is this a compiler bug ? v0.69 win32

int main( char[][] args ) {
 bit dostuff;
 if ( dostuff = (args.length > 2) ) {
  return 0;
 }
 return 1;
}

//bit_test.d(4): '=' does not give a boolean result

after all the bit is bool postings etc, I would have expected to be able to
write the above code
and not
 if ( (dostuff = (args.length > 2)) == true ) {
  return 0;
 }

This is not about bit types, but more about disallowing '=' as the operator for a boolean test, due to it being usually a typo for '=='.


"Mike Wynn" <mike.wynn@l8night.co.uk> wrote in message news:bi1me1$1n9p$1@digitaldaemon.com...
> is this a compiler bug ? v0.69 win32
>
> int main( char[][] args ) {
>  bit dostuff;
>  if ( dostuff = (args.length > 2) ) {
>   return 0;
>  }
>  return 1;
> }
>
> //bit_test.d(4): '=' does not give a boolean result
>
> after all the bit is bool postings etc, I would have expected to be able
to
> write the above code
> and not
>  if ( (dostuff = (args.length > 2)) == true ) {
>   return 0;
>  }
>
>

On Wed, 20 Aug 2003 23:21:09 -0700 (08/21/03 16:21:09)
, Walter <walter@digitalmars.com> wrote:

> This is not about bit types, but more about disallowing '=' as the operator
> for a boolean test, due to it being usually a typo for '=='.
>
>
> "Mike Wynn" <mike.wynn@l8night.co.uk> wrote in message
> news:bi1me1$1n9p$1@digitaldaemon.com...
>> is this a compiler bug ? v0.69 win32
>>
>> int main( char[][] args ) {
>> bit dostuff;
>> if ( dostuff = (args.length > 2) ) {
>> return 0;
>> }
>> return 1;
>> }
>>
>> //bit_test.d(4): '=' does not give a boolean result
>>
>> after all the bit is bool postings etc, I would have expected to be able
> to
>> write the above code
>> and not
>> if ( (dostuff = (args.length > 2)) == true ) {
>> return 0;
>> }

Or maybe ..

 int main( char[][] args ) {
 bit dostuff;
 if ( dostuff = (args.length > 2), dostuff ) {
   return 0;
 }
 return 1;
 }


-- 
Derek

"Derek Parnell" <derek.parnell@no.spam> wrote in message news:oprt8s4jv759ej19@news.digitalmars.com...
> On Wed, 20 Aug 2003 23:21:09 -0700 (08/21/03 16:21:09)
> , Walter <walter@digitalmars.com> wrote:
>
> > This is not about bit types, but more about disallowing '=' as the
> > operator
> > for a boolean test, due to it being usually a typo for '=='.
> >
> >
> > "Mike Wynn" <mike.wynn@l8night.co.uk> wrote in message news:bi1me1$1n9p$1@digitaldaemon.com...
> >> is this a compiler bug ? v0.69 win32
> >>
> >> int main( char[][] args ) {
> >> bit dostuff;
> >> if ( dostuff = (args.length > 2) ) {
> >> return 0;
> >> }
> >> return 1;
> >> }
> >>
> >> //bit_test.d(4): '=' does not give a boolean result
> >>
> >> after all the bit is bool postings etc, I would have expected to be
able
> > to
> >> write the above code
> >> and not
> >> if ( (dostuff = (args.length > 2)) == true ) {
> >> return 0;
> >> }
>
> Or maybe ..
>
>   int main( char[][] args ) {
>   bit dostuff;
>   if ( dostuff = (args.length > 2), dostuff ) {
>     return 0;
>   }
>   return 1;
>   }
>
I personally hate that form, like that it does not look too bad, but i've seen some "wonderful" uses of comma operator in if's in the past.

> > > This is not about bit types, but more about disallowing '=' as the
> > > operator
> > > for a boolean test, due to it being usually a typo for '=='.
> > >
> > >
> > > "Mike Wynn" <mike.wynn@l8night.co.uk> wrote in message news:bi1me1$1n9p$1@digitaldaemon.com...
> > >> is this a compiler bug ? v0.69 win32
> > >>
> > >> int main( char[][] args ) {
> > >> bit dostuff;
> > >> if ( dostuff = (args.length > 2) ) {
> > >> return 0;
> > >> }
> > >> return 1;
> > >> }
> > >>
> > >> //bit_test.d(4): '=' does not give a boolean result
> > >>
> > >> after all the bit is bool postings etc, I would have expected to be
> able
> > > to
> > >> write the above code
> > >> and not
> > >> if ( (dostuff = (args.length > 2)) == true ) {
> > >> return 0;
> > >> }
> >
> > Or maybe ..
> >
> >   int main( char[][] args ) {
> >   bit dostuff;
> >   if ( dostuff = (args.length > 2), dostuff ) {
> >     return 0;
> >   }
> >   return 1;
> >   }
> >
> I personally hate that form, like that it does not look too bad, but i've seen some "wonderful" uses of comma operator in if's in the past.
>
>

Maybe we should extend if (and also while) such to allows the
following syntax (similar to for):

if (declaration; condition)
{
}

Otherwise yo can always do

{
    declarations;
    if (condition)
    {
    }
}

On Thu, 21 Aug 2003 10:01:43 -0400 (08/22/03 00:01:43)
, Philippe Mori <philippe_mori@hotmail.com> wrote:

>> > > This is not about bit types, but more about disallowing '=' as the
>> > > operator
>> > > for a boolean test, due to it being usually a typo for '=='.
>> > >
>> > >
>> > > "Mike Wynn" <mike.wynn@l8night.co.uk> wrote in message
>> > > news:bi1me1$1n9p$1@digitaldaemon.com...
>> > >> is this a compiler bug ? v0.69 win32
>> > >>
>> > >> int main( char[][] args ) {
>> > >> bit dostuff;
>> > >> if ( dostuff = (args.length > 2) ) {
>> > >> return 0;
>> > >> }
>> > >> return 1;
>> > >> }
>> > >>
>> > >> //bit_test.d(4): '=' does not give a boolean result
>> > >>
>> > >> after all the bit is bool postings etc, I would have expected to be
>> able
>> > > to
>> > >> write the above code
>> > >> and not
>> > >> if ( (dostuff = (args.length > 2)) == true ) {
>> > >> return 0;
>> > >> }
>> >
>> > Or maybe ..
>> >
>> >   int main( char[][] args ) {
>> >   bit dostuff;
>> >   if ( dostuff = (args.length > 2), dostuff ) {
>> >     return 0;
>> >   }
>> >   return 1;
>> >   }
>> >
>> I personally hate that form, like that it does not look too bad, but i've
>> seen some "wonderful" uses of comma operator in if's in the past.
>>
>>
>
> Maybe we should extend if (and also while) such to allows the
> following syntax (similar to for):
>
> if (declaration; condition)
> {
> }
>
> Otherwise yo can always do
>
> {
> declarations;
> if (condition)
> {
> }
> }
>
>
>

Possibly, but I was thinking more along the lines of this ...

As the form if ( <EXPRESSION> )  is valid then it might be just simple enough to do ...

 if ( (a = b) )

in other words, if the assignment expression is enclosed in its own parenthesis, it is considered to be a simple single-term expression, which does provided a valid boolean result.

-- 
Derek

"Philippe Mori" <philippe_mori@hotmail.com> wrote in message news:bi2j25$31bd$1@digitaldaemon.com...
> > > > This is not about bit types, but more about disallowing '=' as the
> > > > operator
> > > > for a boolean test, due to it being usually a typo for '=='.
> > > >
> > > >
> > > > "Mike Wynn" <mike.wynn@l8night.co.uk> wrote in message news:bi1me1$1n9p$1@digitaldaemon.com...
> > > >> is this a compiler bug ? v0.69 win32
> > > >>
> > > >> int main( char[][] args ) {
> > > >> bit dostuff;
> > > >> if ( dostuff = (args.length > 2) ) {
> > > >> return 0;
> > > >> }
> > > >> return 1;
> > > >> }
> > > >>
> > > >> //bit_test.d(4): '=' does not give a boolean result
> > > >>
> > > >> after all the bit is bool postings etc, I would have expected to be
> > able
> > > > to
> > > >> write the above code
> > > >> and not
> > > >> if ( (dostuff = (args.length > 2)) == true ) {
> > > >> return 0;
> > > >> }
> > >
> > > Or maybe ..
> > >
> > >   int main( char[][] args ) {
> > >   bit dostuff;
> > >   if ( dostuff = (args.length > 2), dostuff ) {
> > >     return 0;
> > >   }
> > >   return 1;
> > >   }
> > >
> > I personally hate that form, like that it does not look too bad, but
i've
> > seen some "wonderful" uses of comma operator in if's in the past.
> >
> Maybe we should extend if (and also while) such to allows the
> following syntax (similar to for):
>
> if (declaration; condition)
> {
> }
>
> Otherwise yo can always do
>
> {
>     declarations;
>     if (condition)
>     {
>     }
> }
>
and how exactly is that better (or different) from
if ( expr, var ) { .... }
and for IS an extended while!

my complaint was if ( bit_assign ) is an error
I agree `if ( object_assign ) ...` or `if ( int_assign ) ...` are not `if
( boolean )` and thus an error (or potential error.

> Maybe we should extend if (and also while) such to allows the
> following syntax (similar to for):
>
> if (declaration; condition)
> {
> }

Wow, I thought it did support this!

It must do.

> Otherwise yo can always do
>
> {
>     declarations;
>     if (condition)
>     {
>     }
> }

This is the ugly C++ workaround, but not good enough for D, methinks.

Another imperfection in C++ is the fact that you can declare variables of only one type in a for statement.

We would like to be able to do the following in D:

  for(int i = 0, j, k = 0, long l = 0, m, n = 0; ...)
 {

 } // i, j, k, l, m, n all cease to exist here

Walter, does this present any serious problem for the parser? (I'm assuming it'd be a breeze for a man of your talent.)

"Matthew Wilson" <matthew@stlsoft.org> wrote in message news:biojkj$ah0$1@digitaldaemon.com...
> > Maybe we should extend if (and also while) such to allows the
> > following syntax (similar to for):
> >
> > if (declaration; condition)
> > {
> > }
>
> Wow, I thought it did support this!

Nope.

> It must do.

Why?

> > Otherwise yo can always do
> >
> > {
> >     declarations;
> >     if (condition)
> >     {
> >     }
> > }
>
> This is the ugly C++ workaround, but not good enough for D, methinks.
>
> Another imperfection in C++ is the fact that you can declare variables of only one type in a for statement.
>
> We would like to be able to do the following in D:
>
>   for(int i = 0, j, k = 0, long l = 0, m, n = 0; ...)
>  {
>
>  } // i, j, k, l, m, n all cease to exist here
>
> Walter, does this present any serious problem for the parser? (I'm
assuming
> it'd be a breeze for a man of your talent.)

There are some ambiguity problems using typedef'd names after the ,.