Re: use of boost::tie and OUT-arguments.

2004-02-25 Thread Andre Poenitz
On Mon, Feb 23, 2004 at 03:47:44PM +, Angus Leeming wrote: > Andre Poenitz wrote: > >> class tull; > >> std::pair ReturnVar(); > > > > This is illegal IMHO even if g++ and Comeau (*gosh*) accept it. > > [Have we had this discussion before?] > > Don't think so. Show me where it says that the t

Re: use of boost::tie and OUT-arguments.

2004-02-25 Thread Andre Poenitz
On Mon, Feb 23, 2004 at 03:27:50PM +0100, Lars Gullik Bjønnes wrote: > this is mine: > > -- > #include > > class tull; > std::pair ReturnVar(); > > class tull {}; > > int main() { > std::pair ptt = ReturnVar(); > } > > std::pair ReturnVar() { > return std::make_pai

Re: use of boost::tie and OUT-arguments.

2004-02-23 Thread Angus Leeming
Andre Poenitz wrote: >> class tull; >> std::pair ReturnVar(); > > This is illegal IMHO even if g++ and Comeau (*gosh*) accept it. > [Have we had this discussion before?] Don't think so. Show me where it says that the template must be instantiated in a forward declaration. Even code like this is

Re: use of boost::tie and OUT-arguments.

2004-02-23 Thread Andre Poenitz
On Mon, Feb 23, 2004 at 04:26:10PM +0100, Andre' Poenitz wrote: > On Mon, Feb 23, 2004 at 03:27:50PM +0100, Lars Gullik Bjønnes wrote: > > Andre Poenitz <[EMAIL PROTECTED]> writes: > > > > | I am fine with this in this particular case, but not with the underlying > > | principle ("out parameters a

Re: use of boost::tie and OUT-arguments.

2004-02-23 Thread Andre Poenitz
On Mon, Feb 23, 2004 at 03:27:50PM +0100, Lars Gullik Bjønnes wrote: > Andre Poenitz <[EMAIL PROTECTED]> writes: > > | I am fine with this in this particular case, but not with the underlying > | principle ("out parameters are bad") for the usual reasons: losing the > | ability to forward-declare

Re: use of boost::tie and OUT-arguments.

2004-02-23 Thread Lars Gullik Bjønnes
Andre Poenitz <[EMAIL PROTECTED]> writes: | I am fine with this in this particular case, but not with the underlying | principle ("out parameters are bad") for the usual reasons: losing the | ability to forward-declare the parameter (which does not apply here) You do? An example please. this is

Re: use of boost::tie and OUT-arguments.

2004-02-23 Thread Andre Poenitz
On Sun, Feb 22, 2004 at 04:46:00PM +0100, Lars Gullik Bjønnes wrote: > Alfredo Braunstein <[EMAIL PROTECTED]> writes: > > | Lgb wrote: > > > >> I have never liked OUT arguements... and when I see code like: > >> > >> ParagraphList::iterator beg, end; > >> getSelectionSpan(cur, *this, beg, end

Re: use of boost::tie and OUT-arguments.

2004-02-22 Thread Lars Gullik Bjønnes
Alfredo Braunstein <[EMAIL PROTECTED]> writes: | Lgb wrote: > >> I have never liked OUT arguements... and when I see code like: >> >> ParagraphList::iterator beg, end; >> getSelectionSpan(cur, *this, beg, end); >> >> I get a chilly feeling (in the shoulder region). > | [...] > >> (also why c

Re: use of boost::tie and OUT-arguments.

2004-02-22 Thread Alfredo Braunstein
Lgb wrote: > I have never liked OUT arguements... and when I see code like: > > ParagraphList::iterator beg, end; > getSelectionSpan(cur, *this, beg, end); > > I get a chilly feeling (in the shoulder region). [...] > (also why coudn't cur and text be const refs in this funtion?) Both val