn't break...in this case of course I could simply add
> in replacement functions for these in my own code. But the point
> being that functions should not be removed at this point.

I'll ask Don about it, who has been maintaining std.math.

David L. Davis wrote:
> Walter Bright Wrote:
> 
>> Some new stuff to make writing templates easier.
>>
>> http://www.digitalmars.com/d/1.0/changelog.html
>> http://ftp.digitalmars.com/dmd.1.031.zip
>>
>> http://www.digitalmars.com/d/2.0/changelog.html
>> http://ftp.digitalmars.com/dmd.2.015.zip
> 
> Walter thanks for the updates!
> 
> In testing my D v1.0 project source code, I found that two functions were removed between v1.028 and the v1.029 releases, the bool isPosZero(real) and bool isNegZero(real) functions that were in std.math. Since D v1.0 shouldn't be changing much, and just mainly bug fixes for the most part, I normally wait a few releases before testing my code against them. Thus looking at std.math I see a lot of nice improvements, but feel the non-private labeled functions should remain in now that D v1.0 as been reached.
> 
> A two simple wrapper functions should be put back in so existing D v1.0 code doesn't break...in this case of course I could simply add in replacement functions for these in my own code. But the point being that functions should not be removed at this point.
> 
> I think the following code could replace these two functions using the new changes found in std.math:
> 
> bool isPosZero(real x)
> { return isIdentical(x, 0.0); }
> 
> bool isNegZero(real x)
> { return isIdentical(x, -0.0); }
> 
> Thanks, keep up the good work!
> David L. Davis
> 

isPosZero() and isNegZero() were private, and used only for internal unit tests. (If they were exposed, it was a bug).
Your code above is correct, though.

Don Wrote:

> David L. Davis wrote:
> > A two simple wrapper functions should be put back in so existing D v1.0 code doesn't break...in this case of course I could simply add in replacement functions for these in my own code. But the point being that functions should not be removed at this point.
> > 
> > I think the following code could replace these two functions using the new changes found in std.math:
> > 
> > bool isPosZero(real x)
> > { return isIdentical(x, 0.0); }
> > 
> > bool isNegZero(real x)
> > { return isIdentical(x, -0.0); }
> > 
> > Thanks, keep up the good work!
> > David L. Davis
> > 
> 
> isPosZero() and isNegZero() were private, and used only for internal
> unit tests. (If they were exposed, it was a bug).
> Your code above is correct, though.

Hi Don,

   Got it! Thanks for your reply...I'll add the two now missing/modified functions into my existing code. Leave it to me to find public scoped privates to use in my code. :P But you're right, the comments beside the function names did state that they were used only for unit-tests. So with that thought, they were clearly meant to be privates based on the comments alone.

   Please keep on the good work on std.math!!

David L. Davis

David L. Davis wrote:
> Don Wrote:
> 
>> David L. Davis wrote:
>>> A two simple wrapper functions should be put back in so existing D v1.0 code doesn't break...in this case of course I could simply add in replacement functions for these in my own code. But the point being that functions should not be removed at this point.
>>>
>>> I think the following code could replace these two functions using the new changes found in std.math:
>>>
>>> bool isPosZero(real x)
>>> { return isIdentical(x, 0.0); }
>>>
>>> bool isNegZero(real x)
>>> { return isIdentical(x, -0.0); }
>>>
>>> Thanks, keep up the good work!
>>> David L. Davis
>>>
>> isPosZero() and isNegZero() were private, and used only for internal unit tests. (If they were exposed, it was a bug).
>> Your code above is correct, though.
> 
> Hi Don,
> 
>    Got it! Thanks for your reply...I'll add the two now missing/modified functions into my existing code. Leave it to me to find public scoped privates to use in my code. :P But you're right, the comments beside the function names did state that they were used only for unit-tests. So with that thought, they were clearly meant to be privates based on the comments alone.

Phew! I'm glad they don't need to be retained, hopefully no-one else was bitten by that bug.
> 
>    Please keep on the good work on std.math!!
> 
> David L. Davis