fstream constructor

[Chris Theis]

Hello,

as stated by the standard the fstream derivatives (ifstream & ofstream)
expect a const char*
parameter indicating the file name associated with the stream to be opened.
I wondered why there is no constructor accepting a const string& parameter,
which forces the user to convert any filename to an old C-style string
before passing it to the constructor. Is this simply a historical legacy or
is there some deeper problem, which I might fail to see. Probably somebody
could shed some light on this issue.

Thanks
Chris

[Pete Becker]

Re: fstream constructor

Chris Theis wrote:[color=blue]
>
> as stated by the standard the fstream derivatives (ifstream & ofstream)
> expect a const char*
> parameter indicating the file name associated with the stream to be opened.
> I wondered why there is no constructor accepting a const string& parameter,
> which forces the user to convert any filename to an old C-style string
> before passing it to the constructor. Is this simply a historical legacy or
> is there some deeper problem, which I might fail to see. Probably somebody
> could shed some light on this issue.
>[/color]

There are seventeen functions in the C++ standard that take const char*
but don't have analogs that take strings. Versions taking strings will
probably be added in the next standard.

--

Pete Becker
Dinkumware, Ltd. (http://www.dinkumware.com)

[Chris Theis]

Re: fstream constructor

"Pete Becker" <petebecker@acm .org> wrote in message
news:qqSdnUOTof snlhbenZ2dnUVZ_ sCdnZ2d@rcn.net ...[color=blue]
> Chris Theis wrote:[color=green]
>>
>> as stated by the standard the fstream derivatives (ifstream & ofstream)
>> expect a const char*
>> parameter indicating the file name associated with the stream to be
>> opened.
>> I wondered why there is no constructor accepting a const string&
>> parameter,
>> which forces the user to convert any filename to an old C-style string
>> before passing it to the constructor. Is this simply a historical legacy
>> or
>> is there some deeper problem, which I might fail to see. Probably
>> somebody
>> could shed some light on this issue.
>>[/color]
>
> There are seventeen functions in the C++ standard that take const char*
> but don't have analogs that take strings. Versions taking strings will
> probably be added in the next standard.
>[/color]

Thanks Pete for confirming what I had already expected. I just had expected
that the revision of the standard in 2003 would/could have introduced this,
but naturally there are certainly more important things to cover.

Cheers
Chris