This is the mail archive of the
libstdc++@gcc.gnu.org
mailing list for the libstdc++ project.
Re: patch - implement stdio_fstream
Sounds good to me. I'll add the new ctor and resubmit the patch.
Robert
On Fri, 2004-02-20 at 04:42, Gabriel Dos Reis wrote:
> Jonathan Wakely <cow@compsoc.man.ac.uk> writes:
>
> | On Thu, Feb 19, 2004 at 09:11:06PM -0500, Robert Schweikert wrote:
> |
> | > Here is a patch implementing stdio_fstream, and extension that is
> | > usefull for applications requiring legacy APIs or applications wanting
> | > to attach a stream to a pipe or an open file.
> | >
> | > I am still waiting for the copyright assignment paper work but thought
> | > I'd post the patch to allow people to comment or let me if I need to
> | > make any corrections.
> |
> | How about making this ctor take a std::string ?
> |
> | stdio_ifstream(const char* __s,
> | std::ios_base::openmode __mode = std::ios_base::in)
> |
> | I know there are good reasons not to do this for std::fstream, but this
> | class is an extension anyway.
>
> I suggest overload for both forms. The "const char*" version for
> similarity with other file-stream constructors and the natural
> "const string&" version to remove the pain for the sneaky .c_str()
> construct.
>
> | Thoughts?
> |
> | Similarly, any plans to resurrect Phil's patch to make fstream take
> | std::string, or is that being permanently dropped as a non-standard
> | extension? c.f. http://gcc.gnu.org/ml/libstdc++/2001-06/msg00008.html
>
> I think these issues will be discussed in the revision of the
> standard library.
>
> -- Gaby
--
Robert Schweikert <rjschwei@cox.net>